flHHlllllHlllllIllllHlillllllllllllllllHHllHlllllUllHfl

3 1293 01027 6032

 

_UBRARY
M'Chigan State
University

 

 

 

PLACENHE‘I’URNBOXto woman othbchockodfmmywrocord.
INE Mummorbdoroddoduo.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1   .
BEE?  

E

 

 

 

 

 

 

#1

MSU IOAnAfflrmd AdlonlEqu Opponu flylmtltuion

‘f‘

MODIFICATION OF THE COMPLEX FIGURE TEST
AND THE EVALUATION OF AGE AND SEX DIFFERENCES IN VISUAL MEMORY

By

Philip Stephen Fastenau

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Depart-ent of Psychology

1993

ABSTRACT

MODIFICATION OF THE COMPLEX FIGURE TEST
AND THE EVALUATION OF AGE AND SEX DIFFERENCES IN VISUAL MEMORY

By

Philip Stephen Fastenau

Recognition and matching trials were designed for the Rey-
Osterrieth Complex Figure Test (CFT) to help clinicians differentiate
construction, perception, encoding, consolidation, and retrieval
functions in visual-spatial memory performance. The trials were used to
examine age and sex differences in visual memory using 90 community-
dwellers (equal numbers of men and women in all but one quinquennium,
ages 30 to 80). Alpha was set at .001 for all analyses to control
experimentwise Type I error.

The Recognition Trial has 30 multiple-choice items for different
elements of the figure. Alpha reliabilities were .84, .61, .59, and .66
for the Total, Global, Left Detail, and Right Detail scales,
respectively. Total Score correlated .71-.74 with CFT recall and .50-
.57 with Visual Reproductions (VR). Subscale design was confirmed by
factor analysis, and Total Score loaded on visual-spatial memory
factors. Scores distributed normally with strong item-total
correlations and item difficulties.

The Matching Trial has 10 multiple-choice matching-to-sample items.
Alphas were .58, .40, and .08 for the Total, Right Detail, and Left
Detail scales. Total Score correlated .39 with VB Matching and .46 with
Judgment of Line Orientation. Zero-variance items precluded analysis of
the subscale structure. Total Score loaded on broad cognitive factors

which included visual-spatial perception. Test scores were skewed

negatively, and item difficulty indices were low; corrected item-total
correlations were moderate.

These tests and others helped determine whether reduced attention
capacity explains age-related decline in secondary memory. 0n auditory-
verbal memory, no age effect was observed on consolidation;
encoding/storage varied by age, but the effect was due entirely to
attention deficits; only retrieval showed age differences that could not
be explained by attention. On visual-spatial memory, no age effects
were observed in encoding/storage, consolidation, or retrieval after
perception and construction were controlled. Consequently, older memory
may best be sharpened by increasing attention capacity and using visual
imagery.

Finally, men outperformed women on CFT Immediate recall only
(52:.09). This result adds to many inconsistent and small sex

differences in visual-spatial memory research.

Copyright by

PHILIP STEPHEN FASTENAU

1994

DEDICATIONS

This work is dedicated to the glory of God in the name of my Lord and
Savior Jesus Christ. I also dedicate this to my parents, Doris P.
Fastenau and the late Rev. Emmett H. Fastenau, who have always
encouraged me to learn and grow, and to my brothers and sisters who have
supported me in every possible way throughout my graduate training. -

This is further dedicated to my wife, Dana Atkinson Fastenau.

Finally, this dissertation is dedicated to all of my many mentors in
life, but especially James F. "Mr. D" DeBruhl, E. Keith Bramlett,
Michael J. Fimian, Kenneth D. Jenkins, Norman Abeles, and Lauren Julius

Harris.

ACKNOWLEDGMENTS

I recognize Alexander A. Manning and the staff/administration of
Broughton State Hospital, Morganton, North Carolina, for their support
in the pilot study; Jane Holmes Bernstein for her review of the original
recognition items; Natalie L. Denburg for her close collaboration on
this study and on related projects; Norman Abeles, Lauren Julius Harris,
Neal Schmitt, and Bertram Karon for their invaluable contributions to
the design of the standardization study and the interpretation of these
results; Lidia A. Domtrovic and Michael J. Finton for testing some of
the subjects in the standardization sample; Linda E. Sloan, Eric A.
Fertuck, and Jennifer Winer who provided substantial clerical assistance
in the standardization study; the APA Science Directorate for partially
funding the standardization study with a research award; Aaron Beck, The
Psychological Corporation, and Consulting Psychologists Press for
subsidizing the testing aaterials; and Roger Halley for his
resourcefulness in supporting this research. I also appreciate the
participation of the members of Holt Church of the Nazarene, Holt, MI;
Our Savior Lutheran Church, Lansing, MI; Saint Luke Lutheran Church,
Haslett, MI; and the Shaarey Zedek Congregation, East Lansing, MI.
Finally, I am indebted to my wife, Dana Atkinson Fastenau, for her

loving support and encouragement.

vi

TABLE OF CONTENTS

List of Tables ................................................... xi
List of Figures ........ ......... ... ..... ......................... xiii
Introduction ...................................................... 1
Rationale for Modifying the CFT .................................. 3
Limitations of the OFT ...................................... 5
Clinical Remedies ........................................... 7
Alternative Measures of Visual Memory .................. 7
Inferences from Other Tests ............................ 10

Extended CFT: A Synthesis of Assets ........................ 11
Neuroanatomical Correlates of Relevant Cognitive Behaviors ....... 12
Perception .................................................. 13
Attention ................................................... 15
Construction ................................................ 21

nguage 0.....0...0.0..00.0.0...0..........OOOOOOOODOOOOOOO. 23

memory ......OOOOOOOOOOO.....IOOOOOOOOOOOOOOOO.......OOOOOOOO 25

su-lary 00.00....0...........OOOOOOOOOOOOOO......OOODODOOO... 27

Age Differences in Visual-Spatial Memory and Related Variables ... 29
Memory Stores ............................................... 30
Memory Processes ............................................ 32

Consolidation .......................................... 32

EDCOding and Retrieval 00.0.0000...OOOOOOOOOOOOOOOOOOOOO 32

vii

Variables That Affect Memory Performance .................... 37
Attention .............................................. 37
Perception and Production .............................. 38
Health Factors ......................................... 39
Emotional Influences ................................... 43

Sex Differences and Visual-Spatial Memory ........................ 45
Development of the New Measures .................................. 50

Item Construction ........................................... 50
Constructional Elements ................................ 50
Constructional Errors .................................. 50

Pilot Study ................................................. 52

Expert Appraisal ............................................ 54

Final Revision .............................................. 55

Method ........................................................... 59

Subjects .................................................... 59

Instruments ................................................. 60

Procedure ................................................... 75
Recruitment ............................................ 75
Scheduling and Everyday Memory Questionnaires .......... 76
Introduction ........................................... 76
CFT Segment ............................................ 78
Break Between Segments ................................. 86
HMS-R Segment .......................................... 86

Results .......................................................... 95

Prediction 4a ............................................... 96

Prediction 4b ............................................... 97

Predictions 00.......0000...0.0.......OOOOOOOOOOOOOOOOOOOOO 100

viii

Prediction 6 ................... ...........................
Prediction 1 ....... .. ...................................
Prediction 2 .................. . ......... . ....... . . .. .
Prediction 3 . ..................... .... ...... ... ........
Prediction 7 . . .............. . . .......... . .. .....
Prediction 8a ............. .. . ......... ...... . . ......
Prediction 8b ...... . . ........... .. .... ......... . ........ .
Prediction 8c .............. . ............ . ................. ..
Prediction 8d ............ .. . ............ . . . .........
Prediction 8e ......... . ............. . .......................
Prediction 8f ..... . ........ . ......... . ..... .. ......... ......
Prediction 9a . ...... . ........ ...... .............. ... .. .
Prediction 9b . . .............................. ....... ...
Prediction 9c . . . .......... ....... ............ . . . ..... .
Prediction 9d . . .. . . . ....... .. ............... . .. .
Prediction 9e . . .. ........... . ........ . ..................
Prediction 9f ......... ........ .. . ...... . .. ..... ... . .
Discussion ............................ . .. . . . ... ...

Age Differences (Predictions 1 - 6)

Sex Differences (Prediction 7)

Psychometric Properties of the CFT (Predictions 8 and 9) ....

Appendix A:
Appendix B:
Appendix C:
Appendix D:
Appendix E:

Appendix F:

Neuropsychological Survey
Letter Cancellation Stimuli

Symbol Cancellation Stimuli

ix

Scoring Criteria for the CFT Drawings .

Matching Items for the Complex Figure Test

Recognition Items for the Complex Figure Test

100
100
103
107
109
111
113
115
124
129
131
136
137
137
138
138
139
142
142
150
150
156
187
198
203
210

214

Appendix G: HMS-R VR Recognition/Matching Stimuli ............... 218
Appendix H: Cowboy Story Stimuli and Scoring .................... 224
Appendix 1: Confirmation Letter to Participating Organizations ... 228
Appendix J: Consent Form ........................................ 230
Appendix K: Debriefing Form ..................................... 232

List of References ............................................... 234

LIST OF TABLES

Table 1: Classification of Recognition Items by Element

and by Quadrant ....................................... 53
Table 2: Demographic Data for the Sample, by Age Group ......... 61
Table 3: Frequency of Health Risks, by Age Group ............... 62
Table 4: Listing of Tests by the Constructs They Purportedly

Measure ............................................... 63
Table 5: Sequential Ordering of Tasks and Approximate Nuaber of

Minutes Required ...................................... 64
Table 6: Correlations with Age for Attention and Perception, by

Modality (Prediction 4a) .............................. 98
Table 7: Correlations Between Age and Production, by Modality

(Prediction 4b) ....................................... 99
Table 8: Correlations with Age for Health and Emotional

Factors, by Direction of Expected Relationship

(Predictions 5 and 6) ................................. 101
Table 9: Age Effects in Consolidation, by Dependent Variable

(Prediction 1) ........................................ 104
Table 10: Age Effects in Retrieval Before Controlling for

Attention, by Dependent Variable (Prediction 2) ....... 106
Table 11: Age Effects in Encoding/Storage, by Dependent Variable

(Post Hoe) 00............OOOOOOOOOO.......OOOOOOOOOOOOO 110

xi

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

12:

13:

14:

15:

16:

17:

18:

19:

20:
21:

22:

23:

25:

26:

Cronbach's Alpha Reliabilities for CFT Scales and

Subscales, by Trial ................................... 112
Intercorrelations Among CFT Scales and Subscales,

by Trial .............................................. 114
A Priori Factor Structure of Selected Measures, by

Trait and Modality .................................... 116
Descriptive Data for Test Scores Used in

Factor Analysis ....................................... 117
Correlations Among the Test Scores Used in

Factor Analysis ................................... 118-119
Test Loadings on an Eight-Factor Solution Using

SPSS Factor Analysis .................................. 121
Test Loadings on a Two-Factor Solution Using SPSS

Factor Analysis ....................................... 123
Test Loadings on a Four-Factor Solution Using SPSS

Factor Analysis ....................................... 125
Descriptive Data for the CFT Recognition Items ........ 126
Correlations Among the CFT Recognition Items, by

Item Number ....................................... 127-128
Recognition Item Loadings on Global and Detail

Factors Using LISREL .................................. 130
Descriptives for CFT Recognition (Raw Scores),

by Scale .......................................... 132-134
Discrimination Ability of the CFT Recognition Items ... 135
Descriptives for CFT Matching Total Scale

(Raw Scores) .......................................... 140
Discrimination Ability of the CFT Matching Items ...... 141

xii

FIGURE

Figure 1: Complex Figure Test Stimulus ..........................

xiii

INTRODUCTION

Although there are many clinical instruments for assessing visual-
spatial memory, all of them have limitations. In this investigation, I
modify one of those tests, the Complex Figure Test (CFT), to be a more
comprehensive tool for clinical assessment and for the theoretical
investigation of memory functions. This project had three purposes.
First, I designed supplementary trials for the CFT and examined their
psychometric properties using a sample of 90 healthy, community-dwelling
adults ages 30-80.

Second, as an application of the modified instrument, I used it to
test a hypothesis about memory decline in advanced age. It has been
observed that the ability to learn new information (secondary memory)
declines with age, beginning in the 503 and becoming more evident in the
603 (Albert, 1984; Poon, 1985). This decline in secondary memory
parallels age-related decline on some measures of attention. Because
secondary memory involves effortful processing, it was hypothesized here
that age-related decline in secondary memory is mediated, at least in
part, by age-related declines in attentional resources. After
controlling for age-related decline in other cognitive abilities that
contribute to memory performance (perception and production), I used the
modified CFT to decompose secondary memory performance into three
putative processes (encoding/storage, consolidation, and retrieval) in

each of two modalities (auditory-verbal and visual-spatial).

2
Finally, because some sex differences have been observed on some
visual-spatial tasks, the performances of men and women were compared on
the CFT data. The measurement of age and sex differences was also used

to guide stratification of norms for the extended CFT for clinical use.

RATIONALE FOR MODIFYING THE CFT

The Complex Figure Test (CFT; Rey, 1941; Osterrieth, 1944) was
designed to measure both perceptual organization and visual memory in
brain-injured persons (Lezak, 1983). It is a standard component in many
neuropsychological batteries (Kaplan, 1988; Lezak, 1983; Squire, 1986;
Heintraub & Mesulam, 1985). In the administration of this test,
individuals are asked to copy a complex geometric stimulus (Figure 1) on
a blank sheet of white paper using five or six colored pens, which are
presented one at a time by the exaliner. The exaliner can track the
subjects’ constructional approach by listing the sequence of lines as
they are drawn; perceptual organization is inferred from the type of
constructive procedure employed by the subjects during the copy
production. The examiner also notes the time that lapsed between start
and finish. Immediately following the copy trial, the stimulus and copy
are removed. The examiner provides a new blank sheet and asks the
subjects to draw the image from memory (immediate recall). After 20 to
60 minutes, the memory trial is repeated (delay recall); the length of
the delay is inconsequential within these limits (Lezak, 1983).

Research and clinical applications of the CFT have shown it to be
sensitive to certain neuropsychological functions. For example, Haber
and Holmes (1985, 1986) used the instrument with children to describe
developmental patterns in organization and construction. The CFT has

also been used with adults, not only to differentiate brain-injured

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Eign£g_1. Complex Figure Test stimulus. Reproduced with permission of
Swets & Zeitlinger BV, The Netherlands.

5
groups from nonimpaired comparison groups but also to differentiate
between groups with different pathologies. Such studies have
demonstrated some correspondence between the location/type of
neurological insult and certain characteristic patterns of
fragmentation, neglect, rotation, and distortion on this task (Binder,
1982; Brouwers, Cox, Martin, Chase, & Fedio, 1984; Kaplan, 1988; Lezak,

1983; Milberg, Hebben, & Kaplan, 1986).

Limitations of the CFT

Although the CFT can be an invaluable tool in the neuropsycho-
logist’s repertoire, it has limitations. First, its dependence on
constructional ability can confound interpretations regarding memory.
Individuals who have trouble drawing the figures in the copy trial are
destined to score low on the immediate and delay memory trials. For
example, among neurological patients with hemiparesis or with apraxia,
low memory scores may be indicative of drawing or construction
limitations only, thereby masking intact visual memorization skills. To
circumvent these obstacles, supplementary measures of visual memory
should be included that are independent of complex motor responses;
alternative approaches should allow for a pointing or even spoken
response.

Second, in its present use, the test does not discrilinate between
encoding and post-encoding deficits. For patients with low memory
scores, the problem may be getting information into memory
(encoding/storage) or making the memory trace more permanent
(consolidation) or accessing it later (retrieval). There is evidence

that the encoding and post-encoding processes are independent of one

6

another or that they at least operate upon different types of
information (Craik, 1979). This has been observed in developmental
research comparing the efficiencies of these processes at different ages
(Craik a McDowd, 1987; Howe & Hunter, 1985) and in research with
learning disabled children (Howe, O’Sullivan, Brainerd, & Kingma, 1989).
Because a recognition task provides additional cues to facilitate
retrieval, comparisons between free recall and recognition performances
clarify the relative contributions of encoding and post-encoding
processes to total memory performance. Consequently, it is important to
include a measure of recognition as well as a measure of free recall.

Jacoby (1984) argued against separate assessment of the three
processes (encoding, storage, and retrieval), proposing instead that
there are no differences between stages with regard to the type of
processing; strategies, if employed, are active and elaborative
irrespective of the stage at which they are invoked. Differences in
types of amnesias, he claims, are due to whether the patient engages
such processing at one or all stages. This explanation supposedly casts
the focus away from individual stages and onto a comprehensive
metamemory (that is, the executive regulation of memory processes).
However, Jacoby’s argument is deceptive; although his revised paradigm
looks parsimonious, the fact that some patients spontaneously apply
elaborative processing at one stage and not another underscores the
necessity for examining each of the stages.

In addition to its theoretical significance, the assessment of
processes at different stages has clinical relevance. First, in
rehabilitation settings, the specific nature of the deficiency--encoding

versus post-encoding--can have important implications for intervention.

7
Second, including both recall and recognition measures increases the
diagnostic sensitivity of the memory test:
Failure to recognize words [or figures, in the case of nonverbal
memory] as having been previously presented is a more reliable sign
of memory disorder than failure to recall, and recognition failure
denotes a more severe disorder as well. The use of recall and

recognition tests together makes it possible to detect subtle,
early signs of impairment (Squire, 1986, p. 280).

Clinical Remedies

Alternative Measures of Visual Megggy

To address these problems of visual memory, clinicians may use
nonverbal memory tests other than the CFT. These measures, however,
have their own shortcomings when sensitive memory evaluation is the
goal. One of these measures was devised by Kimura (1963). The task
effectively discriminated left- from right-temporal patients.
Unfortunately, the nature of this task (purely a recognition measure)
precludes an assessment of recall, without which the relative
efficiencies of encoding and post-encoding processes cannot be compared.

Another recognition task, originally developed for non-human
primates, was modified for clinical use with dementing patients. The
Delayed Recognition Span Test (DRST) measures recognition across five
types of stimuli: verbal, spatial, color, pattern, and faces (Moss,
Albert, Butters, & Payne, 1986). Despite the advantages of testing so
many relevant stimulus domains, this instrument--1ike the Kimura test--
does not include a recall component.

The Benton Visual Retention Test (BVRT; Benton, 1974) is similar to
the CFT. Like the CFT, however, it measures only recall and requires
the subjects to draw their responses. The BVRT has also been criticized

for the simplicity of its stimuli (Hemsley, 1974; Lezak, 1983). It has

8
been found that simple visual images are frequently encoded in verbal
concepts (Reed, 1974). Many investigators have argued that the BVRT
stimuli, in particular, are susceptible to verbal encoding so that
recall performance on this task is not necessarily a pure measure of
visual memory. In fact, BVRT scores improve with language recovery in
aphasics (indicative of verbal mediation); in addition, performance on
this test correlates more with measures of constructional ability than
with measures of memory (Erickson & Scott, 1977; Lezak, 1983).
Consequently, although this test is not without diagnostic utility,
evidence that it can measure and differentiate visual memory processes
is equivocal.

Finally, the most widely used measure of visual memory besides-~or
in addition to--the CFT is the Wechsler Memory Scale. The revised form
(Wechsler, 1987) contains two subtests of visual memory. The Visual
Reproductions subtest (VR) involves studying geometric stimuli and then
recalling (i.e., drawing) them from memory. There is no recognition
component for this subtest. A second subtest, Figural Memory (FM), uses
abstract stimuli in a recognition paradigm but without a measure of
recall. Ideally, however, one would like to know, for a given stimulus,
how much is available for recall and how much can then be recognized;
when the two tasks are conducted on separate stimuli, the two processes
cannot be analyzed in relation to one another.

Kaplan (1988) designed an unpublished recognition task to be a
supplement for the Visual Reproductions subtest. That task was intended
to accomplish the same objectives of the present study, except by
modifying the WMS rather than the CFT:

Examining a patient’s visual reproductions in this manner obviously
can be very helpful in determining the impact that such factors as

9
encoding, retrieval, visuomotor ability, and perceptual functioning
have on performance. Thus if a patient fails to adequately
reproduce a design that he or she has just seen but can recognize
it and has no problem copying the design except for a segmented
approach, a retrieval problem secondary to deficient encoding may
be postulated. If a patient’s impaired recall, copy, and delayed
recall are all comparably flawed, but both multiple choice
recognition tests [one following immediate recall and one
following delayed recall] are unimpaired, one may assume that
visuomotor dysfunction is implicated rather than a memory problem
(Kaplan, 1988, p. 146).
The HMS-VB, however, is inadequate, even when the unpublished
recognition items are added. First, the stimuli are too simple.
Consequently, they are likely to be coded verbally (Reed, 1974); it is
not surprising, therefore, that this subtest has consistently loaded on
a factor with two verbal tasks in the battery, Logical Memory and Paired
Associates (Wechsler, 1987), possibly reflecting patients’ tendency to
use verbal labels. The simplicity of the figures may further compromise
the usefulness of these tests by insufficiently taxing the upper range
of visual memory capacities. Palmer (1977) reported that
with more complex figures and/or more demanding tasks, many more
levels [of perceptual organization and processing] may be required.
The requirement of multiple levels of structural units rules out
standard template or iconic forms of representation, since only
one level of functional unit is defined for a pattern. It also
rules out any feature-list representation that has only one level
of structural features (p. 469).
Gazzaniga and LeDoux’s (1978) evaluation of clinical data echoed
Palmer’s conclusions. "The interpretation here would seem to be that
both the left and right half-brains have substantial capacities for
visual recognition, with the right excelling mainly where the upper
perceptual limits are tested" (p. 68). Therefore, with simple stimuli,

processes in the left hemisphere may compensate for, and thereby

disguise, any impairment of the right hemisphere.

10

Other findings confirm these speculations. King (1981) found very

modest relationships between "percent recall" scores (1 recall delay

.22).

recall score / copy score x 100) for the WMS-VR and the CFT (g
This supports the hypothesis that the two instruments measure the domain
differently, most likely because of the difference in stimulus features
(i.e., the complexity of the CFT figure).

A second weakness of the HMS-YR is the small number of items in
Kaplan’s unpublished recognition task; there are only four items for
immediate recall and four for the delay trial. Although this may
suffice as a rough screening, it may be insufficient for diagnosis and
treatment planning.

n e n 0 her Memor Tea 8

Another way to address issues of visual encoding and post-encoding
processes is to draw inferences from tests other than visual memory
measures. This can often be inconclusive, however, particularly if the
measures of retrieval deficits are designed for verbal stimuli. The
reason is that memory for verbal material (processed primarily by the
language-dominant hemisphere of the cerebral cortex, usually the left)
is largely dissociated from nonverbal memory (which appears to be
localized predominantly in the opposite hemisphere, usually the right).
Although some language functions are subserved by the right hemisphere,
testing with commissurotomy patients reveals short-term verbal memory
deficits when information is presented only to the right hemisphere
(Zaidel, 1985). In addition, research with epilepsy patients undergoing
unilateral temporal lobectomies indicates that left temporal damage
selectively impairs verbal memory whereas right-temporal damage impairs

nonverbal memory (Milner, 1971, 1975, 1978; Rausch, 1985; Taylor, 1969).

11
This suggests that strengths or weaknesses in verbal learning and/or
retrieval may be unrelated to nonverbal memory functions.

Furthermore, the study by Moss et a1. (1986), cited earlier,
demonstrated how memory dysfunction can be quite modality-specific,
beyond simple auditory-verbal and visual-spatial distinctions. On the
DRST, Huntington’s disease patients were unimpaired on measures of
verbal material despite memory loss across four types of nonverbal
stimuli; this feature, in fact, distinguished this form of dementia from
Alzheimer’s and alcoholic Korsakoff’s syndrome. Consequently,
assessment across modalities is paramount for sensitive
neurodiagnostics, and nonvisual measures cannot be used to describe

processes of visual memory.

Extended CFT: A Synthesis of Assets

This discussion has demonstrated that any neuropsychological
battery needs a test of visual-spatial memory that: (l) is modality-
specific, (2) employs a complex stimulus, (3) includes not only a
measure of free recall but also a recognition task with enough items to
be discriminating, and (4) includes a matching trial to verify accurate
perception of the memory stimuli. In this project, recognition and
matching trials were developed to supplement the Rey-Osterrieth CFT.
Because the design of these instruments and their validation derive from
a neuropsychological model, a brief description of neuroanatomical

correlates follows.

NEUROANATOMICAL CORRELATES OF

RELEVANT COGNITIVE BEHAVIORS

Neuropsychologists attempt to relate behavior to neurological
substrates. The data come from clinical, developmental, and animal
neurosciences, with ongoing efforts to integrate these areas (e.g.,
Heilman a Valenstein, 1993; Mesulam, 1985b; Olton, Gamzu, a Corkin,
1985; Squire a Butters, 1985). More and more efforts are underway to
compile findings from all domains in which memory phenomena are studied,
not only from the neurosciences but also from cognitive, developmental,
and social psychology (Solomon, Goethals, Kelley, & Stephens, 1989).

The recognition and matching tasks under development are intended
for use with neurologically impaired patients as well as for
understanding memory changes in older adulthood. The rationale for
designing these two tasks derives predominantly from neuropsychological
models of cognition and from research with clinical populations.
Consequently, a review of that research will be included here, even
though hypotheses with neurological patients will not be tested here.

The brain is divided into many anatomical regions based on
topographical features, histological distinctions, and functional
relationships. Some of these regions are specialized; that is, they
exercise primary control over certain behaviors and process certain
kinds of information more efficiently than other regions. The following

review is not intended to be an exhaustive account of functional

12

13

neuroanatomy. Rather, it encompasses those findings that are relevant
to the threefold purpose of this project: developing recognition and
matching trials for the CFT, assessing the role of attention in age-
related memory decline, and examining sex differences on the CFT.

Performance on the CFT involves attention, perception,
construction, and memory (encoding/storage, consolidation, and retrieval
processes of secondary memory). Because language mediation has been
implicated in the memorization of simpler visual stimuli, it too is

described in the model that follows.

Perception

Perception is the interpretation of a sensation, the ascription of
meaning or identity to a sensory stimulus. Each of the senses has a
discrete unimodal locale in the cortex where sensory data from that
modality are first processed consciously (usually after coursing from
the sensory organ through relay nuclei in the thalamus). For each of
the senses, there are unimodal primary and secondary areas and a
unimodal association area on the cortex. Primary and secondary areas
register the physical attributes of the stimulus (e.g., wavelength,
amplitude), and the unimodal association area for that sense attaches
meaning to the pattern of firing (e.g., "blue," "bright"). In addition,
there are multimodal areas where perceptual data from several unimodal
association areas are integrated (Barr & Kiernan, 1988; Mesulam, 1985b).

Visual sensation is represented in V1 and V2 (Brodmann areas 17 and
18, respectively) along the calcarine sulcus of the occipital lobe;
visual perception is achieved by the adjacent visual association cortex

(VA, area 19). Somatosensory information is first registered in SI and

14
82 (areas 1, 2, and 3), just posterior to the Rolandic sulcus in the
anterior parietal lobe; interpretation of those sensory data is achieved
in SA (area 5). Auditory data are first consciously recorded among the
cells of the anterior transverse temporal gyrus (Heschel’s gyrus, A1 and
A2, areas 41 and 42) and interpreted in Hernicke’s area on the plenum
temporale (AA, Area 22). Primary gustatory information projects to
orbitofrontal cortex, just anterior to the insula, and olfaction
projects to the piriform cortex. Both of these senses have association
cortex in the paralimbic regions (Barr & Kiernan, 1988; Mesulam, 1985b).

In addition to the unimodal association areas that correspond with
each of the primary sensory areas in neocortex, there are two prominent
heteromodal association areas where information from several unimodal
association areas is exchanged and integrated with other forms of
processing such as language, memory, and motor movement. One of these
areas is located at the juncture of the temporal and parietal lobes,
primarily comprised of the angular and supramarginal gyri (areas 39 and
40, respectively). Another prominent multimodal area lies in the
prefrontal cortex and encompasses Brodmann areas 8, 9, 10, 11, 12, 45,
46, and 47. This area integrates perceptual, motor, and processing
functions to achieve the higher-order intellectual behaviors
characteristic of humans (Barr a Kiernan, 1988; Mesulaa, 1985b).

As a final organizational principle, visual and somesthetic
sensation and perception are represented contralaterally on the cortex.
Stimuli in the right visual field of each eye project to V1 and V2 in
the left hemisphere, whereas stimuli in the left visual fields project
to the right hemisphere. Similarly, somesthesis from the right side of

the body projects to 81 and 82 in the left hemisphere; left-side

15

afferents cross to the right hemisphere. Audition is less lateralized,
due to substantial bilateral projections and to commissural connections
between the inferior colliculi (Barr & Kiernan, 1988; Mesulam, 1985b).

A lesion anywhere in the circuit between the sense organs and the
primary sensory area can result in unawareness of contralateral
sensation despite appropriate reactivity by the sense organ itself. In
vision, this is sometimes called "cortical blindness." A lesion of the
unimodal association area or of the projections between unimodal and
heteromodal association areas ("disconnection") may result in some forms
of agnosia, an inability to identify the stimuulus or to appreciate its
significance despite an accurate description of the physical properties
of that stimulus (Bauer, 1993). Lesions involving the temporo-parietal
heteromodal association area can result in a variety of disorders of
perception and orientation (Mesulam, 1985b; Damasio & Anderson, 1993).

The integrity of these primary, secondary, and association areas is
critical to effective performance on the CFT. Pillon (1981; in Lezak,
1983) reported that individuals with parietal-occipital pathologies
displayed profound deficits in spatial organization, which were
ameliorated by the provision of reference points to guide their
drawings. There remains a question as to whether such deficits in
perception and organization (manifest not only in the copy trials but
also in the recall trials) hinder encoding of the stimulus in memory.

This issue may be better addressed with the use of a recognition task.

Attention
Two theories of attention based on clinical case studies and

experimental animal research have converged on a coherent, albeit gross,

16
neuropsychological model of attention. A third theory, which derives
from cognitive neuroscience, leads to a rather different model.
Similarities between the models suggest that there are some reliable
principles to the organization of attention; the many differences in
definitions and substrates demonstrate that our conceptualization of
these phenomena is still very primitive.

As described above in the "Perception" section, visual, auditory,
and somatosensory inputs project via thalamic nuclei to the primary
sensory cortices, then to unimodal association cortex, and finally to
multimodal association cortices in the temporoparietal and prefrontal
regions. These multimodal association areas, in turn, have extensive
connections to the inferior parietal lobule and the cingulate gyrus;
Heilman, Watson, and Valenstein (1993), who developed the first model to
be presented here, proposed that the same sensory inputs that stimulate
thalamic and cortical centers simultaneously stimulate the cholinergic
tracts arising from the mesencephalic reticular formation (MRF). These
neurons synapse on the nucleus reticularis in the thalamus to inhibit
its inhibitory influence over the other thalamic nuclei, thereby
facilitating transmission of sensory information through the relay
nuclei within the thalamus. The ascending MRF neurons also project
diffusely to the cingulate and to neocortex, including all of the
primary sensory cortices and the unimodal and heteromodal association
areas. These cholinergic tracts appear to desynchronize electrical
activity in these regions, thereby potentiating transmission of the
sensory signals.

Mesulam (1985a) constructed a very similar model. He called the

MRF-nucleus reticularis-neocortex component of attention the "matrix or

17
state" function, or "tonic attention." The other network, which
involves the sensory relays and specific cortical projections, controls
what he called the "vector or channel" function, or "selective
attention." Tonic attention appears to be maintained primarily by the
MRF with some executive regulation by the frontal lobes; injury to the
MRF produces a disruption of overall attentional tone (e.g., a delirious
state). Selective attention appears to involve an interchange (via
thalamic and striate connections) between the sensory information of the
primary sensory cortex, the motivational information of the cingulate
(e.g., the emotional and adaptive relevance of the stimulus), and the
motor control of the posterior frontal cortex (e.g., orientation to the
stimulus by the frontal eye fields, area 8). Lesions in any of these
areas can produce inattention to part of the perceptual field (Mesulam,
1985a).

A third and quite different model has been proposed by Posner and
his colleagues (Posner & Peterson, 1990; Posner & Presti, 1987). Posner
divides attention into three functions. The first component of
atttention, "alerting" or vigilance, is similar to Mesulam’s tonic
attention. However, Posner proposes that this is a function of a
noradrenergic pathway that arises out of the locus ceruleus at the level
of the midbrain and pens. This pathway courses through frontal areas
and then projects caudally to the posterior attention system that serves
visual orienting; these tracts are more lateralized to the right
cerebral cortex.

Posner’s second component, "orienting," corresponds to Mesulam’s
selective attention, but Mesulam and Posner emphasize different

subsystems. Mesulam subdivides this function into sensory information

18
processing (a function of sensory cortex), motivational processing
(cingulate gyrus), and motor movements (posterior frontal cortex).
Posner defines orienting in terms of three other functions (and
different structures): (1) disengaging from the present focus (a
function of the parietal lobe), (2) shifting attention to a new location
(a function of the midbrain), (3) and registering the contents of the
new location (mediated by the pulvinar nucleus of the thalamus).
Finally, Posner’s third component is "target detection"; research with
both auditory-verbal and visual-spatial stimuli has implicated the
anterior cingulate gyrus as a common substrate for this function (Posner
& Peterson, 1990).

All of these theorists state that visual attention, like visual and
somesthetic sensation and perception, is lateralized. That is, there is
a functional asymmetry between the left (LH) and right (RH) cerebral
hemipheres. However, only the LH follows the contralateral principle
described above; the RH processes stimulation bilaterally. According to
Heilman and Mesulam, the LH controls attention to the right visual
fields; the RH, on the other hand, controls attention to stimulation
arising from eithe; hemispace (Heilman & Van Den Abell, 1979, 1980;
Mesulam, 1985a). Therefore, LH lesions impair attention to the right
hemispace, but this is partially compensated by ipsilateral projections
to the RH. By contrast, RH lesions impair attention to the left
hemispace, with no compensatory ipsilateral projections to the LH.

These hemispheric differences are believed to account for the
clinical finding that left neglect syndromes are both more severe and
more enduring than right neglect syndromes (Cummings, 1985). Past

research has indicated that in addition to being more severe and

19

enduring, left neglect syndromes are also more common than right

 

neglect. However, this claim has been challenged. Ogden (1987) found
that just as many left-brain-damaged (LBD) patients as right-brain-
damaged (RBD) patients experienced contralateral neglect.

Ogden (1987) proposed that the relative incidence rates were
misconstrued for several reasons. First, it appears that patients with
anterior left lesions frequently experience concomitant aphasia;
investigators tended to exclude these subjects from their studies.
Secondly, some researchers focused on large figural units for evaluating
neglect, but they ignored omissions of smaller details (which are more
common in LBD patients). In addition, many tests are more sensitive to
left neglect than to right neglect, perhaps because of the difference in
severity. Finally, because neglect in the left side of visual space
does tend to be more enduring, the rate of left neglect may appear
higher with greater time intervals between injury and assessment.

When these problems were controlled, Ogden (1987) found that on
some measures there were as many patients with right neglect as there
were with left neglect. The former patients typically suffered anterior
LH lesions, whereas the latter typically suffered posterior RH lesions.
Consequently, it can be expected that, if the instrument is sensitive
enough, anterior LBD and posterior RBD will result in equal frequencies
of neglect. In spite of this qualification for the frequency of the two
syndromes, it is still maintained that left neglect is typically more
severe and more enduring than right neglect.

Other aspects of attention are also believed to be lateralized.
Heilman, Watson, and Valenstein (1993) argue that the RH may play more

of a role in arousal than does the LH. As mentioned earlier, Posner and

20
Peterson (1990) concur, pointing to evidence that the noradrenergic
innvervation of the posterior attention systems is itself more
lateralized to the RH. In addition, Posner and his colleagures posit
that the RH may be more specialized for attending to global features of
a stimulus (e.g., gestalt, contours, or melodies) and that the LH may be
more specialized for attending to local, or detail, elements (e.g.,
individual line segments or musical notes).

These three models differ in how they subdivide attention; they
also differ in some of the substrates they implicate. Nonetheless, some
common principles emerge. First, attention involves a general arousal -
function that maintains attentional tone across the cortex and "alerts"
or potentiates specific regions for special processing of incoming
nervous signals. This function appears to have its center in the
mesencephalon and is achieved either via ascending noradrenergic tracts
or via cholinergic innervation. Second, at the cortical level there are
several component processes involved in directing attention to a
specific input for priority processing. However these functions are
defined, they are sure to involve the posterior right hemisphere,
especially for processing bilateral visual stimuli and gestalt features,
and the left hemisphere, especially for processing right-side visual
stimuli and detail features. Third, there appears to be a motivational
component achieved by the cingulate gyrus and its connections to limbic
structures. Finally, there is a motor component to selective attention
that is served by motor and premotor cortex in the posterior frontal
lobes.

From these models of attentional representation, it could be

predicted on the CFT that RBD would more likely result in neglect of the

21
left side of the constructions, and LBD would be manifest to a lesser
degree in the right side, if at all. In fact, this has been observed in
CFT copy and recall drawings (Kaplan, 1988; Milberg, Hebben, & Kaplan,
1986). Patients with RBD neglect the left side of the figure in their
copy and recall drawings, whereas left-lesioned patients frequently show
a progressive loss of detail (not a total neglect) in the right side of
their drawings. Among stroke patients, Binder (1982) discovered that
left cerebrovascular accidents (CVAs) resulted in right-side hemi-
inattention. Patients with right CVAs neglected the left side of the

figure and produced more severe spatial distortions.

Construction

Construction, or constructional praxis, is the ability to assemble
materials into a whole object. This term is used to refer to a
diversity of tasks, ranging from two-dimensional pencil drawings to
jigsaw puzzles to three-dimensional block configurations. Assessment of
the construction function itself assumes the integrity of component
functions such as perception, attention, and motor control.
Construction ability is the actual synthesis process. Benton and Tranel
(1993) argue that this concept has been applied too broadly to be
clinically useful. They demonstrate that a plethora of studies localize
"construction" to many different parts of the brain, depending on the
response required (especially two-dimensional versus three-dimensional),
how the stimulus was presented (two-dimensional versus three-dimensional
model), and the difficulty of the task (especially the complexity of the
stimulus). Studies of laterality are difficult to compare because they

use different tasks and they do not control for location or severity of

22
injury. Benton and Tranel’s (1993) review and Mesulam’s (1985b) review
suggest a convergence of localization findings: There is substantial
agreement that posterior lesions are more devastating than anterior
lesions, and a preponderance of the evidence implicates the right
parietal region in more severe cases of constructional apraxia.

Investigators using the CFT with frontal lobe patients have
observed that planning and organizational processes controlled by the
frontal lobe also play a role in construction for this two-dimensional
task. Bennett-Levy (1984) reported that, during the CFT copy and recall
administrations, frontal patients exhibited a loss of programming and
lack of organization yielding poor recall scores. Klicpera (1983) used
the CFT with dyslexic children and concluded that visuomotor and
visuospatial anomalies in this population appear to be less of a
function of lateralization than of the overall integrity of the frontal
lobes. Compared with normal controls, his sample of learning disabled
children produced the copy and recall trials with very poor organization
and a less systematic approach. Therefore, problems on the constructive
trials of the OFT may have reflected poor organization of both the
stimulus and the response.

There is little consensus on how construction should be
subclassified and what that might mean for localization. This
discussion has shown that constructional disability has been correlated
with both left and right pathology and with both anterior and posterior
lesions. Regardless of the cerebral substrate, it is reasonable to
assume that how the CFT copy is organized and synthesized into a picture
is likely to have an impact on the encoding/storage and retrieval of the

memory trace. Therefore, it is possible that the extent to which these

23
deficits actually impair encoding of the memory trace can be examined

with the use of a recognition task.

Language

Language is significantly lateralized. In virtually all right-
handed individuals and in the majority of left-handers (about 70%), the
left hemisphere exercises substantially more control over language
functions and is therefore called the "language-dominant" hemisphere.
People who sustain brain damage early in life are much less homogeneous
with respect to lateralization of language functions; in this group, up
to 30% of right-handers and 801 of left-handers are right-dominant or
bilaterally represented for language (Springer & Deutsch, 1985).

It should be noted that dominance does not mean exclusive control.
Although most major language functions are controlled by the language-
dominant hemisphere, language processing is represented to some degree
in been hemispheres (Benson, 1985; Zaidel, 1985). This point has been
argued in a banter of polemics revolving around the split-brain findings
of the ’608 and ’70s (Gazzaniga, 1983a, 1983b, 1984; Levy, 1983; Myers,
1984; Zaidel, 1983). Challenges to the conclusions of those early
studies have hinged on issues such as: the extent to which the measures
that were used effectively circumvented intact communication circuits
between hemispheres (e.g., bilateral audition); bias in subject
selection, nalely, choosing to study those commissurotomy patients who
gig show right-hemisphere language because they exhibited that
presumably uncommon characteristic; the extent of surgical sectioning
and, more specifically, the role of the often-spared anterior

commissure; and definitions of "linguistic competence." The consensus

24
of this sample of critical scholarship appears to be that (1) any
linguistic skills that exist in the language-nondominant hemisphere are
rudimentary at best and that (2) whatever skills that can be described
reliably and validly in these early patients can be generalized only
with extreme caution because of the limited number of cases and because
of their idiopathic histories.

A recent review of lexical processes was conducted to further
elucidate the role of each hemisphere in language. To avoid the
interpretive pitfalls of research employing split-brain and other
neurological samples, Chiarello (1988) focused exclusively on normal,
right-handed adults who were neither bilingual nor reading-disabled.
The strategy most frequently employed with the normal population is the
manipulation of word presentations using a visual half-field technique.
Based on her thorough examination of this research base, she concluded
that lateralization is manifested in different ways and to varying
degrees at each of three stages of lexical processing. The most
substantial differences exist in prelexical operations (encoding). At
this stage, if the letters are presented in the normal reading format,
the left hemisphere perceives the stimuli (i.e., words) in a very rapid,
automatic process whereas the right hemisphere employs a slow, serial
analysis; however, the relative efficiency of the right hemisphere
increases when the words are presented in an unusual format (e.g., in a
vertical array). In actual lexical access, the second stage of this
model, lateralization effects were not observed; there is no evidence
that the right hemisphere is limited in the size of "its" word stores,
nor is the right hemisphere less efficacious in its access (relative to

left-hemisphere processing). Finally, postlexical processing, like

25
prelexical processing, contains some cortical asymmetries; however,
these lateralized differences interact with the demands of the task and
with the type of lexical information involved and cannot be easily
distilled into interpretable generalizations.
In conclusion, for the majority of the population the LH excels in

processing most verbal stimuli. Although the RH exercises some control

 

over some language functions, the LH appears to be more skilled and is

 

used preferentially, especially for verbal stimuli presented in standard
language form (e.g., left-to-right arrays for English-speaking people).
With regard to memory-performance, therefore, verbal codes for visual
stimuli could implicate LH functioning instead of the RH patterns

described below.

Memory

The investigations cited thus far support a model of complex
interaction between the hemispheres in the processing of language and
spatial information. Nonetheless, when we examine our present body of
eeeeey data, the generalizations of left language dominance and right
spatial dominance seem to prevail (Delaney, Rosen, Mattson, & Novelly,
1980; Kaplan, 1988; Kimura, 1963; Milberg, Hebben, & Kaplan, 1986;
Milner, 1971, 1978; Rausch, 1985; Squire, 1982, 1986; Taylor, 1969;
Zaidel, 1977, 1985).

Generally stated, patients with left brain dysfunction (LBD)
experience impairments of verbal memory even when nonverbal memory is
completely spared; individuals who suffer right brain dysfunction (RBD)
display more deficits of figural memory, that is, when the visual

material is complex enough to preclude verbal mediation. Although King

26
(1981) found no differences on the CFT when comparing patients with left
hemisphere lesions and those with right hemisphere lesions (both groups
equally impaired relative to noninjured controls), other evidence
suggests that this task is very sensitive to hemispheric differences,
both quantitatively (in terms of the alount recalled) and qualitatively
(how the figure was drawn during recall).

Milner (1971, 1978) found that, across a variety of tasks, patients
with right temporal lobe epilepsy who subsequently had partial
lobectomies experienced impaired visuospatial memory compared to their
counterparts with partial lobectomies of the left temporal lobe. Taylor
(1969) demonstrated this with the CFT among patients who had left- and
right-temporal excisions.

Kaplan (1988; Milberg, Hebben, & Kaplan, 1986) uses the CFT as part
of her "process approach" to assessment. As would be predicted from the
above-described behavioral geography of the brain, LBD patients
experienced less impairment on this spatial memory task than RBD
patients. The drawings by the two patient groups also differed
qualitatively. The RBD patients’ figures typically lacked proper
contours, integration, and organization, whereas the LBD patient’s
figures were typically drawn in segmented fashion and were more likely
to lack internal detail.

Certain subcortical regions have also been implicated in memory
processes, although it is difficult to be specific because a lesion of
one of these areas is typically accompanied by damage to its neighboring
structures (Barr & Kiernan, 1988). Nonetheless, some patterns have
emerged in clinical neuropathology and in animal science to elucidate

the relative contributions of some of these deep nuclei.

27

Squire (1982, 1986) proposed that there are two types of
subcortical amnesias. Patients with "diencephalic amnesias" (those
involving the mammillary bodies and dorsomedial thalamic nuclei) exhibit
a normal rate of forgetting, which suggests that theirs is an encoding
deficit; retention is relatively unimpaired, however, such that
information that can be acquired can also be retained. Patients with
"bitemporal amnesias" (hippocampal and amygdaloid lesions), by contrast,
exhibit an abnormally rapid rate of forgetting, which implicates
deficient post-encoding processes of consolidation and/or elaboration,
which are critical for effective retrieval. These findings are limited -
to declarative knowledge (information); procedural knowledge, or
perceptual-motor ability, is not affected by these local pathologies.

A review of clinical human studies and experimental animal research
(Winocur, 1984) supports the conclusions that thalamic damage affects
encoding whereas hippocampal injury affects consolidation and subsequent
retrieval. Furthermore, hippocampal lesions may contribute to

interference effects observed in memory processing.

Summary

All of these findings help describe the organization of salient
cognitive functions from a neuropsychological frame of reference.
Throughout this discussion I demonstrated the importance of assessing
cognitive processes through at least two modalities, auditory-verbal and
visual-spatial. This discussion also revealed the diagnostic
implications for differentiating among attention, perception,
construction, and memory deficits. Finally, I argued that it can be

useful to distinguish between encoding and post-encoding processes in

28
memory functioning. Adding recognition and matching trials to the Rey-
Osterrieth Complex Figure Test can be an invaluable asset to extant
neuropsychological batteries. The additional trials will enable
neuropsychologists to rule out perceptual and constructional confounds
to the memory assessment, and comparisons between free recall and
recognition performance can help distinguish between encoding and post-
encoding deficits.

To make future clinical investigations and applications meaningful,
this project will lay the psychometric groundwork for the new
recognition and matching trials and provide age-appropriate norms.
Furthermore, this project will explore the interplay of sex and age on
the different trials of the CFT (i.e., Copy, Immediate Recall, Delay
Recall, Recognition, and Matching), in hopes of shedding light on
developmental differences among the various memory processes for males
and females. In the sections that follow, I review the literature on
memory differences that have been observed between different age groups
and between males and females, with emphasis on visual memory and its
measurement. The specific predictions of this project are embedded in

those discussions.

AGE DIFFERENCES IN VISUAL-SPATIAL MEMORY

AND IN RELATED VARIABLES

Developmental changes in memory are receiving more and more
attention in the research literature for several reasons. First,
complaints of memory difficulties are common among older adults,
beginning in the 508 and becoming especially frequent after age 60
(Albert, 1984). (In this discussion, "older adults" refers to people in
their late 508 and older unless otherwise indicated.) This has led many
investigators to examine the validity of subjective reports and to
determine the nature and extent of reported age-related memory
impairments. Second, pattern of memory loss in older adulthood has
proven to be diagnostically useful, for example, differentiating
dementia (organic memory loss) from pseudodementia (temporary memory
loss secondary to depression). Consequently, there has been an
increasing need for sensitive measures of memory processes and for age-
appropriate norms, especially as more and more people thrive beyond the
seventh decade.

In this section, I present a cognitive model of memory that has
been frequently applied to explain memory phenomena. The discussion
begins with four theoretical memory stores (sensory memory, primary
memory, secondary memory, and tertiary memory) and the putative
information processes that act upon the memory traces within and between

those stores (selective attention, encoding/storage, consolidation, and

29

 

 

 

30

retrieval). Then I address variables that affect memory performance:
attention, perception, production, health, depression, and anxiety.
Research findings on age differences are documented throughout the
discussion. Finally, I relate these findings to the study at hand, the
development of recognition and matching trials for the CFT.
Memory Stores

Since the early information-processing theories in the 1960s,
models of memory have incorporated putative storage cells of varying
capacities and functions. The smallest store holds sensory stimuli in
close-to-original form for a very brief time; it is frequently labeled
"sensory memory" or the "sensory register." It is engaged
unconsciously, has a fixed storage capacity, and is further
characterized by rapid decay of the stimuli. The memory traces are
called "echoes" when they derive from auditory stimuli and "icons" when
the stimuli are visual-spatial. These traces are maintained for only a
few seconds at most, until the individual can selectively attend to a
subset of stimuli that require further processing. These processes
proceed more-or-less automatically and at an unconscious level.

The next store is called "working memory, short-term memory," or
"primary memory." Like the sensory register, primary memory has a
relatively fixed capacity; in contrast to the sensory register, primary
memory involves conscious processing, and traces can be maintained as
long as they are active in the store. It is postulated that primary
memory serves as a workbench where new memories are maintained until
they are no longer needed; the traces may then be displaced by new

incoming material or encoded for longer storage and for later retrieval.

Primary memory can hold approximately seven "chunks" of information,

31

although each chunk can represent many smaller units of data; for
example, the numbers 1, 4, 9, 16, 25, 36, and 49 could be rehearsed as
one chunk, "perfect squares." Traces can be held in this store as long
as the individual consciously maintains them, such as through rehearsal.

More permanent storage conventionally has been called "long-term
memory (LTM)." This storage is believed to have infinite capacity and
can maintain at least some memories for an indefinite period of time.
Memory traces are encoded here and are stored, becoming more and more
consolidated (that is, resistant to forgetting) with time. Retrieval
processes make these memories accessible to primary memory for future
re-use. Encoding and retrieval both demand substantial attentional
resources, whereas consolidation appears to proceed automatically.
Forgetting from LTM appears to be due, at least in part, to improper
encoding and/or inefficient retrieval; whether LTM traces are subject to
decay is still a subject of debate.

Since the original models were proposed, LTM has been subdivided.
A classification scheme has been introduced that distinguishes between
long-term memories that are more contemporary ("recent memory" or
"secondary memory") and memories that comprise one’s personal archives
from the distant past ("remote memory" or "tertiary memory"). An
example of secondary memory would be a newly learned list of words or
story; tertiary memory, on the other hand, would house autobiographical
information (name, birthdate, etc.) and overlearned material (e.g., the
alphabet or nursery rhymes). Whether these two "types" of memory
actually represent different stores or whether they are simply anchors

on either end of a consolidation gradient has yet to be determined.

32
Nonetheless, inasmuch as they are affected differently by pathological
processes and by aging, the distinction would appear to be warranted.

In a recent review of age differences in memory, Poon (1985)
summarized findings that had addressed the relative capacities of each
of these stores across the lifespan. He concluded that age-related
impairments in sensory, primary, or tertiary stores are negligible or
absent altogether. However, in secondary memory, older adults performed
significantly worse than younger adults. Differences in this domain
implicate deficiencies in encoding and/or retrieval processes. These
processes have been investigated at length, especially using auditory-
verbal stimuli.

Meeery Processes

Cogsoligation. Consolidation is the unconscious, automatic process
that transforms secondary memory traces into tertiary memories by making
the traces more resistant to disruption. Because there are no age
differences in tertiary memory (Poon, 1985) and because consolidation is
largely automated thereby requiring fewer attentional resources, it
seems probable that consolidation is not affected by the aging process.

This hypothesis is reflected in the following prediction:

 

Prediction 1: When the variance associated with consolidation in
secondary memory is isolated, age will not explain a

significant portion of that variance.

 

Encoding and retrieval. The bulk of research has attempted to

compare encoding and retrieval efficiencies across the age span. The

33
findings of many studies suggest that attentional resources are central
to understanding age differences.

Smith (1977), manipulated cues in an auditory-verbal learning task
in an attempt to investigate the effectiveness of different cues to
ameliorate age differences in memory performance. When appropriate cues
were provided at input, thus facilitating encoding, older adults
performed as well as younger adults did. By contrast, providing cues at
retrieval benefitted all age groups equally. These results suggest that
memory decline in older adults is due to faulty or incomplete encoding.

In a later paper, Smith (1980) reviewed the literature and found
other evidence for encoding deficits. Older adults were impaired across
three different modes of processing (organization, elaboration, and
imagery); however, they differed from younger adults not in their
ability to process the stimuli but instead in their spentaneoes use of
such strategies. This was also the finding of Perlmutter and Mitchell
(1982), based on their own data and based on a review of 12 other
studies. Unfortunately, the present research design does not lend
itself to the isolation of encoding functions; consequently, a

prediction will be made about only retrieval mechanisms:

 

Prediction 2: Age will explain a significant portion of variance in

retrieval operations of secondary memory.

 

Craik and his colleagues (Craik & Byrd, 1982; Rabinowitz, Craik, &
Ackerman, 1982) explored these encoding deficits further. They found
that older adults tend to encode core semantic (general) features rather

than distinct characteristics of stimuli. The former is believed to

34
involve more automated processing, whereas the latter is assumed to
require more effortful processing. Therefore, the investigators
concluded that memory differences are due to the reduction in processing
resources with increasing age. This is consonant with other findings
indicating that automaticity is spared in aging and that effortful
processing declines (Hasher & Zacks, 1979).

Waugh and Barr (1982) also found evidence for an attentional
deficit hypothesis. Specifically, older adults registered less
information than younger adults and failed to encode specific details
valuable for later retrieval. Thus, it appeared that older adults
encode more selectively and less efficiently, attending to some items at
the expense of others.

Although most of these studies employed auditory-verbal material,
similar investigations used visual-spatial stimuli. Craik’s (1977)
review of the literature indicated that visual-spatial memory is
susceptible to the same processing deficits as auditory-verbal memory
and may be even more compromised than memory for auditory-verbal
material. One reason may be that auditory-verbal input can be processed
with less effort and more automaticity. "When the input is visual,
however, attention is required; older subjects have less processing
capacity to spare in the divided attention situation, thus their
“performance on visual memory tasks is especially poor" (Craik, 1977, p.
392). Findings by Winograd and Simon (1980) support this conclusion.
They found that older adults did worse in forming and/or retrieving
pictorial codes as opposed to auditory-verbal encoding/retrieval. Craik
(1977) also reported that comparisons between recall and recognition

with visual-spatial stimuli (geometric designs) are similar to such

35
comparisons with auditory-verbal stimuli. Specifically, recognition,
although impaired with aging, is more resistant to decline than is
recall. This suggests that auditory-verbal and visual-spatial retrieval
are similarly facilitated by the decreased processing demands of a
recognition format.

Some investigators have used the Complex Figure Test with older
adults. Tombaugh and his colleagues (Tombaugh, Hubley, Faulkner, &
Schmidt, 1990) administered the test in a modified fashion; they
instructed subjects to memorize the figure during a timed exposure
(intentional learning, viewing the stimulus rather than drawing it) and
they administered the copy trial last, following the immediate and delay
recall trials. They recruited a large community-dwelling sample that
had equal numbers of subjects in each 10-year age band from 20-79.
These investigators found that adults in their 608 scored lower than
younger adults on immediate and delay trials, despite equivalent
performance on the copy trial; adults in their 70s scored even lower
than adults in their 608 on those same trials. These findings were
corroborated by Berry, Allen, and Schmitt (1991), who used the standard
administration of the test with healthy, community-dwelling adults
ranging from age 50 to age 79. This is further evidence that the
secondary memory deficits observed with auditory-verbal material affect
visual-spatial material in a similar way.

Read (1987) suspected that the original stimulus was too difficult
for the older adults to draw and that difficulty on the copy trial
resulted in poor recall. (No data were reported to support this
proposition.) Consequently, he presented individual elements of the

design in a structured, sequential manner so that the subjects could

36
copy the figure effectively. Cues were provided at the time of
retrieval in a similar fashion. Copy reproductions and recall improved
substantially. Nonetheless, even with the modified administration and
equivalent copy performance, older adults performed worse on recall than
did their younger counterparts. Facilitating encoding and retrieval,
and thus reducing the attentional resources required, failed to
eliminate all age differences, suggesting that age-related changes in
attention capacity play a significant but not solitary role in age-
related memory decline. Again, because the present research design does
not permit the dissection of encoding processes, I proffer a prediction

about only retrieval mechanisms:

 

Prediction 3: After controlling for the effects of attention, age will
explain significantly less variance in retrieVal

operations of secondary memory.

 

In summary, sensory memory and tertiary memory operate mostly at an
unconscious level by very automated processes (selective attention and
consolidation); these are unaffected in healthy aging. Primary memory
is also virtually unaffected in healthy aging. By contrast, secondary
memory ability declines significantly with healthy aging. The two main
information manipulations in this store, encoding and retrieval, are
engaged with substantial effort. For older adults, an age-related
decline in attention capacities has been consistently observed,
especially in tasks where attention is divided between two or more
activities. This pattern of attentional decline closely coincides with

the pattern of decline in secondary memory. It seems plausible,

37
therefore, that impairment of encoding and retrieval in secondary memory
is due, at least in part, to limited attentional resources and that
attentional decline is the primary mediating agent of memory decline
among older adults. Because visual-spatial processing is more effortful
than auditory-verbal processing, the attentional deficit may
disadvantage visual—spatial memory even more than auditory-verbal
memory.
Variables that Affect Memory Performance

Memory performance is dependent on other cognitive abilities:
attention, perception, and production. Individuals must attend to the
specific stimulus in their environment and then accurately perceive it
in order to encode and store it as a memory trace. They also must be
able to verbally or graphically report the memory trace upon retrieval.
Because these skills decline with age, they must be considered as
potential confounds to analyses of aging memory. In addition, many
health conditions are known to impair memory performance, and some of
these occur with greater frequency among older adults such that
pathological memory decline might be misattributed to normal, healthy
aging. Finally, clinical levels of depression and anxiety have been
shown to impair performance on memory tests; therefore, the relationship
between age and these emotional states is also salient to this
investigation.

Attegtien. Attention for effortful processing was shown to decline
with age in a classic study by Hasher and lacks (1979). More recently,
Stankov (1988) examined 11 different tests of attention as part of an
extensive battery of other cognitive measures. These 11 measures formed

three attentional factors; performance on all factors declined with age

38
(£8 ranged from -.43 to -.48). Furthermore, these changes in attention
explained most of the variance in the fluid intelligence factor and in
the crystal intelligence factor in that study.

Craik and Byrd (1982) reviewed several lines of evidence for
compromised attention in older adults, including studies of divided-
attention, depth of processing, and automaticity. They concluded: "We
postulate that reduced attentional resources lead to an attenuation or
shrinkage in the richness, extensiveness, and depth of processing
operations at both encoding and retrieval" (p. 208).

Most of the research evidence for age decline comes from cross-
sectional studies, which means that the differences may represent cohort
and time-of-measurement effects in addition to or possibly even instead
of aging per 8e. Crossley, Hiscock, and Beckie (1991) conducted a
longitudinal follow-up to a cross-sectional analysis and found age-
related decline in multi-task processing across the adult lifespan. The
longitudinal findings, in conjunction with cross-sectional evidence,
indicate that these differences are indeed a product of aging,
independent of cohort and time-of—measurement effects.

Pereeptieg egd proguction. Perception and production also decline
with age; the age effect increases with the complexity of the stimuli.
In the auditory-verbal domain, perception (language comprehension for
sentences) declines mildly with aging. Production also changes, with
significant declines in the number of words generated during fluency (or
listing) tasks and with significant increases in repetition, redundancy,
and personalization during discourse (La Rue, 1992).

In the visuospatial domain, simple perceptual skill (e.g.,

judgments of line angles) begins to decline in the fourth decade, but

39
even in the ninth decade performance is only marginally lower than that
in the sixth decade. The decline in the performance of complex tasks
(facial recognition and embedded figures) is more substantial. Similar
declines occur on tests of visual-spatial production. For example,
older adults can accurately copy simple two-dimensional designs but they
experience more difficulty with complex two-dimensional tasks and with
three-dimensional stimuli (La Rue, 1992, chapter 3). Although the CFT
is a relatively complex stimulus, testing of adults ages 20 to 79
(Tombaugh, Hubley, Faulkner, & Schmidt, 1990) and of adults ages 50 to
79 (Berry, Allen, and-Schmitt, 1991) failed to identify any age
differences on the copy trial. Therefore, in this study I expect the

following outcomes:

 

Prediction 4a: Attentional skill and perceptual accuracy will correlate
negatively with age in both auditory-verbal and visual-
spatial modalities.

Prediction 4b: Production ability will correlate negatively with age in
the auditory-verbal modality but will not correlate with

age in the visual-spatial modality.

 

Health factors. Memory also depends on the physical integrity of
the central nervous system. Age-related changes in brain structure,
metabolism, and neurotransmission have been studied extensively, and
correlations between biological indices and cognitive performance in
older adulthood have also been documented (for review, see La Rue, 1992,
chapter 2). Because no biological measures of brain structure and

function were used in this study, this body of research will not be

40
presented here. However, many medical conditions can compromise nervous
system efficiency and impair cognitive performance; in the current
study, these conditions were grossly measured through a structured
interview.

A variety of direct insults to the brain can affect cognition. In
the current study, questions were asked about 1088 of consciousness
(LOC) following traumatic head injury, cerebrovascular disease (CVD;
ischemia and cerebrovascular accidents [CVAs or "stroke8"]),
hydrocephalus, and epilepsy. The structured interview also inquired
about a variety of systemic factors that are known to impair cognition.
These include heavy alcohol use, smoking, hypertension (HTN), coronary
heart disease (CHD), diabetes mellitus (DM), pulmonary insufficiency,
renal insufficiency, and hepatic dysfunction. Several of these same
variables are more likely to show up in older adults’ histories for
three reasons: (1) Older adults have lived longer and thus have longer
histories to report. (2) Physical decline and vulnerability to disease
increase from young adulthood to older adulthood. Finally, (3) older
cohorts did not have the benefit of contemporary health care, preventive
education, and nutrition when they were younger.

Vascular disorders are more common among older adults than among
younger adults and have been shown to impair some elements of cognition.
HTN has been diagnosed in 60 percent of non-Hispanic white Americans who
are over 60 years old (National High Blood Pressure Education Program,
1992). HTN alone has been shown to affect attention and memory and
possibly perception and constructional ability (Waldstein, Manuck, Ryan,
& Muldoon, 1991). In addition, HTN substantially increases risk for

CHD, CVD, and renal disease (National High Blood Pressure Education

41
Program, 1992), which also increase with age and which also have
detrimental effects on cognition.

The prevalence of ischemic heart disease increases with age, rising
to 50 percent among people over age 60 (Bienenfeld, 1990a).
Neuropsychological impairment has been unequivocally demonstrated in
patients with CHD, at least for those with disorders severe enough to
warrant surgical intervention (Bornstein & Kelley, 1991).

CVA is more frequent in older adults, occuring in 5.1 percent of
Americans over the age of 65, compared to 1.0 percent between the ages
of 46 and 65 and only 0.7 percent below age 45. In 47 percent of these-
CVA patients, intellectual impairment is observed; in 50 percent of
theee patients, cognitive decrements are severe enough to fulfill
criteria for dementia (Cummings & Mahler, 1991).

In addition to HTN, CHD and CVD, non-insulin dependent diabetes
mellitus (DM) increases in frequency with age as glucose uptake by
target organs diminishes (Bennett, 1990; Bienenfeld, 1990a). Prevalence
begins to increase around age 40 and reaches 15 to 20 percent after age
65 (Davidson, 1991). DM predisposes individuals to much higher rates of
HTN, CHD and CVD as well as higher rates of acute fluctuations in mental
status secondary to metabolic instability. Cognitive consequences of
the vascular complications (HTN, CHD, CVD) were described above.
Cognitive sequelae due to DM itself have not been conclusively
demonstrated; research in this area has been virtually nonexistent until
recently, and studies so far have been limited to people under age 45
(Bornstein & Kelly, 1991).

In contrast to these other risk factors, use of alcohol and illicit

substances (both frequency and amount) decreases with age, especially

42
after age 70 (Bienenfeld, 1990b). The proportion of current drinkers in
1990 declined across cohorts, from 68.8 percent in 25-44 year-olds to
57.6 percent in 45-64 year-olds to 41.4 percent in adults 65-and-older
(U. S. Department of Health and Human Services, 1993). In addition,
alcohol use is not so predictive of cognitive impairment a8 is long-
standing alcohol abuse. Chronic substance abusers are underrepresented
in older community-dwelling adults because of early mortality and high
comorbidity of conditions leading to institutionalization (Bienenfeld,
1990b; National Institute on Alcohol Abuse and Alcoholism, 1990).

Cigarette smoking is also observed least among older Americans (age
65 and older; 12—15 percent) compared to other adult cohorts (25-34, 35-
44, and 45-64 year-old cohorts range 25-33 percent; U. S. Department of
Health and Human Services, 1993). Related to cigarette smoking is
hypoxemia (pulmonary insufficiency). Even though current smoking may
decline with age, residual pulmonary disease may continue to impair
cerebral oxygenation. In a review of two major investigations of
chronic obstructive pulmonary disease, Bornstein and Kelley (1991)
concluded that severity of cognitive impairment is related to the
severity of hypoxemia. Chronic obstructive pulmonary disease is most
frequently identified in the seventh decade.

Epilepsy can also contribute to cognitive impairment. It is
estimated that 10 percent of the population has a seizure at some time
in their lives. The age of onset for seizures is linked to the
etiology. Among older adults, it is not uncommon for CVA patients to
develop seizures; CVA is the most common cause after age 50. However,
prevalence of seizures is much less common in older adults than among

younger adults and children. Approximately 50 percent of persons who

43
experience epilepsy have their first seizure before age 20; most people
have one or a few seizures without recurrence (Epilepsy Foundation of

America, 1992; National Institutes on Health Consensus Development

Conference, 1990).

 

Prediction 5a: The frequency of HTN, CHD, CVD, DM, and pulmonary
disease will correlate positively with age.

Prediction 5b: Current alcohol and tobacco use will correlate
negatively with age.

Prediction 5c: Nonsignificant correlations are expected between age and
hydrocephalus, renal disease, hepatic disease, seizures,

and LOC that follows traumatic head injury.

 

Emotional influences. Finally, an inventory of depression and an
inventory of anxiety were included in the battery. Clinical levels of
depression can compromise memory performance (Fisher, Sweet, & Pfaelzer-
Smith, 1986; Fromm & Schopflocher, 1984); this effect appears to be due
to a cerebral mechanism rather than motivation deficits (Richards &
Ruff, 1989). However, depression is not expected to confound age
comparisons in this study because clinical depression is no more
prevalent among noninstitutionalized older adults than among younger
adults (Nakra & Grossberg, 1990).

Anxiety, also, can adversely affect cognitive test performance
(King, Hannay, Masek, & Burns, 1978; Wrightsman, 1962). Clinical levels
of anxiety are no more frequent among older adults than among younger
adults (Ruskin, 1990), and adults ages 50-69 have even been shown to

have slightly less reactive ("state") and chronic ("trait") anxiety than

44
adults ages 19-39 (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983,
p. 5). In testing situations, however, older adults are more prone to
reactive anxiety due to repeated confrontation with failure on tasks
that they could perform better at a younger age. Also, older cohorts
tend to be more intimidated by testing because they have less formal

education than younger cohorts (Albert, 1988, pp. 62-63).

 

Prediction 6a: Depression scores and trait anxiety scores will not
correlate with age in this community-dwelling sample.

Prediction 6b: State anxiety scores will correlate positively with age.

 

SEX DIFFERENCES AND VISUAL-SPATIAL MEMORY

Harris (1978, 1981, 1985) documented the history and findings of
sex differences across cognitive domains. Although many of the data are
inconsistent and controversial, some differences have been more robust:
Males tend to excel in spatial tasks, and females tend to excel in
verbal fluency tasks (Harris, 1978, 1981, 1985). Even here, there is
substantial variation within each group and considerable overlap between
groups, indicating that the differences reflect "sex-related
variability" rather than "sexual dimorphisms" (Harris, 1985, p. 296).
Harris (1978, 1981) concluded that this sex-related variability may have
several foundations, both physiological and environmental in nature.

As a clue to a possible physiological mechanism, there is some
evidence for sex differences in the strength of lateralization of
function. Although sex differences often do not appear, where they do
occur it is such that males are more lateralized than females.
Specifically, males show more right-hemisphere (RH) specialization for
spatial tasks and more left-hemisphere (LH) specialization for
linguistic tasks, whereas females exhibit more bilateral representation
of these same skills. Witelson (1976) has reported differences of this
sort in childhood, already at age six. Others have found that the
differences endure into adulthood (e.g., McGlone, 1978). However, the
lateralization data are inconsistent and controversial (see commentaries

in McGlone, 1978).

45

46

There is some evidence for environmental influences. Hyde (1990)
conducted two meta-analyses on research that examined sex differences in
verbal and visual-spatial functioning. Comparing results from the first
meta-analysis, which used studies published before 1973, to results from
the second meta-analysis, which used studies published after 1973, she
found a decline in the magnitude of sex differences in the later study.
She suggested as one explanation that this trend may reflect changes in
sex roles.

It is probable that both physiological factors and environmental
factors contribute to sex differences in the facility and manner of
spatial processing. As one possible scenario, Harris (1978, 1981)
speculated that because females acquire language earlier than males,
they may come to rely on verbal mediation to a greater extent than males
do and, thus, may develop an approach to memory and problem~solving that
draws more heavily on verbal mediation. If so, a preference for verbal
strategies, linked to maturation and molded by environmental forces, may
account for many observed individual and sex differences.

Mixed results have emerged from sex comparisons on the Complex
Figure Test, but some findings suggest that intricate processing
differences may play a role. Waber and Holmes-Bernstein have
investigated sex differences in CFT performance among children. With
regard to copy scores, Weber and Holmes (1985) found no sex differences
on accuracy or organizational approach in a large sample (ages 5 to 14).
Memory performance was examined in a much smaller sample of 5th- and
8th-grade children (Waber, Bernstein, & Merola, 1989). In that study,
the encoding condition (i.e., whether the child simply viewed the figure

or copied the figure during exposure to the stimulus) affected some

47
features of recall, but this effect depended on the child’s sex and age.
In general, older children, especially older boys, benefited from
copying the stimulus as opposed to simply viewing it. These findings
require replication because of the limited sample size (approximately 10
per cell). Nonetheless, they raise several questions. First, are
measures of memory that require visual encoding (e.g., the WMS-R,
Wechsler, 1987) comparable to those whose administration involves motor
encoding (e.g., the OFT)? Second, how do age and sex interact
throughout the lifespan on measures of visual memory?

With factory line workers ranging in age from 17 to 49 (M = 29),
Bennett-Levy (1984) found significant differences between men and women.
Men exceeded women on measures of symmetry, good continuation, strategy
(an organizational score derived from both symmetry and continuation),
and recall. However, subsequent multiple regression analyses showed
that sex did not make an independent contribution to the total variance
in recall; rather, strategy, copy score, and age were the best
predictors.

A thorough examination of the data presented by Bennett-Levy (1984)
suggests that the relationship between strategy use and sex was very
salient to the interpretation of these results. Given that sex did not
make an independent contribution to recall, its effect must be mediated
by another variable. Three variables predicted recall: the copy score,
age, and the strategy score. The copy score and age were found to be
sex-independent, whereas the strategy score (reflective of good
organization) was significantly related to sex (; = -.28).

Consequently, it seems that differences in recall were best

explained not by sex per se but by strategy differences. That is, sex

48
affected the level of organization such that many (but not all) men
organized their copies better than many (but not all) women. Those
subjects who organized their drawings better (a mixed-sex group composed
of more men than women) went on to recall more components of the figure
on later trials (3 = .65 or .69, depending on the stringency of the
scoring rules applied). This re-evaluation supports Harris’ (1985)
conclusion that sex differences are sex-related and not dimorphic;
furthermore, it is consistent with Harris’ (1978, 1981) proposal that
sex differences, at least in visual memory, might be mediated through
different processing mechanisms. -

In study using college students, Snyder (1993) found that men’s
memory constructions were superior to women’s but only on the delay
recall trial. This difference accounted for less than two percent of
the total variance.

With older adults, Read (1987) also found that men performed
slightly better than women on delay recall. He also found that women
performed slightly better than men on the copy trial. In his study, the
examiner provided sequential cues on the copy trial to facilitate
organization; consequently, the results are not representative of
performance on standardized administrations. As in Snyder’s study, sex
differences were very small, accounting for less than two percent of the
total variance. In another study with older adults, Berry, Allen, and
Schmitt (1991) found no evidence for sex differences. These results
suggest that even if sex differences persist throughout adulthood, they
do not account for much of the variance in visual memory performance.

In summary of the CFT data, children show a complex interaction

between sex, age, and format of stimulus presentation; when given the

49
opportunity to draw the stimulus and not just view it, pubescent
children improved whereas younger children did not, and boys improved
more than the girls in the pubescent group (Waber, Bernstein, & Merola,
1989). With adults, men tended to outperform women on the delay memory
trial only, but differences were minimal when they appeared at all
(Bennett-Levy, 1984; Berry, Allen, & Schmitt, 1991; Read, 1987; Snyder,
1993). These differences in memory performance may be mediated by
organizational style and other intervening variables, and the form and
degree of mediation may change across the lifespan. A study that spans

adulthood to provide age- and sex-appropriate norms seems critical.

 

Prediction 7: The only sex difference on the CFT will be on the delay

recall trial, where men will outperform women.

 

DEVELOPMENT OF THE NEW MEASURES

Item Conetruction

 

A set of 20 recognition items were initially designed based on
findings in the literature and using patient test protocols in the
neuropsychology department of a North Carolina psychiatric hospital.
Each item was classified by two descriptors: the part of the figure
used (constructional element) and the type of error that was predominant
among the distractors, that is, the incorrect choices in the array
(constructional error).

Qegetguctienal elemente. Waber and Holmes (1985, 1986) described
developmental changes in the production of the CFT and quantified
organization and style. In doing so, they classified the original 18
scoring units (Lezak, 1983) into four groups. This system guided the
design of test items for the recognition instrument under construction
here. The base rectangle (BR) and main substructure (MS) contain the
more global elements of the figure: the large rectangle, the diagonal
cross, and the horizontal and vertical midlines. More specific elements
comprise the outer configuration (00; e.g., cross at far left, diamond
at far right) and internal detail (ID; e.g., circle with three dots,
five horizontals in upper left quadrant).

Constructional errors. The distractors for the original items were
derived using types of errors that have been frequently observed during

the construction of the complex figure and similar stimuli (Binder,

50

51

1982; Kaplan, 1988; Koppitz, 1964; Waber & Holmes, 1986). Rotatien (R)
items contain choices that are rotated varying numbers of degrees (90,
180, 270) from the accurate orientation as well as one choice that is a
mirror image. The choices on integretion (I) items lack connections at
intersections, lack good continuation, and lack closure. Choices on
other items are characterized by errors of distortion (D) such as
malalignment, disproportion, and misplacement. Finally, some items were
designed to be more sensitive to neglect (N). The responses on these
items were arranged in a vertical array; the features that discriminated
the correct choice from incorrect choices were concentrated on one side
of all the choices in order to evaluate the extent to which the subject
could attend to that side of the page and to that side of each drawing.

By recording the sequence of the lines drawn during the copy and
memory administrations of the CFT, some investigators have observed that
after age 13 the base rectangle and main substructure become
increasingly salient as organizational units. These units are typically
copied and recalled first, and then the more specific elements are added
(Milberg, Hebben, & Kaplan, 1986; Waber & Holmes, 1985, 1986). For this
reason, the items of the recognition task were arranged so that the
global parts are presented first (BR and MS items) followed by items
that incorporate the specific parts (00 and ID, except number 14). In
addition, within each half of the test, different constructional
elements were alternated as much as possible; that is, the first items
were in the sequence BR, MS, BR, MS, etc., and the latter items were
arranged 00, IC, 00, IC, etc. This was intended to guard against
comparisons across consecutive items and to limit the extent to which

previous choices could provide cues in successive responses. Where it

52
became necessary to place two 008 or two IDs together, stimuli were
chosen that contained different detail features, again to prevent
comparisons across items. This ordering is reflected in the final set
of items (Table 1).
Pilot Study

Using the 20 items described above, Fastenau and Manning (1992)
conducted a pilot study to examine the psychometric properties of the
new recognition items. The sample consisted of 42 adult volunteers on
staff at a North Carolina state psychiatric hospital. These subjects
were employed full-time, most in skilled positions. Education varied
from 12 to 18 years with a mean of 13.6 (SD = 1.6), age ranged from 18
to 55 with a mean of 32.5 (SD = 12.2), and the mean estimated IQ based
on a demographic formula (Wilson, ROsenbaum, & Brown, 1979) spanned the
Average-High Average range (91 to 116) with a mean of 104.5 (SD = 5.0).

Based on the preliminary data collected in that pilot study, the
recognition task merited continued use and further development. The
test data yielded point-biserial correlations that were mostly in the
range of .300 to .600. The coefficient alpha was .68 (p < .001). Raw
scores distributed fairly normally between 12 and 20 (M = 16.2, SD =
2.2), with a slight ceiling effect.

Some revisions were made based on the point biserial correlations
and based on other research. Specifically, one item was dropped because
the point biserial correlation was negative, indicating that those
subjects with good overall performances tended to fail that item. In
addition, the first item was correctly completed by virtually all
subjects; nonetheless, this item was retained to orient subjects to the

task. Finally, right neglect items were added based on Ogden’s (1987)

Table 1

53

Classificetion of Recogeition Items by Element and by Quadreg .

 

 

 

Item No. Element‘ Quadb Item No. Element"I Quadb
1 BR NS 16 ID RI
2 MS NS 17 ID RS
3 BR NS 18 ID LS
4 MS NS 19 0C RG
5 00 NS 20 ID LI
6 MS NS 21 QC RI
7 BR NS 22 ID RS
8 MS NS 23 OC L8
9 0C NS 24 OC LI

10 ID LS 25 ID RI
11 OC LG 26 ID RS
12 ID R1 27 0C LI
13 OC RS 28 OC RG
14 ID L8 29 ID NS
15 0C LI 30 0C RG

 

llBR=Base Rectangle, MS=Main Substructure, ID=Internal Detail,
OC=Outer Configuration
bLS=Left Superior Quadrant, LI=Left Inferior Quadrant, LG=Left

Global (Both Left Quadrants), RS=Right Superior Quadrant, RI=Right

Inferior Quadrant, RG=Right Global (Both Right Quadrants),

NS=Non8pecific

54
findings, cited previously, that after controlling for speech
impairments in studies of neglect, right hemisphere neglect was as
frequent as left hemisphere neglect, although the latter tended to be
less severe. Because the recognition task can circumvent language
deficits, it seemed especially important to design right neglect items
into the task. These revisions expanded the set to 27 items.

The constructional error labels were modified after the pilot study
because it was difficult to classify some items as exemplifying purely
one error in the absence of another error (e.g., rotated but not
distorted). Neglect was an exception because the items that were
classified this way met a reliable, objective construction criterion.
Specifically, all of the stimuli for a given item were identical on one
side of the midline. In the 27-item set, there were equal numbers of
left- and right-neglect items (six each) and these were equally
distributed between BR (one each), MS (one each), OC (three each), and
ID items (one each). The results of the pilot study are reported
elsewhere (Fastenau & Manning, 1992).

Expert Appraisal

Dr. Jane Holmes Bernstein of Children’s Hospital, Boston, has used
the CFT extensively with children (Waber, Bernstein, & Merola, 1989;
Waber & Holmes, 1985, 1986). She reviewed the 27 items generated from
the pilot study and suggested modifications based on her experience with
the CFT and based on her research in visual memory. Her first
recommendation was that I design the neglect items to allow for more
systematic examination of the different quadrants. Her second
recommendation was based on informal observations in her clinical work

with children with a variety of medical conditions. She observed that

55

the children occasionally shifted the left and right halves vertically
along the vertical midline. They also tended to separate the two halves
horizontally to leave a narrow alley in the center of their drawings,
which then was often left devoid of detail. Based on this information,
I modified several items to include distractors with empty alleys and
midline shifts. Finally, the items were revised to better reflect the
errors committed by the normal pilot group. It was expected that this
would eliminate the ceiling effect observed in the pilot study.
Fipal Revision

The final revised set contains 30 recognition items (Appendix A). -
Four items were designed for each of the four constructional quadrants
(left superior, left inferior, right superior, and right inferior);
these account for 16 of the 30 items. Three additional items reflect
gross right-side features (i.e., specific to the right side of the
figure but not to one quadrant or the other); another item contains
gross left-side features. Ten other items lack lateralization qualities
in their design. The proportions of Base Rectangle items (.10), Main
Substructure items (.13), Outer Configuration items (.40), and Internal
Detail items (.37) were made to match the same proportions in the
criteria used for scoring copy and recall constructions (.06, .17, .44,
and .33, respectively; derived from Taylor’s criteria reported in Lezak,
1983, p. 400). The classification of each item by constructional
element and constructional error is depicted in Table 1.

The matching items are reproduced in Appendix B. The matching set
includes one base rectangle item and one main substructure item. The
other eight consist of two items from each of the four quadrants. This

representative sample of the different quadrants was meant to provide an

56
additional means for assessing visual neglect. During testing, each
matching item is placed one at a time beside its corresponding
recognition stimulus card.

The new recognition and matching trials are expected to be reliable
and valid. Repeated testing for temporal consistency (test-retest
reliability) will not be conducted as part of this study. Because the
recognition and matching trials are scored objectively, interrater
indices will not be necessary. In addition, only one form of each trial
has been developed, so alternate-form reliability (consistency across
forms) will not be presented. The only measure of reliability will be
homogeneity of content, or internal consistency. This should be high
for the Total Scale and moderate for the subscales, which will have
fewer items and thus limit the size of Cronbach’s alpha coefficient.

Convergent validity will be demonstrated by significant
correlations with analogous measures and by confirmatory factor loadings
with tests of similar taxonomy (that is, visual-spatial secondary memory
and perception for recognition and matching respectively). Construct
validity will be evidenced by confirmatory factor loadings of each item
on its respective scale (Global element, Left Detail, and Right Detail).
Exceptions are Items 5, 9, and 29 which are specific to neither the left
side nor the right side of the stimulus. On the Matching Task, two
items (Items 1 and 6) are global-element items but these are too few to
construct a scale. Consequently, these two items will be omitted from
the factor analysis; the other eight items are expected to load onto one
of two factors, Left Detail and Right Detail.

For the recognition task, I expect this healthy sample to produce a

normal distribution of scores, with no floor or ceiling effects. For

57
the matching task, however, simplicity of the items for healthy persons
is expected to produce a negatively skewed distribution with a
substantial ceiling effect. A broad range of item difficulties is
expected by design: Some items were created to be mastered by
neurological patients so as to foster some degree of self-esteem and a
sense of competence within a potentially discouraging examination; these
should be answered easily by most or all healthy subjects. Other items
were intentionally designed to discriminate among higher functioning
individuals and, therefore, to be difficult even for healthy subjects.
Good discrimination will be measured for each item by a corrected item-
total correlation, the correlation between that item and the Total Score
after correcting for the direct contribution of the item to the Total
Score. A high item-total coefficient is expected on most items; modest
coefficients, however, are expected on items with low item difficulties
because success on an item by all subjects restricts the range of the

point-biserial. These expectations are summarized below.

 

Prediction 8a: The CFT Recognition Task will have homogeneous content,
as indicated by a strong Total Scale alpha reliability
and by moderate subscale alpha reliabilities.

Prediction 8b: The CFT Recognition Total Score will have good
convergent validity, that is, will correlate positively
with the WMS-R Visual Reproductions recognition score.

Prediction 8c: The CFT Recognition Total Score will have good
construct validity in that it will load on a factor

with other visual-spatial secondary memory test scores.

Prediction

Prediction

Prediction

Prediction

Prediction

Prediction

Prediction

Prediction

Prediction

8d:

8e:

8f:

9a:

9b:

9c:

9d:

9e:

9f:

58
As further evidence of construct validity, each item of
the CFT Recognition Task will load on the a priori
scale for which it was designed (Global, Left Detail, or
Right Detail), except items 5, 9, and 29.
CFT Recognition Total Scores will distribute normally
with no ceiling or floor effects.
Most items on the CFT Recognition Task will have
significant, positive corrected item-total correlations.
The CFT Matching Task will have homogeneous content, as
indicated by a strong Total Scale alpha reliability and
by moderate subscale alpha reliabilities.
The CFT Matching Total Score will have good convergent
validity, that is, will correlate positively with the
WMS—R Visual Reproductions matching score and with the
Judgment of Line Orientation test score.
The CFT Matching Total Score will have good construct
validity in that it will load on a factor with other
visual-spatial perception tests, and not with auditory-
verbal perception tests or with memory tests.
As further evidence of construct validity, each item of
the CFT Matching Task will load on the a priori scale
for which it was designed (Left Detail or Right Detail),
except items 1 and 6.
CFT Matching Total Scores will be negatively skewed
with a prominent ceiling effect.
Most items on the OFT Matching Task will have

significant, positive corrected item-total correlations.

 

METHOD

Subjects

The normative sample consisted of 90 healthy community-dwelling
adults. These individuals reported no active central nervous system
conditions and were living independently in the community at the time of
testing. Volunteers with uncorrectable visual impairment, uncorrectable
hearing impairment, or impaired use of the preferred hand were excluded
from the sample.

Subjects were solicited from three churches and one synagogue in
and near East Lansing, Michigan. Participating organizations received a
cash donation for each of the participants from their group. In
addition, participating organizations were provided a financial
incentive to recruit equal numbers of men and women from each of 10 5-
year age bands (30-34, 35-39, ... , 70-74, 75-and-over). With this
incentive, I created an age- and sex-stratified sample and minimized
socioeconomic bias within the age and sex groups. That is, because each
organization is equally represented among men and women and across age
groups, potential differences between organizations in socioeconomic
status are unlikely to confound age and sex analyses.

The total sample consists of 38 men and 52 women, fairly equally
represented (approximately 58% women) from age 30 through age 79+,
except in the 60-64 age group where the number of women far exceeded the

number of men (82% women). Efforts to remedy this disparity were

59

60

unsuccessful. In at least one large church, review of the census
indicated that men in their early 608 were very disproportionately
underrepresented. This same age-sex cell was difficult for all of the
groups to fill; therefore, this may reflect a broader cohort bias.
Based on 1990 census data, females consistently comprise 48.7 percent of
the population among the 1966 and 1990 birth-cohorts (0- and 24-year-
olds in 1990) and those in between. From the 1965 cohort backward, the
proportion of women increases curvilinearly across the seven successive
10—year age bands (49.6, 50.0, 50.7, 52.3, 55.9, 62.6, and 72.5,
respectively; U.S. Department of Health and Human Services, 1993).
Although no data are available for the specific 1929-1933 birth-cohorts
that were underrepresented here, the steady increase across the lifespan
and the relatively modest disproportion of women in this cohort in 1990
(52.3 percent) renders it unlikely that selective mortality could
account for the huge disparity (82 percent women) observed in this
sample. The reason for this bias is uncertain. Demographic data and
health data for the 90 subjects are presented in Tables 2 and 3,
respectively.
Instruments

The following descriptions are arranged by instrument. Some of the
instruments (for example, the Wechsler Memory Scale-Revised) contain
multiple subtests, each of which may test different functions from the
other subtests. It is parsimonious to discuss different tests within
these batteries within the same section, even though separate subtests
may measure very different processes. As an aid to the reader, Table 4
lists each test under the construct it is supposed to measure; Table 5

depicts the order of the tests within the testing sequence. The reader

61

 

 

 

 

 

Table 2
Demegrephic Data for the Sample, by Age Group.
Variable Mean Median Mode SD Min Max

Total Group (N = 90)
Age 55.9 54.5 ---- 14.1 29 88
Education 15.2 15.0 16 3.0 8 25
Vocabulary 12.5 12.5 12 2.3 5 19
Percent Women 57.8

Younger Adults (p = 47)
Age 43.5 43.2 ---- 7.4 29 54.9
Education 15.7 16.0 16 2.9 12 25
Vocabulary 12.3 12.0 12 2.6 7 19
Percent Women 55.3

Older Adults (p = 43)
Age 67.5 66.6 ---- 7.4 55 88
Education 14.6 14.0 12 3.1 8 21
Vocabulary 12.7 13.0 12 2.1 6 17
Percent Women 60.5
Nete. Education is reported in number of years completed. Vocabulary

scores are age-corrected scale scores from Wechsler (1981).

62

Table 3

Frequency of Health Risks, by Age Group.

 

Percentage of Group

 

Total Youngerl Older2

History of . . . N=90 p=47 p=43

 

Intracranial Conditions

Loss of consciousness following
a blow to the head 16 17 14

Unexplained loss of consciousness 8 6 9

Cerebrovascular accident or

transient ischemic attack 4 0 9*
Seizures , 1 2 0
Neurosurgery (intracranial) 1 2 0
Hydrocephalus 0 0 0

Systemic Conditions

Hypertension 31 23 40
Heart Disease 18 9 28
Diabetes 10 4 16
Renal disease 16 13 19
Hepatic disease 2 2 2
Pulmonary disease 10 9 12

 

1 Younger subjects ranged from 30 to 55 years old.
2 Older subjects ranged from 55 to 80+ years old.

' p < .001. All other comparisons, p 3 .005.

63

Table 4

Lieting of Tests by the Constructs They Purpprtedly Meesure.

Affective States
Beck Depression Inventory ("BDI")

State Trait Anxiety Inventory ("STAI"; State scale only)

Perception
Speech Sounds Perception Test ("SSPT"; Series A, B, and C only; AV)
Judgment of Line Orientation ("JOLO"; VS)

CFT Matching (VS)
WMS-R Visual Reproductions ("VR") Matching (VS)

Response Speed
WMS-R Alphabet Recitation (AV)

Counting 1-20 (AV)
Gross Letter Cancellation (VS)
Gross Symbol Cancellation (VS)

Attention and Concentration
Discriminative Letter Cancellation (AV)

Discriminative Symbol Cancellation (VS)

WMS-R Digit Span ("DS") Backward (AV)

WMS-R Visual Memory Span ("VMS") Backward (VS)
Phoneme Matching (AV)

Shape Matching (VS)

Production
Letter Fluency: C, F, L (AV)

Category Fluency: Animal, Fruit, Vegetables (AV)
CFT Copy (VS)
WMS-R Visual Reproductions ("VR") Copy (VS)

Primary (Working) Memory and Simple Attention
WMS-R Digit Span ("DS") Forward (AV)

WMS-R Visual Memory Span ("VMS") Forward (VS)
Secondary (Recent Long-Term) Memory, Immediate Free Recall
WMS-R Logical Memory I ("LM-I"; AV, Intentional)

Cowboy Story (AV, Incidental)
WMS-R Visual Reproductions I ("VR-I"; VS, Intentional)
CFT Immediate Recall (VS, Incidental)

Secondary (Recent Long-Term) Memory, Delay Free Recall
WMS-R Logical Memory 11 ("LM-II"; AV, Intentional)
Cowboy Story Delay Recall (AV, Incidental)

WMS-R Visual Reproductions II ("VR-II"; VS, Intentional)
CFT Delay Recall (VS, Incidental)

Secondary (Recent Long-Term) Memory, Delay Recognition
WMS-R Logical Memory Recognition (AV, Intentional)
Cowboy Story Recognition (AV, Incidental)

WMS-R Visual Reproductions Recognition (VS, Intentional)
CFT Delay Recognition (VS, Incidental)

Note. AV=Auditory~Verbal modality; VS=Visual-Spatial modality.
Abbreviations in quotation marks are provided for cross-reference; these
are conventions used in the field and/or abbreviations used in the text.

64
Table 5

Segpential Ordering of Tasks and Approximate Number of Minutes Reguired.

 

Tasks Time

 

Intro, Consent, Neuropsychological Questionnaire 10
CFT Segment (65 minutes)

Cowboy Story Immediate Free Recall

CFT Copy and Immediate Free Recall

HRB Speech-Sounds Perception Test

Judgment of Line Orientation

Cowboy Story Delay Free & Cued Recall, Recognition, Reading)
Symbol Cancellation (Gross and two different shapes)

CFT Delay Recall, Recognition, Matching

RAVLT (one trial only), 7/24

Category Fluency (Animals, Fruits, Vegetables)

HH

H
UIO'IO'IUINOIUIOU

Break (15 minutes)

Beck Depression Inventory 5
State Trait Anxiety Inventory 10

WMS-R Segment (65 minutes)

H

Logical Memory I (Immediate Free Recall)

Visual Reproductions I (Immediate Free Recall)

Digit Span (Forward 5 Backward)

Visual Memory Span (Forward 5 Backward)

Count 1 to 20

Alphabet tasks (Recitation, Phoneme Matching, Shape Matching)
Letter Cancellation (Gross, "E," "H")

Logical Memory 11 (Delay Free & Cued Recall, & Recognition)
Visual Reproductions II (Delay Free Recall)

Visual Reproductions Recognition, Matching, & Copy

WAIS-R Vocabulary

Letter Fluency ("C," "F," and "L")

H
wommmmwummmo

 

Total Time 155-180

 

Note. Most subjects completed the testing in 150 minutes. Total Time

 

ranged from 120 minutes to 180 minutes, with a few exceptions: Three

of the oldest subjects required two 2-hour sessions.

65
may find it useful to consult these tables while reading the following

descriptions.

Neuropsychological Questionnaire. Appendix C contains a copy of
the survey that was used to collect demographic and medical information
for the study. Questions address the participant’s sex, age,
educational background, and medical history. Age was computed in years
(from birth date to testing date); computations used ages that were
rounded to the nearest one-thousandth of a year. Education was recorded
in whole years completed, beginning with lst grade. A graduation
equivalency diploma (GED) was recorded as 12 years. College and
technical training were recorded in full-time, academic-year
equivalents; for example, half-time attendance for four semesters was
recorded as one year. Sex was coded 0 for men and 1 for women.

Medical history variables were also derived from questions on the
questionnaire. Hypertension (HTN) was assessed by the question, "Have
you ever had high blood pressure?"; coronary heart disease (CHD), "Have
you ever had a heart attack, heart disease, or any other heart
problems?"; diabetes mellitus (DM), "Have you ever had diabetes?";
pulmonary disease, "Have you ever had problems with your lungs?"; renal
disease, "Have you ever had problems with your kidneys?"; and hepatic
disease, "Have you ever had problems with your liver?" Cerebrovascular
disease (CVD) was assessed by the question "Have you ever had a stroke
or a transient ischemic attack (sometimes they’re called TIAs or mini-
strokes)?"; hydrocephalus by the question "Have you ever had
hydrocephalus (too much fluid in the brain)?"; seizures, "Have you ever
been treated for seizures or seizure prevention?"; and loss of

consciousness (LOC) following traumatic head injury, "Can you remember

66
the WQR§1 time that you hit your head or suffered a blow to your head?
[If yes] Were you unconscious?"

Current alcohol use was assessed by the question "On average, how
many alcoholic drinks do you have PER WEEK?" Follow-up inquiries
translated the subjects’ responses to a standard drink size of
approximately 0.6 ounces of pure alcohol (12 ounces of beer = 5 ounces
of wine = 3 ounces of port = 1.5 ounces of liquor; Albert, 1988, p. 78).
Current tobacco use was assessed by the question "Have you ever smoked
cigarettes? [If yes] At present, how many packs do you smoke PER WEEK?"

WAIS-R Vocabulary. As an index of broad intelligence, examiners
administered the Vocabulary subtest of the Wechsler Adult Intelligence
Scale-Revised (WAIS-R; Wechsler, 1981). This subtest is desirable
because of its ease of administration and because of its high
correlation with the Full Scale IQ (r = .85). It was also selected
because it is the most robust measure of intelligence across the
lifespan and is stable even following many forms and locations of
neurological insult (Russell, 1987; Wechsler, 1958); thus this provides
an IQ index that can be compared across normal and neurological groups
in different studies.

Letter Fluency and Category Fluency. A8 a measure of language
production skills, subjects were asked to say as many words as possible
that begin with a designated letter of the alphabet or within a given
category. Three trials were conducted using different stimulus letters
(C, F, L); another three trials used different stimulus categories
(animals, fruits, vegetables). A one-minute time limit was imposed.

More information about these tasks is available in Lezak (1983, pp. 329

ff).

67

HRB_§peeeh;§pppge Perception. Auditory—verbal perception was
assessed using the Halstead-Reitan Battery Speech-Sounds Perception Test
(Reitan & Wolfson, 1993). This measure was chosen because it provides
an auditory analogue that is close to the visual measures of perception.
Specifically, in the Speech-Sounds test, a tape-recorded nonsense word
is presented, and the subject is to identify that word from four
choices. The test score is the number of errors; only three of the six
10-item trials were administered so the total possible score was 30.

This test was chosen because the multiple-choice match-to-sample
format parallels that used in all three of the visual-spatial perception
measures included here (CFT Matching, WMS-R VR Matching, and JOLO).
Also, this task uses verbal stimuli instead of tones or rhythms; this
was important for verifying the ability to perceive similar verbal
stimuli used in the auditory-verbal memory tasks. The Seashore Rhythm
Test (Reitan a Wolfson, 1993) and the Woodcock Johnson-Revised Sound
Patterns (Woodcock & Johnson, 1989) are nonverbal and are therefore less
appropriate. Furthermore, Woodcock Johnson-Revised Sound Blending and
Incomplete Words (Woodcock a Johnson, 1989) lack the matching-to-sample
format desired for this study.

Jpggpent of Line Orientetion Test (JOLO). This is a well-
standardized and validated measure that measures perception for the
orientation of lines in two-dimensional space (Benton, Hamsher, Varney,
& Spreen, 1983). In this test, subjects are shown two cards. The top
card contains two stimulus lines; each line lies on the card in one of
11 angles, varying from 0° to 180° from the edge of the card. On the
bottom card are 11 sample lines radiating from a common vertex, spaced

equally from 0° to 180° from the bottom edge of the card. The subject

68
is instructed to choose the two lines on the bottom card that point in
the same direction as the two stimulus lines on the top card. There are
30 items; the score is the number of completely correct items (i.e.,
both lines matched correctly). This matching-to-sample format parallels
the administration of the CFT and VR matching tasks.

Alppebet trecking. The examiners administered two attention tasks
used by Coltheart, Hull, and Slater (1975) in their first experiment.
These investigators employed two simple tasks to measure relative
efficiencies of verbal and visual tracking:

For the verbal task, subjects were asked to proceed mentally

through the alphabet from A to Z, counting the number of letters

containing the sound "ee," including E. No external aids such as
speaking or writing were permitted. The subjects were asked to

perform as rapidly as possible and the time between the beginning
of the task and the utterance of the solution was measured. The
visual task, performed under the same conditions, was to proceed
mentally through the alphabet from A to Z, counting the number of
letters containing a curve in their upper-case form. Since no

information about the shape of a letter could assist in deciding

' and since no information

whether its name contains the sound "ee,'
about the sounds constituting the name of a letter could assist in
deciding whether its printed form contains a curve, we considered
that these tasks were to a sufficient degree purely verbal and
purely visual (Coltheart et al., 1975, p. 439).
These two alphabet tracking tasks were modified by DeLuca and Cicerone
(1991). In the revised administration, subjects first proceeded through

the alphabet silently and counted every letter containing the sound "ee"

69
(silent phoneme). Then they repeated the task, saying each letter aloud
as they proceeded (voiced phoneme). Next, subjects silently counted all
the curved letters (silent shape). They then repeated the task vocally
(voiced shape). A speeded recitation of the alphabet was also
administered.

Because silent phoneme and silent matching do not permit analysis
of the number of errors, these two trials were not analyzed in this
study. For the recitation and two voiced trials, speed and accuracy
were combined into a single variable, "seconds per target" (SPT):

Time

sp'r = --

 

Number Correct
Recitation SPT served as an index ofresponse speed. Phoneme Tracking
SPT and Shape Tracking SPT measured verbal and visual attention,
respectively.

Counting 1-20. As another measure of verbal response speed,
examiners asked the subjects to perform a presumably automated function,
similar to the alphabet recitation. Subjects were instructed to count
from one to 20 as fast as possible. Counting SPT will be computed as
described above.

Letterpgencelletion. The examiners administered six timed trials
of a letter cancellation task, similar to tasks described by Lezak
(1983, chapter 17) as measures of attention. Stimuli were large print
prose paragraphs (Appendix D). For the first two trials, the subjects
crossed out every letter on the page (gross cancellation). On the next
two trials, the subjects crossed out all of the e’s on the page. On the

last two trials, the target letter was "h." The two "e" trials

70
contained equal numbers of targets, and the two "h" trials also had
equal numbers of targets. In addition, each trial had equal numbers of
silent and pronounced targets.

Speed and accuracy were combined into a single variable, "seconds
per target" (SPT), computed by applying the same formula used "Alphabet
Tracking." SPTs were computed for each trial. SPTs for the first two
trials were summed to provide an index of manual response speed (Gross
Letter SPT). SPTs for the four discriminative trials were summed to
provide a measure of attention in the auditory-verbal modality (Letter
Tracking SPT).

Sypbol Cancelletion. As a visual-spatial analogue to the letter
cancellation tasks described above, I designed three timed cancellation
trials using three pages of geometric symbols as stimuli (Appendix E).
The examiners administered three timed trials of symbol cancellation,
similar to the tasks described in the "Letter Cancellation" section
above. For the first trial, the subjects crossed out every symbol on
the page (gross cancellation). On the next trial, the subjects crossed
out every occurrence of a specific symbol that the examiner circled in a
key at the top of the page. On the last trial, a different symbol in
the key was circled as the target.

Speed and accuracy were combined into a single variable, "seconds
per target" (SPT), computed by applying the same formula for "Alphabet
Tracking." SPTs were computed for each trial. SPT for the first trial
provided an index of manual response speed (Gross Symbol SPT). SPTs for
the two discriminative trials were summed to provide a measure of

attention in the visual-spatial modality (Symbol Tracking SPT).

71

QEI. The Complex Figure Test (CFT) has been described already.
The original copy trial, immediate recall trial, and delay recall trial
(20-minute delay) were supplemented by the recognition (Appendix A) and
matching trials (Appendix B) that were developed as part of this
project. Because of the unreliability of other scoring systems, I
developed my own scoring criteria (Appendix F) for the free recall and
copy drawings by modeling the criteria for WMS-R Visual Reproductions
(Wechsler, 1987). The interrater reliability on 23 complete protocols
spanning a wide range of ability was very good for the copy drawings (p
= .90) and exceptional for immediate and delay recall drawings (p = .97
for both; Fastenau & Denburg, 1994).

flM§;R. Several subtests from the Wechsler Memory Scale-Revised
were administered; psychometric information for this battery is
documented in the manual (WMS-R; Wechsler, 1987). Digit Span and Visual
Memory Span assessed verbal and visual attention, respectively; however,
these attention skills appear to be synonymous with primary memory.

There are four measures of secondary memory. Two are auditory-
verbal in nature: Logical Memory I (LM-I) for immediate recall and
Logical Memory II (LM-II) for delay recall. I wrote 10 multiple-choice
questions for each of the two LM stories to measure auditory-verbal
recognition memory. These questions were modeled after questions
written by Edith Kaplan (formerly of Boston Veteran’s Administration
Medical Center) in her "Process Approach" to the original version of the
Wechsler Memory Scale (Wechsler, 1945). The distractors for my set of
questions were designed using the scoring criteria (frequently occurring

incorrect responses) for LM in the WMS-R manual.

72

Two other measures of secondary memory are visual-spatial in
nature, using simple geometric designs for stimuli: Visual
Reproductions I (VR-I) for immediate recall and Visual Reproductions II
(VR-II) for delay recall. VR-II was followed by a recognition trial and
a matching trial; Hanger and his colleagues recently developed these
stinuli (Hanger, Montague, & Smith, 1991) and provided the items for
inclusion in this study (Appendix G). Subjects then copied the VR
stimuli a8 a demonstration of constructional ability (Fastenau & Sloan,
1993).

Digit Span Backward (DS-B) and Visual Memory Span Backward (VMS-B)
were included as measures of verbal and visual attention, respectively.
The use of these as attentional indices is based on a discussion by
Lezak (1983, pp. 550-551).

Finally, I had the subjects recite the alphabet as quickly as
possible, using the instructions from the Mental Control subtest of the
WMS-R. This overlearned and automated recitation served as a measure of
simple verbal response speed. Prorated time (SPT) was computed.

Cowboy Story. This memory task has been used in mental status
exams for many decades (Franz, 1919; Talland, 1965, ch. 9). Originally,
it was administered in a procedure very similar to WMS-R LM-I, as an
intentional memory task for aurally presented material; that is, the
patients were told in advance to try to remember the story and then they
listened to the story as it was read to them. However, in the present
study the subjects themselves silently read the passage from an 8 1/2" x
11" white page with large black print. The examiner gave the subjects
no indication that they should remember it. Immediately afterward, the

examiner removed the stimulus and asked the subjects to recall as much

73
as possible. After a delay of no fewer than 20 minutes, subjects were
asked again to recall as much of the story as possible. Thus, the
presentation was modified from an intentional aural administration to an
incidental reading administration, as an auditory-verbal analogue to the
ECFT.

I designed criteria for scoring the free recall reports, modeling
my criteria after the LM system in the WMS-R manual (Wechsler, 1987).

In addition, ten multiple-choice recognition items recorded in Talland’s
chapter were modified to more contemporary language. The story and
questions are appended (Appendix H). Some psychometric data are
presented in Talland (1965). Talland, however, used the standard
procedure (intentional, aural presentation of the story); there are no
data for the visual, incidental presentation used in this study.

Rey Auditory Verbal Leerning Test (RAVLT). In the standard
administration of the RAVLT (Rey, 1964, cited in Lezak, 1983; Taylor,
1959), subjects are read a list of 15 words and are asked to repeat as
many as words as possible. The first trial was believed to measure
immediate memory span, similar to other primary memory tasks like the
WMS-R and WAIS-R Digit Span subtests (Lezak, 1983, pp. 426-427).
Additional trials provide information about rate of learning and
interference. However, in this study, only one trial was administered.
The examiner first said, "I am going to read a list of words. Listen
carefully, for when I stop you are to say back as many words as you can
remember. It doesn’t matter in what order you repeat them. Just try to
remember as many as you can" (Lezak, 1983, p. 423). The following words
were then read to the subject at a rate of one per second, in accordance

with the standardized administration: Drum, Curtain, Bell, Coffee,

74

School, Parent, Moon, Garden, Hat, Farmer, Nose, Turkey, Color, House,
River. Because this measure exceeds the primary memory span and shows
primacy and recency effects in recall (indicative of mixed primary and
secondary memory involvement), these data will not be analyzed in this
study. It is described here so that the reader will know what other
tasks the subjects completed as part of the battery.

Z[2_. Barbizet and Cany (1968; reported in Lezak, 1983, p. 456-
459) developed a measure of visual memory that was modified by Rao,
Hammeke, and Huang (1982; also reported in Lezak, 1983) to be analogous
to the RAVLT. As with the RAVLT, recall on the first trial of 7/24 was
originally believed to represent the immediate memory span, but for
visual material. The white stimulus card contains a 4 x 6 rectangular
checkerboard with black dots in seven of the 24 squares. The examiner
says, "in a moment, I’m going to show you a card that looks like this,"
placing a blank 4 x 6 checkerboard answer sheet in front of the subject,
"only there will be black dots in some of the spaces. You will have 10
seconds to look at the card. Then I will take it away and let you fill
in this plepk one from memory. Try to remember which squares have dots
in them so that, afterwards, you can place each dot in the correct
square on the blank card. Don’t begin to draw it until I say ’Go.’
Ready?" The examiner reveals the stimulus card for 10 seconds and then
removes the card saying, "Now mark the boxes that had circles in them."

Beck Depression Inventory. In addition to the cognitive measures
described thus far, two instruments were administered to assess
emotional states that are reportedly salient to memory performance. The
Beck Depression Inventory (BDI; Beck, 1978) is a standardized and well-

validated self-inventory of depression. It includes four statements of

75
hierarchical intensity for each of 21 depressive symptoms. Subjects are
asked to indicate which one of the four statements best describes how
they have been feeling over "the past week, including today." The
weights of all the endorsed statements are summed for a Total Score,
which can be compared to normative data. A recent meta-analysis of
research with this instrument indicated that the BDI correlates well
with other inventories of depression and with physiological and
behavioral symptoms of depression. Furthermore, BDI Total Scores
effectively discriminate psychiatric from nonpsychiatric patients (Beck,
Steer, & Garbin, 1988). These findings were corroborated by other
reviews (Rehm, 1988).

Spete Tpeit Anxiety Inventory. Another emotional measure used was
the State Trait Anxiety Inventory (Spielberger, Gorsuch, Lushene, Vagg,
& Jacobs, 1983). Like the BDI, this is a standardized and well-
validated self-report inventory. Each anxiety scale, State and Trait,
consists of 20 statements accompanied by four-point intensity rating
scales. The State scale assesses how subjects feel "right now, at this
moment"; the Trait scale assesses how they "generally feel." This

instrument is reported to have strong psychometric properties (Chaplin,

1984).

W
Mass:
The investigator contacted five Christian churches and one
synagogue to solicit participation for the study. The purpose of the
study and the need for healthy adults living in the community were

described. Also explained were the testing (two and one-half hours of

76

pencil-and-paper tests of memory and related skills such as attention,
hearing, and drawing); the assurance of confidentiality; and the
subject’s right to decline any part of the testing and to withdrawal
from the project. Volunteers with uncorrected vision, uncorrected
hearing, or impaired use of their preferred hand were excluded from
participation. I offered each organization $10 for each volunteer who
completed testing; in addition, there was a $50 bonus for every 20
volunteers who matched the minimum age-sex distribution (i.e., 1 man and
1 woman in each of 10 5-year age bands). One church declined to
participate and another declined to coordinate the volunteer sign-up
based on their opposition to "fund-raising" in the church. Three
organizations generated 18 to 20 volunteers each, and another generated
33 volunteers; prorated bonuses were awarded to each organization for
having carefully attended to the age-sex criteria. Appended is the
confirmation letter that was mailed to the contact person at each
organization to be distributed to interested persons (Appendix 1).
Scheduling and Everyday Memory Questionnaires

Subjects’ names and phone numbers were provided to me by the
coordinator at each organization. I or one of the research assistants
called the subjects, explained the project to them, answered any
questions, scheduled an appointment at the MSU Psychological Clinic, and
then administered the Subject’s Everyday Memory Questionnaire. If the
subject gave permission, the investigator phoned the designated
informant for the Relative’s Everyday Memory Questionnaire.
Introduction

All subjects were tested individually. Most subjects came to the

MSU Psychological Clinic for testing, and testing was conducted in

77
quiet, private testing rooms. Five subjects with limited mobility were
tested in their homes; steps were taken to maintain the same quiet,
private testing conditions created at the Clinic.

Most subjects completed the testing in one session, lasting from
two to three hours. Three older subjects required two sessions for
optimal testing; in each case, the subject completed one full section
("CFT" or "WMS-R") followed by the "Break" section during the first
visit; this ensured that depression and anxiety measures immediately
followed the first section of testing for all subjects.

The following description of the testing session is also outlined
in Table 5. When subjects arrived for testing, the examiners introduced
themselves. They then presented the consent form, which explained in
writing the nature of the study, the confidentiality of the data, and
the subject’s right to terminate participation at any time (Appendix J).
After the consent form was signed, the examiner asked a standard set of
questions regarding the subject’s neuropsychological history (Appendix
C). These preliminary steps helped establish rapport with the subject
before formal testing began.

The examiner prefaced the actual testing by saying, "We’ll work for
about one hour. Then, we’ll have a break before we begin again. Would
you like to get some water or use the bathroom now, before we get
started?" When the subject was ready to begin, the examiner said, I’m
going to ask you to do a lot of things today. Some of them may seem
very easy to you, perhaps even silly. Other tasks may challenge you a
bit. There may even be some things that you can’t do at all. On every

task, just do your best, and then we’ll move on to the next one."

78
C Se ment

The CFT segment and the WMS-R segment were counterbalanced. That
is, some subjects took the CFT segment before the break, followed by the
WMS-R segment (Table 5). The other subjects took the WMS-R segment
first, followed by the CFT segment. The assignment to these two
conditions was stratified within age-sex cells.

Cowboy Story, Immediete Recell. The CFT segment began with the
Cowboy story. The examiner said, "Take a moment and read this story to
yourself. Tell me when you are done." This was not introduced as a
memory task. After the subject finished, the examiner removed the
stimulus page from view and said, "Now what did you just read? Tell me
everything and begin at the beginning." The examiner recorded the
subject’s report verbatim. Then the examiner asked the subject to tell
the story one more time to verify that nothing was omitted. The
examiner would not change or delete anything from the first record but
would only add material that had not been recorded in the first report.
Using the WMS-R Logical Memory scoring criteria as a model, I developed
scoring criteria for the Cowboy Story; these criteria were used for
delay recall also (Appendix H).

0E1, Copy. The examiner placed a clean 8 1/2 x 11 sheet of white
paper (marked with a "C" and the subject number in one corner) on the
table in front of the subject and said, "I have something I want you to
copy VERY,CAR§FULLY. As you work, I will hand you different colored
pens. When I hand you a new pen, place the other pen aside and continue
with the pen I hand you." The examiner handed the subject the first

pen, turned the stimulus card over, and said, "Draw this picture as

79
ACCURATELY as you can. Try to make your drawing look identical to the
picture in every way."

A stopwatch, visible to the subject but not emphasized, was then
started and stopped when the subjects gave clear indication that they
were finished. If subjects indicated that they could not draw (e.g.,
"I’m no artist"), the examiner responded, "Do your best. Draw it as
accurately as you can." If the subject expressed concern about being
timed, this same examiner response was provided. Examiners changed pens
at approximately 20-30 second intervals. The new pens were introduced
smoothly but assertively when the subject lifted his or her hand/pen
between lines. If the subject ignored the new pen for more than five
seconds, the examiner said, "Now take this pen." The examiner recorded
the subject’s hand preference, the Color sequence of the pens, and the
time to completion. Drawings were scored only for accuracy (Appendix
F), not for organizational style.

CFT, Immediate Recall. Following the copy trial, the examiner
removed the stimulus card and the copy drawing and placed them out of
sight. Then, the examiner laid down a new, clean sheet of paper (marked

"I" and the subject number in one corner). Handing the subject

with an
the first pen, the examiner said, "Now draw as much of it as you can
from memory. Draw the picture as ACCURATELY as you can." Pens were
introduced at approximately 15-20 second intervals initially, and then
the timing was gauged by the subject’s speed. Color sequence was
recorded by the examiner after each pen was introduced. Again, a timer
was started. When the subject finished, the examiner said, "I always

tell people to take another moment in case something else comes to

mind." The time was recorded at the second indication that the subject

 

80
had finished. The subject was not told that there would be another
memory trial. A delay was then imposed by the administration of the
following tasks.

Speech-Sound§ Perception Teet. This test was administered and
scored in standardized fashion (Reitan & Wolfson, 1993, pp. 267-271) but
using only the first three trials instead of all six. Consistent with
the manual and with other studies using this measure, scores on this

task reflect the number of errors, with 30 points possible.

 

Judgment of Line Orientation. This test was also administered and
scored in standardized fashion (Benton, Hamsher, Varney, & Spreen, 1983,
ch. 5). Scores reflect the number of items that are completely correct,
with 30 points possible.

gewboy Story, Delay Recall. The examiner said, "Do you remember
the little story you read a few minutes ago? Now I want you to tell me
that story again. Tell me everything; begin at the beginning." If the
subject could not remember the story, the examiner said, "The story was
about a cowboy." However, no credit was given for the first phrase ("A
cowboy") if the reminder was required. The examiner recorded the
subject’s report verbatim and then asked the subject to tell the story
one more time to verify that nothing was omitted. The examiner did not
change or delete anything from the first record but only added material
that had not been recorded in the first report. Finally, the examiner
handed the original Cowboy stimulus to the subject again and said,
"Okay, now I would like you to read this story again, but this time read
it out loud." While the subject read, the examiner noted any errors.
When scoring memory trials, the recall was compared to what was actually

read.

81

Rev Auditory VerbelpLeerning Teet. The examiner introduced the
next task by saying, "I am going to read a list of words. Listen
carefully, for when I stop you are to say back as many words as you can
remember. It doesn’t matter in what order you repeat them. Just try to
remember as many as you can." The following words were read at a rate
of one word per second, and not one second between words: drum,
curtain, bell, coffee, school, parent, moon, garden, hat, farmer, nose,
turkey, color, house, river. The examiner recorded the order in which
the words were repeated and wrote any intrusions verbatim. Plurals were
accepted for credit; intrusions and omissions were scored as errors.

Gross Symbol Cancellation. The examiner placed the stimulus sheet
(SC-1) and a pencil in front of the subject and said, "This is a test to
see how quickly you can draw a line through each symbol on this page,
like this." Then, the examiner drew a line through each symbol in the
"Key," from left to right. "When I tell you to start, you do the rest
of them. Begin here (pointing to the top left) and draw a line through
every symbol as quickly as you can, one after the other, without
skipping any. Meke eure theteyour line touchee one and only one sympel.
When you finish this line (sweeping finger across the top line), go on
to the next one (pointing). Keep working until you reach the end
(pointing). Work as quickly as you can without making mistakes. Ready.
Go!" Timing was started immediately and was stopped as soon as the
subject cancelled the last symbol (bottom right) or gave clear
indication of being finished. If subjects asked whether they could
sweep left-to-right down one line and then right-to-left down the next,
the examiner repeated the corresponding direction with emphatic sweeping

gestures; however, if subjects used such a zigzag motion during the

82

timed trial, they were not interrupted. The time to completion was
recorded; later a template was used to count omissions. Omissions were
scored if there was no line present for a target or if less than 50% of
the line closest to a target was within the box of that target; a line
that touched two targets received credit for only one target.

Discriminative Symbgl Cencelletion. SC-l was removed immediately
so that subjects could not review their performance. SC-2 was then
placed on the table. "Good. Let’s try another one. This time, I’m
going to circle one of the symbols in the key. I want you to go across
the page, same as before, but this time cross through only the symbol I
circle. Make sure that your line touchee one and only one symbol.
Begin here (pointing to the top left) and draw a line through that
symbol each time you see it. When you finish this line, go on to the
next one. Keep working until you reach the end. You are looking for

this symbol (circling #2 in the Key: the horizontal with a dot

 

underneath). Work as quickly as you can without making mistakes.
Ready. 00!" The time to completion was recorded; later a template was
used to count omissions.

SC-3 was presented in similar fashion. "Good. Let’s try one more.
This time, I’m going to circle a different symbol in the key. You are

looking for this symbol (clrcling #5, the center symbol, in the Key: the

 

"T" that is turned clockwise 90°). Work as quickly as you can without
making mistakes. Ready. Go!" Time and omissions were recorded. For
all three symbol cancellation trials, time and omissions were combined
into a summary score, "Seconds Per Target" (SPT), described earlier

("Instruments" section).

83

CFT, Delay Recall. At this point, the examiner placed the color
markers on the table with a clean sheet of white paper and said, "Do you
remember that picture you copied for me? I want you to draw as much of
it as you can from memory. Draw the picture as ACCURATELY as you can."
Again, a timer will be started. Pens were changed at approximately 10—
15 second intervals initially; the timing was then gauged by the
subject’s speed. The sequence of the pens was recorded. When the
subject finished, the examiner said, "Take another moment and see if
anything else comes to mind." The time was recorded at the second
indication that the subject is finished. Scoring was described earlier.-

CEI, Recognition. The examiner presented the first recognition
item and said, "Which one of these was part of the ORIGINAL picture,
which you copied a little while ago? Point to the right one." If the
subject was correct ("a" is correct), the examiner said, "That’s right.
Of these five, this one looks most like the original picture," and then
recorded the response. If the first response was incorrect (any R

"a"), the examiner said, "No, that’s incorrect. Do you remember

except
the card that was lying on the table? If the card were sitting on this
page facing you (here, the examiners outlined the imaginary card with
their fingers, above the page), how would the picture look to you?" If
the subject was correct on the second attempt, the examiner said,
"That’s right. Of these five, this one looks most like the original
picture," and continue with Number 2. If incorrect again, the examiner
said, "No, the picture on the card had this shape" (pointing to the
correct answer). If the subject appeared confused, the examiner

elaborated as necessary but without describing any more elements of the

stimulus. This was to make sure that the subject understood the task

84
and to ensure that the subject was correctly oriented to the layout of
the drawings. Whether or not the confusion cleared, the examiner
continued with #2, offering no more help.

The examiner presented the second test item saying, "Which one of
these was part of the original picture, which you copied a little while
ago? Point to the right one." The examiner recorded the response but
provided no feedback. These instructions were repeated as necessary for
each item, abbreviating or eliminating the directions as the subject
became more familiar with the task.

Each item answered correctly received one point credit, with all 30
items contributing to the Total Scale. Items 1-4 and 6-8 comprise the
Global Scale. Items 5 and 9-30 comprise the Detail Scale. The Right
Neglect Scale consists of Items 5, 12, 13, 16, 17, 19, 21, 22, 25, 26,
28, and 30. The Left Neglect Scale consists of Items 10, 11, 14, 15,
18, 20, 23, 24, and 27.

CFT, Matching. The first matching stimulus card was placed
alongside the corresponding multiple-choice array (identical to the
array used in the recognition task), and the examiner said, "Which one
of these (pointing to multiple-choice array) looks like this (pointing
to stimulus) picture? Point to the right one." If the subject was
correct, the examiner said, "That’s right. Now try this one." If the
subject was incorrect on the first item, the examiner said, "No, THIS
one (pointing to the correct response) looks the most like this
(pointing to the stimulus) picture. Try this one. Which one of these
looks like this picture (pointing)?" Feedback was not provided on any

remaining items.

85

Each item answered correctly received one point credit, with all 10
items contributing to the Total Scale. The Right Neglect Scale consists
of Items 4, 5, 9, and 10. The Left Neglect Scale consists of Items 2,
3, 7, and 8.

1[_5. The examiner then placed a blank checkerboard answer sheet
(with an asterisk in the upper right hand corner) in front of the
subject and said, "In a moment, I’m going to show you a card that looks
like this, only there will be black dots in some of the spaces. You
will have 10 seconds to look at the card. Then I will take it away and

let you fill in this blank one from memory. Try to remember which

 

squares have dots in them so that, afterwards, you can place each dot in
the correct square on the blank card. Don’t begin to draw it until I
say ’Go.’ Ready?" Then the examiner revealed the design on the
stimulus card by turning the card so that the asterisk was in the
subject’s upper right hand corner. After 10 seconds, the stimulus card
was removed and the subject was instructed, "Now mark the boxes that had
circles in them. Go." Any mark was scored; a circle was not necessary
for credit. Omissions and commissions were recorded separately.

Letter Fluency. Finally, the examiner said, "I’m going to say a
letter of the alphabet and then I want you to tell me all of the words
you can think of that begin with that letter, as fast as you can. But
proper names are not allowed, so if the letter were ’B’ you would NOT
say ’Boston’ or ’Bob.’ (After a pause) The letter is ’C.’ Now tell me
all the words you can think of that begin with the letter ’C.’ Ready.
Begin!" The examiner recorded every word produced for one minute,

noting each 15 second interval with a hash mark.

86
At the end of the first trial, the examiner said, "Okay. Now we’re
going to do the same thing again, but this time tell me all the words
you can think of that begin with the letter ’F.’ Do it as fast as you
can. Ready. Begin!" Again, the examiner recorded the responses for
one minute, noting each 15 second interval with a hash mark.

- The third trial was introduced in similar fashion: "Good. Now,
we’ll do it one more time, but this time tell me all the words you can
think of that begin with the letter ’L.’ Do it as fast as you can.
Ready. Begin!" Responses were recorded as before. On all three
trials, each accurate response received one point credit.

Breek Between Segeeppe

Between the CFT segment and the WMS—R segment, irrespective of the
order of the segments, we gave the subject a break and asked them to
complete the State-Trait Anxiety Inventory (STAI) and the Beck
Depression Inventory (BDI). "Now you can take a brief break. Then take
a few minutes to complete these two questionnaires. Be sure to complete
both sides of each one. Notice that this one (BDI) asks about how
you’ve been feeling over ’THE PAST WEEK, INCLUDING TODAY’ (underlining
that phrase in the directions). This one (STAI) is different on each
side. On this side, describe how you feel ’RIGHT NOW. ... AT THIS

MOMENT’ (underlining). On the other side, describe how you ’GENERALLY

 

EEEL’ (underlining). Please don’t write your name on either of these
forms." Both instruments were scored in accordance with their manuals.
When the subject was ready, testing resumed.
WMS-R Segment

Logical Memory . The WMS-R segment began with Logical Memory (LM)

I, immediate free recall for paragraph-length stories, in strict

87
adherence to the directions and scoring criteria in the manual
(Wechsler, 1987).

Visual Reproductions 1. Visual Reproductions (VR) 1, immediate
free recall drawings for geometric designs, followed LM I and were
administered and scored as instructed in the manual (Wechsler, 1987).

Digit Span. Digit Span, repetition of numerical strings, was
administered next in standardized fashion (Wechsler, 1987).

Visual Memory Span. Visual Memory Span followed (Wechsler, 1987).

Count 1-20. The next task was introduced as follows: "I want to
see how quickly you can count from 1 to 20, like this--1, 2, 3, all the
way to 20. Go ahead!" The timer was started immediately after the
directions were concluded. Time, omissions and commissions were
recorded; time and omissions were combined into the summary score,
"Seconds Per Target" (SPT), described earlier in this chapter
("Instruments" section).

Alphebet Recitetion. Then the examiner said, "Now I want to see
how quickly you can say the alphabet for me, like thi8--A, B, C. Go
ahead!" The timer was started immediately. Time, omissions and
commissions were recorded, and SPT was computed.

Silent Phoneme Matching. The examiner then said, "Now I want you
to silently think of the alphabet and, without saying anything, count
every letter that contains the sound ’ee.’ As quickly as you can, tell
me how many letters contain the sound ’ee.’ Go ahead!" The timer was
started immediately and stopped when the subjects gave their answers.
Time and the subject’s response were recorded. This format only permits
a dichotomous score on accuracy, that is, whether or not the response

matches the correct answer.

88
Voiced Phoneme Matching. The examiner continued, saying, "Now
we’ll do it again, only a little differently. This time, I don’t want

you to count the letters. Instead, start from the beginning and tell me

 

every letter that contains the sound ’EE’ as you go. Do it as quickly
as you can without any mistakes. Go ahead!" The timer was started
immediately and stopped when the subjects provided the first indication
of being finished (e.g., looking up at the examiner or dropping the
pitch in their voices) QR as soon as they reach "Z" (if they recited in
order). Time, omissions, and commissions were recorded, and SPT was
computed.

Silent Shepe Metching. The next task was introduced. "Now we’ll
do a different exercise. Once again silently think of the alphabet, but
this time picture it typed in upper—case form-~that is, all capital
letters. Without saying anything, count every letter that contains a
curve. As quickly as you can, tell me how many letters contain a curve.
Go ahead!" The timer was started immediately after saying "Go ahead"
and stopped when the subjects gave their answers. Time and the
subject’s response were recorded. Again, this format permits only
dichotomous scoring on accuracy.

Voiced Shepe Matching. The examiner continued, saying, "Now we’ll
do it again, only a little differently. This time, I geplp want you to

count the letters. Instead, start from the beginning and tell me every

 

letter that contains a curve as you go. Do it as quickly as you can
without any mistakes. Go ahead!" The timer was started immediately and
stopped when the subject provided the first indication of being finished
(e.g., looking up at the examiner or dropping the pitch in their

voices). Every response was recorded, together with the total time.

89
Subjects were not penalized for reporting letters out of order. Time,
omissions, and commissions were recorded, and SPT was computed.

Gross Letter Cancelletion ("LC-1" and "LC—2"). Next, the examiner
placed the stimulus marked "LC-1" in front of the subject with a pencil
and said, "This is a test to see how quickly you can draw a line through
each letter on this page, like this (drawing a line through each letter
in the key words above the paragraph, from left to right). When I tell
you to start, you do the rest of them. Begin here (pointing to the top
left) and draw a line through every letter as quickly as you can, one
after the other, without skipping any. When you finish this line, go on
to the next one (pointing). As soon as you get to the end, cross
through the word ’FINISHED’ here (pointing). Make sure that your lipe
touchee one end only one letter. Work as quickly as you can without
making mistakes. Ready. Go!"

The examiner started timing immediately and stopped timing as soon
as the subjects canceled the word "Finished" (bottom right) or gave
clear indication that they were finished. If necessary, the following
prompts were used to redirect the subject: "Do them in order. Don’t
skip any. Do this one next" (pointing to the item omitted), or "Make
sure that your line touches one and only one letter." Time was recorded
and a template was used to count omissions (absence of a unique line for
one of the targets). Time and omissions were combined into a summary
score, "Seconds Per Target" (SPT), described earlier in this chapter
("Instruments" section).

Removing LC-l and placing LC-2, the examiner said, "Good. Let’s
try another one. You’ll do this one the same way as the last one.

Begin here; as soon as you get to the end, cross through the word

90
’Finished.’ Mekepepre thet eech line touchee one and only one lette .
Work as quickly as you can without making mistakes. Ready. Go!"
Timing and scoring were the same as on LC-l.

Discriminative Letter Cancellation. The examiner then exchanged
LC-2 for LC-3 and said, "Okay. Let’s try another one. This time, I’m,
going to write a letter at the top of the page. I want you to go across
the page, same as before, but this time cross through only the letter I
write. Begin here (pointing to the top left) and draw a line through
that letter each time you see it. As soon as you get to the end, cross
through the word ’Finished’ here (pointing). You are looking for the -
letter ’E’ (writing ’E/e’ at the top of the page). Work as ui k as
you can without making mistakes. Ready. Go!" Timing and scoring were
the same as on LC-l.

Upon completion of LC-3, the examiner introduced LC-4, saying,
"Good. Let’s try another one. You’ll do this one the same way. You
are looking for the letter ’E’ (writing ’E/e’ at the top of the page).
Work as quickly as you can without making mistakes. Ready. Go!"
Timing and scoring were the same as on LC-l.

The examiner then exchanged LC-4 for LC-5 and said, "Okay. Let’s
try another one. This time, I’m going to write a different letter at
the top of the page. I want you to go across the page, same as before,
but this time cross through only the letter I write. You are looking
for the letter ’H’ (writing ’H/h’ at the top of the page). Work as
quickly as you can without making mistakes. Ready. Go!" Timing and
scoring were the same as on LC-l.

Upon completion of LC-5, LC-6 was introduced saying, "Good. And

now I have just one more. You’ll do this one the same way. You are

91
looking for the letter ’H’ (writing ’H/h’ at the top of the page). Work
as quickly as you can without making mistakes. Ready. Go!" Timing and
scoring were the same as on LC—l.

Legleel Mepory II. Delay free recall for the LM stories was
administered according to the directions in the manual and scored by the
criteria in the manual (Wechsler, 1987).

Vlepel Reproductione II. Delay free recall for the VR designs was
administered according to the directions in the manual (Wechsler, 1987).
The drawings were scored using the manual’s criteria.

Visual Reproductions, Recogpition. The examiner presented the
first recognition item and said, "Which one of these was the original
picture which you copied a little while ago? Point to the right one."
The first response was recorded and then feedback was provided. If the
subject was qorrect (correct R is "d"), the examiner said, "That’s
right. Of these five, this one looks most like the original picture."

If the first response was incorrect (any R other than "d"), the
examiner said, "No, that’s incorrect. Do you remember the card that was
lying on the table? If the card were sitting on this page facing you
(the examiner outlining an imaginary card with his or her finger, above
the page), how would the picture look to you?" If the subject was
correct on the second attempt, the examiner said, "That’s right. Of
these five, this one looks most like the original picture." If the
subject was incorrect again on the eeeppq attempt, the examiner said,
"No, the picture on the card had this shape (pointing to the correct
answer)."

If the examinee appeared confused, the examiner elaborated as

necessary in order to help the subject understand the task and to ensure

92
that the subject was correctly oriented to the layout of the drawings.
Whether or not the confusion cleared, the response was recorded and the
examiner continued with Item 2. No further help was offered.

In presenting Item 2, the examiner said, "Which one of these was
the original picture which you copied a little while ago? Point to the
right one." Instructions were abbreviated as subject became familiar
with the directions. Response were recorded without providing feedback.
Each correct response received one point credit, and item scores were
summed for a maximum Total Scale score of 5.

Visual ReproductionsL Metching. Then, the first matching stimulus
card (Card A from the WMS-R Kit) was placed alongside of the
corresponding multiple-choice array (identical to the array used in the
recognition task), and the examiner said, "Which one of these (sweeping
finger across the multiple-choice array) looks like this picture
(pointing to the stimulus)? Point to the right one."

The response was recorded, and feedback was provided on the first
item. If correct, the examiner said, "That’s right. Now try this one."

If incorrect, the examiner said, "No, this one (pointing) looks the most

 

like this picture (pointing). Try the next one. Which one of these
looks like this picture (point)?" No further feedback was provided.
Response were recorded without providing feedback. Each correct
response received one point credit, and item scores were summed for a
maximum Total Scale score of 5.

Visual Reproductions, Copy. The examiner placed the blank "VR-II
Copy" sheet, ppfolded, on the table in front of the subject. The first
Visual Reproduction card ("Flags") was placed above the top of the page.

The examiner instructed the subject, "Copy this picture as accurately as

93
you can. Try to make your drawing look identical to the picture in
every way." Draw it here (pointing to the first blank space on the
record sheet). If subjects indicated that they could not draw (e.g.,
"I’m no artist"), the examiner responded, "Do your best. Draw it as
accurately as you can." The subject was permitted to erase.
- This procedure was repeated for each of the other three stimulus

' and "Rectangles/Semicircle"), with the

cards ("Circles," "Windows,'
examiner emphasizing accuracy on eyepy stimulus. The drawings were
scored using the WMS-R Visual Reproduction criteria (Wechsler, 1987).

WAIS-ReVocebulery Subteet. This test was administered and scored
according to the manual (Wechsler, 1981).

Category Fluency. The examiner introduced the next task saying, "I
want to see how many different animals you can name. Any animals will
do; they can be from the farm, from the jungle, from the ocean or even
house pets. For instance, you can start with ’dog.’ You have one
minute. Try to say as many as you can until I tell you to stop. Ready.
Begin!" The examiner recorded every word produced for one minute,
noting each 15 second interval with a hash mark.

At the end of the first trial, the examiner said, "Good. Now we’re
going to do the same thing again, but this time I want to see how many
different fpplpe you can name. Any fruits will do. You have one
minute. Try to say as many as you can until I tell you to stop. Ready.
Begin!" Again, the examiner recorded the responses for one minute,
noting each 15 second interval with a hash mark.

The third trial was introduced in similar fashion: "Okay. Now,

we’ll do it one more time, but this time I want you to name as many

different vegetables as you can. Do it as fast as you can. Ready.

94
Begin!" Responses were recorded as before. On all three trials, each
accurate response received one point credit.

Debriefing. At the end of the testing session, the examiner handed
subjects the debriefing form (Appendix K) and asked them to read it.
Upon indication that they had read the debriefing sheet, the examiner
encouraged the subjects to take the form with them and then thanked them

for their participation.

RESULTS

Correlations were used to describe bivariate relationships.
Pearson product-moments were used when both variables were continuous
and interval in nature; point-biserials were computed when one variable
was continuous and interval and the other was discrete and nominal. For
all predicted relationships, the direction was specified so one-tail
tests were applied; when no relationship was expected, two-tail tests
were applied. Because of the excessive number of analyses being
performed on this data set (9 predictions involving multiple analyses
each), a stringent cutoff for Type I error was applied to guard against
alpha inflation. Following the Bonferroni convention (recommended by
Cohen & Cohen, 1983, p. 504), I divided the desired alpha, .05, by the
expected number of significance tests (approximately 55) for a stringent
alpha of .001. At an experimentwise error rate of approximately 1 - (1-
alpha) 8'1, where g is the number of comparisons or significance tests
to be performed, an alpha of .001 over 55 significance tests keeps the
experimentwise Type I error rate at .05. In addition, I collapsed
across trials and across variables to create summary variables wherever
it was prudent, in order to minimize the number of independent variables
(e.g., combining error scores and time into a "Seconds Per Target" score
on Letter Cancellation and taking the mean of all equivalent trials;

Cohen & Cohen, 1983, chapter 4).

95

96

In the conventional, "variance decomposition" model, p2 best
estimates the percentage of variance in the dependent variable that is
explained by the independent variable (Cohen & Cohen, 1983, p. 100;
Ozer, 1985). In "variance partitioning" models, however, the
correlation itself best estimates the variance explained. This model
applies in several situations, including the determination of measured
variables by a latent trait. In concurrent validity, for example, two
tests are assumed to measure the same construct; p, and not p2, can be
interpreted as the amount of variance in those tests that is explained
by that construct (Ozer, 1985). In addition to its use for estimating
variance explained, some advocate that y, rather than £2, is a better
indicator of effect size (Gorsuch, 1991; Nelson, Rosenthal, & Rosnow,
1986; Rosnow & Rosenthal, 1988). Both p and :2 will be reported.

The predictions in this study were numbered according to their
appearance in the literature review. However, for the purposes of
describing the results, another ordering seemed more logical.
Consequently, the results and the discussion begin with Predictions 4,
5, and 6, followed by Predictions 1, 2, and 3. Predictions 7, 8, and 9

complete the results and discussion chapters.

 

Prediction 4a: Attentional skill and perceptual accuracy will correlate
negatively with age in both auditory-verbal and visual-

spatial modalities.

 

Correlations with age were computed for the summary attentional and
perceptual measures in each modality. Attention in the auditory-verbal

domain was measured by Discriminative Letter Cancellation (the mean

97
seconds-per-target [SPT] for trials 3 through 6), Digit Span Backward
(raw score), and Voiced Phoneme Matching (SPT). Visual-spatial
attention was measured by Discriminative Symbol Cancellation (the mean
SPT for trials 2 and 3), Visual Memory Span Backward (raw score), and
Voiced Shape Matching (SPT).

Perception in the auditory-verbal modality was measured by SSPT
(number of errors). Visual-spatial perception was measured by JOLO
(number correct), CFT Matching Total Score, and VR Matching Total Score.
The correlations between age and all of these variables are tabulated
below (Table 6). For eight of the 10 tests, an age-related decline in
skill was observed (p 5 .001). Contrary to the prediction, however, two
measures did not correlate with age: Digit Span Backwards (p =.01) and

CFT Matching (p =.01).

 

Prediction 4b: Production ability will correlate negatively with age in
the auditory-verbal modality but will not correlate with

age in the visual-spatial modality.

 

Correlations with age were computed for the summary production
measures also. Auditory-verbal production was measured by Letter
Fluency (the sum of all three trials) and Category Fluency (the sum of
all three trials). Visual-spatial production was measured by CFT Copy
and VR Copy scores. This prediction was not well supported.
Performance on Category Fluency and Visual Reproductions Copy Trial
declined with age (p < .001), and there was a trend for CFT Copy (p
=.005); however, Letter Fluency did not correlate with age (p 2 .05).

These correlations are tabulated below (Table 7).

98
Table 6

Correletione With Age for Attention and Perception, by Modality
(Prediction 4a).

 

 

 

Test 2 2
Attention
Auditory-Verbal
Letter Cancellation, Discrimination Trials .51 .0005
WMS-R Digit Span Backwards -.26 .01
Phoneme Matching .38 .0005
Visual-Spatial
Symbol Cancellation, Discrimination Trials .51 .0005
WMS-R Visual Memory Span Backwards -.49 .0005
Shape Matching .45 .0005
Perception
Auditory-Verbal
Speech-Sounds Perception Test .46 .0005
Visual-Spatial
Judgment of Line Orientation -.39 .0005
CFT Matching -.25 .01

WMS-R Visual Reproductions Matching -.31 .001

 

99
Table 7
CorreletionepBetween Age and Production, by Modality (Prediction 4b).

 

 

Test E D
Auditory-Verbal
Letter Fluency -.11 >.05
Category Fluency -.43 .0005
Visual-Spatial
CFT Copy -.29 .005
WMS-R Visual Reproductions Copy -.36 .0005

 

Note. All correlations in this table represent decline with age;

 

variables with positive correlations were scored in terms of errors or
time required to completion; variables with negative correlations were

scored in terms of number correct.

100

 

Prediction 5a: The frequency of HTN, CHD, CVD, DM, and pulmonary
disease will correlate positively with age.

Prediction 5b: Current alcohol and tobacco use will correlate
negatively with age.

Prediction Sc: Nonsignificant correlations are expected between age and
hydrocephalus, renal disease, hepatic disease, seizures,
and LOC that follows traumatic head injury.

Prediction 6a: Depression scores and trait anxiety scores will not
correlate with age in this community-dwelling sample.

Prediction 6b: State anxiety scores will correlate positively with age.

 

Correlations with age for health and emotion variables are
presented in Table 8. In this sample, none of the health variables
were related to age at the stringent significance level applied (p =
.001), although DM approached significance (p = .005). Among the
emotional variables, the prediction was supported: State Anxiety alone

correlated with age, and positively (p < .001).

 

Prediction 1: When the variance associated with consolidation in
secondary memory is isolated, age will not explain a

significant portion of that variance.

 

For measures of association between a dependent variable and
multiple independent variables, multiple regression was used. Although
the careful stratification of the sample into discrete age groups lends

the data to analysis of variance (ANOVA) for age and sex predictions,

101
Table 8

Correletione With Age for Heelth and Emotional Factors, by Direction of

Expected Relationehin (Predictionee5 end 6).

 

 

 

 

Variable E 2
Factors Expected to Increase With Age
Hypertension .26 .01
Coronary Heart Disease .24 .01
Cerebrovascular Disease .26 .01
Diabetes Mellitus .28 .005
Pulmonary Disease .08 >.05
Reactive ("State") Anxiety .31 .001
Factors Expected to Decrease With Age
Current Alcohol Consumption .10 >.05
Current Tobacco Smoking -.16 >.05
Factors Expected to Be Unrelated to Age
Hydrocephalus --° --°
Renal Disease .15 >.05
Hepatic Disease .00 >.05
Seizures -.12 >.05
Loss of Consciousness Following Traumatic Head Injury -.08 >.05
Chronic ("Trait") Anxiety .15 >.05
Depression -.01 >.05

 

° Negative history for all subjects, thus coefficient was undefined.

102
the classification of age into a discrete variable (age group) would
sacrifice the natural variability on that variable. Furthermore,
multiple regression yields a Multiple R and 32 that are more
interpretable with regard to the effect size and with regard to the
amount of variance explained by the multivariate association,
respectively. Both R and 32 will be reported in each analysis, by the
same rationale described above.

Power associated with the regression analyses conducted here was
estimated beforehand in consideration of the sample size needed for this
study. For the largest possible analysis, which could involve 15
independent variables (Prediction 3, entering both attention variables
and health/anxiety variables), even very conservative power (.90) and
significance levels (.01) could be maintained with a sample size
exceeding 73 for an estimated effect size exceeding .35 (Cohen & Cohen,
1983, chapter 3).

For Prediction 1, performance on delay free recall trials imposes
the same demands as immediate free recall trials (perception, attention,
encoding/storage, and retrieval), with the added demand of consolidation
of the memory trace. Consequently, variance in memory performance
attributable to consolidation can be isolated by regressing delay recall
scores on immediate recall scores for the same stimuli. If age fails to
enter on the second step (i.e., if it fails to explain a significant
portion of the remaining variance), it can be inferred that age does not
affect consolidation.

In the auditory-verbal domain, Logical Memory II (delay free
recall) scores were regressed on Logical Memory I (immediate free

recall) scores in the first step. Then, age and age2 were entered

103
simultaneously and stepwise to assess whether a linear or curvilinear
(quadratic) relationship exists between age and consolidation of memory
traces (Cohen & Cohen, 1983, chapter 6). Immediate free recall entered
in the first step, as predicted, and explained 24 percent (Adjusted R2)
of the total variance in delay free recall for an effect size of .50
(Multiple R). Age and age2 failed to enter on the second step (pr >
.05), as predicted, indicating that secondary memory consolidation does
not change as function of age, either linearly or quadratically. The
regression weight and significance value for each term in the equation
are presented in Table 9 (R[1, 88] = 366.89, p < .0005).

In the visual-spatial domain, CFT Delay (delay free recall) scores
were regressed on CFT Immediate (immediate free recall) scores in the
first step. Then, age and age2 were entered simultaneously and
stepwise. CFT Immediate entered in the first step, as predicted, and
explained 88 percent of the variance in delay free recall (Adjusted R2)
for an effect size of .94 (Multiple R). Age and age2 failed to enter on
the second step (pr > .05), as predicted. The regression weight and

significance value for each term in the equation are presented in Table

9 (R[1, 88] = 612.18, p < .0005).

 

Prediction 2: Age will explain a significant portion of variance in

retrieval operations of secondary memory.

 

Performance on delay free recall trials imposes the same demands as
delay recognition trials (perception, attention, encoding/storage,
consolidation), with the added demand of unassisted retrieval.

Consequently, variance in memory performance attributable to retrieval

104
Table 9
Age Effecte in Coneelidation, by Dependent Variable (Prediction 1).

 

IUD

Step Variable SE3 pa Mult R Adj R2

 

Logical Memory 11

 

1 Logical Memory I 1.03 0.05 .0001 .495 .236
Constant - 4.17 1.48 .0061
CFT Delay
1 CFT Immediate 0.89 0.04 .0001 .937 .877
Constant 2.93 1.04 .0059

 

Note. A stringent alpha criterion was used for variable entry (pr 5

 

.001). R statistics for the total equations are reported in the text.
° Significance values reported in the table refer to the t-test for the

regression weight, R.

105
can be isolated by regressing delay recall scores on delay recognition
scores (first step) for the same stimuli. If age enters on the second
step, it can be inferred that age does affect retrieval processes in
secondary memory.

In the auditory-verbal domain, Logical Memory II (delay free
recall) was regressed on Logical Memory Recognition (delay recognition)
in the first step. Then, age and agez were entered stepwise. Logical
Memory Recognition explained 30 percent of the variance in Logical
Memory II, leaving 70 percent of the variance unexplained, presumably
representing retrieval ability and other unmeasured variance. Age
entered alone on the next step to explain an additional 10 percent of
the total variance and 14 percent of the unexplained variance. A total
of 40 percent of variance in Logical Memory II was explained by these
variables for an effect size of .64. The regression weight and
significance value for each term in the equation are presented in Table
10 (R[2, 87] = 30.83, p < .0005).

In a post hoc analysis, the potentially confounding health and
emotion variables, history of cerebrovascular disease (CVD) and State
Anxiety, were entered stepwise prior to age and age2 to determine the
extent to which the age effect reflects unhealthy aging rather than
healthy aging. State Anxiety failed to enter (p > .05), but CVD did
enter in the intermediate step between Logical Memory Recognition and
age to explain an additional eight percent of the total variance and 11
percent of the unexplained variance. Age entered even after CVD to
explain an additional seven percent of the total variance and 10 percent
more of the unexplained variance. The final solution accounted for 44

percent of the total variance in Logical Memory II for an effect size of

106
Table 10

Age Effects in Retrieval Before Controlling for Attention, by Dependent
Variable (Prediction 2).

 

Step Variable SE8 p Mult R Adj R2

Iw

 

Logical Memory II (Without Health/Anxiety)

1 Logical Memory Recognition 2.62 0.54 .0001 .556 .301
2 Age - 0.21 0.05 .0001 .644 .401
Constant 12.36 6.31 .0534

Logical Memory II (With Health/Anxiety)

 

1 Logical Memory Recognition 2.70 0.53 .0001 .544 .288
2 Cerebrovascular Disease -11.14 3.78 .0042 .620 .370
3 Age - 0.18 0.05 .0010 .677 .439
Constant 10.38 6.13 .0939
CFT Delay
1 CFT Recognition 1.20 0.13 .0001 .711 .500
Constant 7.60 2.20 .0009

 

Note. A stringent alpha criterion was used for variable entry (pr 5
.001). R statistics for the total equations are reported in the text.
° Significance values reported in the table refer to the t-test for the

regression weight, R.

107
.68. The regression weight and significance value for each term in the
equation are presented in Table 10 (R[2, 87] = 23.98, p < .0005).

In the visual-spatial domain, CFT Delay (free recall) was regressed
on CFT (Delay) Recognition in the first step. Then, age and age2 were
entered stepwise. CFT Recognition entered on the second step to explain
50 percent of the total variance, leaving 50 percent of the variance
unexplained, believed here to represent retrieval variance and other,
unmeasured variance. Age and age2 failed to enter on the next step (pr
= .03). The total effect size was .71; the equation weights and
significance values can be found in Table 10 (R[1, 86] = 86.56, p <
.0005).

 

Prediction 3: After controlling for the effects of attention, age will
explain significantly less variance in retrieval

operations of secondary memory.

 

The rationale and method for isolating retrieval variance was
described under Prediction 2. That method was repeated for Prediction
6, except that relevant attention measures were entered immediately
before age variables.

In the auditory-verbal domain, Logical Memory II (delay free
recall) was regressed on Logical Memory Recognition (delay recognition)
in the first step. Then, auditory-verbal measures of attention and
concentration were entered stepwise: Digit Span Backward,
Discriminative Letter Cancellation, and Phoneme Matching. Finally, age
and age2 were entered stepwise. Although there was a trend for Digit

Span Backward (pr = .005), none of the attention measures entered the

108
equation at the specified alpha, .001. Consequently, the results for
auditory-verbal retrieval were identical to those reported under
Prediction 2. In the visual-spatial domain there was no age effect in
the original analysis; consequently, there was no need to re-analyze the
data with attention variables.

Because attention variables did not explain any variance in
consolidation or retrieval, post hoc analyses were conducted to measure
the contributions of attention to total memory performance. In post hoc
analysis of auditory-verbal memory, Logical Memory I (LM-I) and II (LM-
II) were regressed on measures of auditory-verbal perception (SSPT) and
production (Letter Fluency) to isolate secondary memory ability. Then,
age and age2 were entered stepwise to determine the amount of variance
accounted for by these variables prior to controlling for attention.
Next, age and age2 were removed and auditory-verbal attention variables
were entered stepwise (Digit Span Backward, Discriminative Letter
Cancellation, and Phoneme Matching). Finally, age and age2 were entered
again, stepwise.

For LM-I, neither SSPT nor Letter Fluency entered on the first
step. Age2 entered on the second step to explain 22 percent of the
variance. After removing agez, Phoneme Matching entered alone. Age
entered on the last step to explain only 10 percent more variance after
controlling for attention (p < .0005). Similar findings emerged for LM-
II, except that attention played no role at all. Age2 entered initially
to explain 24 percent of the total variance. After removing agez,
Letter Fluency entered to explain 20 percent of the total variance. No

attention measures entered. Age entered again, to explain only 14

109
percent more variance beyond that which was accounted for by production
skills. The results of these analyses are available in Table 11.

In post hoc analysis of visual-spatial memory, CFT Immediate and
Delay trials were regressed on measures of visual-spatial perception
(CFT Matching, JOLO) and construction (CFT Copy, VR Copy) to isolate
secondary memory ability. Then, age and age2 were entered stepwise to
determine the amount of variance accounted for by these variables prior
to controlling for attention. For Immediate, JOLO and CFT Copy both
entered to explain 41 percent of the variance. For Delay, only CFT Copy
entered on the first step, explaining 32 percent of the variance (p <
.0005). Surprisingly, age and age2 failed to enter on the second step
for either Immediate (p = .01) or Delay (p =.05). Visual-spatial
attention variables were entered stepwise on the last step (VMS
Backward, Discriminative Symbol Cancellation, and Shape Matching).
Visual Memory Span Backward entered alone and for CFT Delay only,
explaining an additional 12 percent of that total variance (p < .0005).
Even when using liberal significance criteria (p = .05), the age effect
(age and age2 together explaining six percent of the total variance)
disappeared after controlling for attention. The results of these

analyses are available in Table 11.

 

Prediction 7: The only sex difference on the CFT will be on the delay

recall trial, where men will outperform women.

 

To test this prediction, each of the CFT trials was correlated with
sex. Sex was coded 0 for males and 1 for females, such that a negative

value is indicative of males exceeding females on the task. Sex

Table 11

Age Effectg in Encoding/Storege. by Dependent Varieble (Poet H00).

110

 

 

 

Step Variable R SEB p° Mult R Adj R2
Logical Memory I
1 Phoneme Matching - 1.71 0.43 .0001 .493 .234
2 Age - 0.18 0.05 .0004 .591 .334
Constant 40.25 2.60 .0001
Logical Memory 11
1 Fluency 0.12 0.33 .0004 .453 .196
2 Age - 0.24 0.05 .0001 .593 .336
Constant 24.12 5.06 .0001
CFT Immediate
1 JOLO 1.00 0.27 .0003 .589 .339
2 CFT Copy 0.59 0.17 .0009 .654 .414
CFT Delay
1 CFT Copy 0.63 0.14 .0001 .570 .317
2 VHS Backward 1.83 0.42 .0001 .670 .435
Constant -17.40 6.15 .0058

 

Note. A stringent alpha criterion was used for variable entry (pr 5

 

.001).

F statistics for the total equations are reported in the text.

° Significance values reported in the table refer to the t-test for the

regression weight, R.

111

explained a small but significant portion of variance on the Immediate

free recall trial (5 = -.30, 52 = .09, p = .001), but none of the other

correlations was significant at the specified alpha: Copy, 3 = -.20 (p
= .03); Delay (free recall), 3 = -.23 (p = .01); Recognition (delay
recognition), p = -.23 (p = .01); and Matching, p = -.14 (p > .05).

Therefore, in partial support of this prediction, no differences were
found for Copy, Recognition, or Matching trials. However, contrary to
the prediction, the sex difference that did appear was observed on the
Immediate free recall trial, not on Delay free recall. The direction of

the relationship (males outperforming females) was preserved.

 

Prediction 8a: The CFT Recognition Task will have homogeneous content,
as indicated by a strong Total Scale alpha reliability

and by moderate subscale alpha reliabilities.

 

On the CFT Recognition Trial, Cronbach’s alpha was computed for the
Total Scale, Global Element Scale, Detail Element Scale, as well as for
the Left Detail and the Right Detail subscales. These are presented in
Table 12. As expected, the Total Scale had a strong interitem
reliability (alpha=.84). The subscales ranged from moderate (.59, .61,
and .66) to high (.81), in rough relation to the number of items on each
scale; this was also consistent with the prediction. Each of the items
contributed substantially to the integrity of its host scale. None of
the items detracted from any of the alpha reliabilities except for Item
27 on the Left Detail subscale, and that alpha was reduced by less than
.02. Consequently, these data support the retention of all items on

their host scales except for Item 27 on the Left Detail subscale.

Table 12

112

Cronbach’s Alpha Reliebilitieeefor CFT Scales and Subscales, by Trial.

 

 

 

No. of
Scale Composite Items Items Alpha
Recognition Trial
Total 1-30 30 .84
Global Elements 1-4, 6-8 7 .61
Detail Elements 10-28 30 20 .81
Left Detail 10 11 14 15 18 20 23 24 27 9 .59
Right Detail 12 13 16 17 19 21 22 25 26 28 30 11 .66
Matching Trial
Total 1-10 10 .58
Left Detail 2 3 7 8 4 .40
Right Detail 4 5 9 10 4 .08

 

113

Intercorrelations among scales and subscales can be found in Table
13; all were significant (p < .0005). Correlations between the Total
Scale and each of other scales are, of course, very large because each
scale is nested within the Total Scale; correlations between the Detail
Scale and its nested subscales, the Left and Right Detail subscales, are
high for the same reason. More important is the corrected correlation
between the two unnested scales, Global and Detail (.75). Because this
relationship is only moderate, we can conclude that the two scales
measure unique variance and, thus, unique constructs. The corrected
correlations of the two Detail subscales with the Global scale (.65 and
, .77) are virtually identical to the Global-Detail correlation; in
addition, the corrected correlation between the two subscales is
virtually perfect (.99). These latter data indicate that the two Detail
subscales are parallel short-forms of the Detail Scale and that they

measure the same construct.

 

Prediction 8b: The CFT Recognition Total Score will have good
convergent validity, that is, will correlate positively

with the WMS-R Visual Reproductions recognition score.

 

Recognition correlated modestly with VB recognition (p = .40, p <
.0005). CFT Recognition also correlated moderately with VR-I and VR-II
(.50 and .57, respectively, p < .0005).

In addition, correlations between CFT Recognition and the other CFT
memory trials were computed. The Total Score on the Recognition Trial
correlated strongly and positively with both Immediate free recall (3 =

.74, p < .0005) and with Delay free recall (3 = .71, p < .0005).

114

Table 13

lntercorrelations Among CFT Scales and Subscales, by Trial.

 

Recognition Trial

GLO DET L-DET R-DET TOT
Global (GLO) .61 .75 .65 .77 1.02
Detail (DET) .53 .81 1.24 1.24 1.18
Left Detail (L-DET) .39 .86 .59 .99 1.15
Right Detail (R-DET) .49 .91 .62 .66 1.18
Total (TOT) .73 .97 .81 .88 .84

 

Matching Trial

L-DET R-DET TOT

Left Detail (L-DET) ‘ .40 3.07 .99
Total (TOT) .82 .92 .58

 

Nete. All values are significant (p < .0005). Values on the
diagonals are reliabilities (Cronbach’s alpha). The values above the
diagonals are corrected for attenuation; the values below the diagonal

are not corrected for attenuation.

115
Correlations with the nonmemory trials were much weaker (Copy, 3 = .38,

p < .0005; Matching, y = .18, p = .05).

 

Prediction 8c: The CFT Recognition Total Score will have good construct
validity in that it will load on a factor with other

visual-spatial secondary memory test scores.

 

For the confirmatory factor analysis, LISREL (Joreskog & Sorbom,
1989) was preferred over the SPSS (SPSS, Inc., 1990) Factor routine
because the former allows (and requires) the investigator to specify the
loadings of variables onto theorized factors; that is, it is truly
confirmatory of an a priori model. Also, the results of the LISREL
analysis provide goodness-of-fit indices, which summarize how well the
data fit the a priori model, and modification indices, which identify
individual variables that would fit better on other factors.

For this prediction, correlations among all of the measures listed
in Table 14 were entered as a matrix and analyzed in a multitrait-
multimodality model using the LISREL program for confirmatory factor
analysis. (The descriptives for each variable and the correlations
among all of the variables are tabulated in Tables 15 and 16,
respectively.) As shown in Table 14, each test is nested within one
trait (attention, perception, production, or memory) and within one
modality (auditory-verbal or visual-spatial). This organization
dictated the parameters for the Lambda X and Phi matrices; the initial
estimates for factor loadings were generated from a review of past
research with these same tests or with very similar tests. However, the

Phi matrix was not positive definite, resulting in premature termination

116

Table 14

A Rriori Fector Structure of Selected Measures, by Trait and Modality.

 

 

 

 

 

 

Auditory-Verbal Measures Visual-Spatial Measures
Perception
Speech-Sounds Perception Test Judgment of Line Orientation
CFT Matching
VR Matching
Attention
Digit Span Backward VMS Backward
Discriminative Letter-Cancellation Discriminative Symbol Cancellation
Phoneme Matching Shape Matching
Production
Letter Fluency ' CFT Copy
Category Fluency Visual Reproductions Copy
Memory
Logical Memory I Visual Reproductions Immediate
Logical Memory II Visual Reproductions Delay
Logical Memory Recognition Visual Reproductions Recognition
Cowboy, Immediate CFT Immediate
Cowboy, Delay CFT Delay

Cowboy, Recognition CFT Recognition

 

117

Table 15

Fector Analyse .

 

 

est Scores Used in

 

Descriptive Data for T

 

Older (p=43)

n=47)

Younger (

Total (R=90)

 

 

 

M.

M:

M:

Variable

 

3.5

5.0

1.9

2.4

3.1

3.7

SSPT

426
41

737
O O O
494

484

269
see

694

905
a e o
31

548
594

CFT, Match
VR, Match

JOLO

380
222

707
O. O
672

123
211

539
e e e
751

237
C O O
221

113

DS, Backward
Letter Canc.
Phoneme Matching

004
221

204
C C O
752

162
o m e
211

039
841

VMS, Backward
Shape Matching

Symbol Canc.

Letter Fluency
Category Fluency

CFT Copy
VB: COPY

dydo
680
I10
I88
IDR
w”,
myyy
a;
mmmcmc

......

VR, Immed
VR, Delay
VR, Recog
CFT, Immed
CFT, Delay
CFT, Recog

 

The Older group was missing two subjects on some measures.

Note.

118
Table 16

Correlations Among the Test Scores Used in Factor Analyse .

 

 

Test 1 2 3 4 5 6 7 8 9 10

1 SSPT

2 JOLO ~49

3 CFT, Match ~49 46

4 VR, Match ~34 32 39

5 DS, Backward ~36 43 37 23

6 Letter Canc. 48 ~47 ~43 ~51 ~39

7 Phoneme Matching 40 ~58 ~55 ~54 ~45 62

8 VMS, Backward ~47 60 29 33 46 ~46 ~39

9 Symbol Canc. 54 ~48 ~50 ~41 ~37 88 52 ~48

10 Shape Matching 63 ~60 ~54 ~41 ~49 58 68 ~52 62

11 Letter Fluency ~24 22 27 9 53 ~36 ~44 28 ~33 ~42-
12 Category Fluency ~37 46 37 28 52 ~47 ~50 46 ~47 ~53
13 CFT Copy ~50 60 65 40 36 ~46 ~55 44 ~55 ~62
14 VR, Copy ~56 49 52 51 42 ~53 ~66 39 ~55 ~65
15 LM I ~34 39 37 27 44 ~40 ~52 55 ~42 ~51
16 LM 11 ~35 37 41 28 43 ~39 ~46 56 ~42 ~47
17 LM Recog ~42 41 48 32 35 ~47 ~63 41 ~43 ~55
18 Cowboy, Inned ~18 32 19 21 23 ~11 ~27 30 ~17 ~31
19 Cowboy, Delay ~27 36 24 27 25 ~23 ~28 36 ~31 ~37
20 Cowboy, Recog ~ 8 11 19 29 5 ~ 2 ~21 10 ~ 2 ~16
21 VR, Immed ~51 60 56 43 47 ~52 ~59 53 ~57 ~61
22 VR, Delay ~48 50 47 38 54 ~56 ~62 49 ~55 ~56
23 VR, Recog ~40 48 37 28 44 ~47 ~46 47 ~41 ~61
24 CFT, Immed ~42 58 27 22 35 ~38 ~42 54 ~40 ~44
25 CFT, Delay ~45 55 29 16 34 ~35 ~38 57 ~39 ~44
26 CFT, Recog ~35 46 18 26 31 ~29 ~30 47 ~30 ~36

 

Note. Decimals were omitted.

Table 16 (cont’d)

119

 

 

 

 

 

Test 11 12 13 14 15 16 17 18 19 20
11 Letter Fluency

12 Category Fluency 59

13 CFT Copy 31 47

14 VR, Copy 38 46 63

15 LM I 40 41 34 49

16 LM II 38 44 39 45 90

17 LM Recog 33 38 39 46 60 56

18 Cowboy, Immed 22 22 17 36 51 52 42

19 Cowboy, Delay 30 27 19 43 55 55 43 92

20 Cowboy, Recog 10 0 6 23 29 35 19 56 57

21 V8, Immed 27 51 53 61 60 60 42 24 34 16
22 VR, Delay 42 60 49 58 58 62 41 24 27 17
23 VR, Recog 33 51 46 42 42 44 30 25 29 7
24 CFT, Immed 23 41 58 50 54 52 23 34 30 12
25 CFT, Delay 24 36 56 48 53 54 23 37 33 13
26 CFT, Recog 14 31 38 39 48 48 19 31 30 19
Note. Decimals were omitted.

Table 16 (cont’d)

Test 21 22 23 24 25 26

21 V8, Immed

22 VR, Delay 79

23 VR, Recog 65 63

24 CFT, Immed 58 62 48

25 CFT, Delay 59 61 46 93

26 CFT, Recog 50 57 40 74 71

 

Note.

 

Decimals were omitted.

120
of the analysis. This appears to have been due to high
intercorrelations among many of the tests and may reflect the influence
of a strong general factor.

SPSS factor analysis, by contrast to LISREL, is confirmatory only
inasmuch as it permits the investigator to specify the number of factors
in the final solution; the investigator cannot specify the constitution
of those factors in advance. This program classifies the variables into
a solution for the number of factors specified, but the variables may or
may not be organized around the principle intended. Because the data
did not conform to the assumptions of LISREL, they were subjected to the
SPSS Factor routine.

A principal factors analysis with varimax rotation of eight factors
(four traits in the auditory-verbal modality and four traits in the
visual-spatial modality) resulted in a solution (Table 17) that
explained an unusually large proportion of the variance (80 percent).
Factors were orthogonal, and the eigenvalues tapered off after the
fourth factor was extracted. The solution was mostly test-specific,
with trials of the same test coalescing onto a common factor. A "Visual
Memory" factor did emerge and explained a large portion of the total
variance; it consisted of the three CFT memory trials and Visual
Reproductions II. A second factor was much more heterogeneous with the
two alphabet tracking tasks (phoneme and shape matching), the two
copying tasks (CFT and Visual Reproductions Copy trials), and two of the
perceptual measures (SSPT and CFT Matching). The three trials of Cowboy
story comprised the third factor. The two fluency tasks and Digit Span
Backward loaded onto the fourth factor. The fifth and sixth factors

were Cancellation (Discriminative Letter and Symbol Cancellation tasks)

Table 17

Test Loadings on an Eight-Factor Solution Using SPSS Factor Apalysi .

121

 

 

 

Factors

Variable h2 1 2 3 4 5 6 7 8

VB, Delay 84 49 24 ~02 41 18 34 19 43
CFT, Immed 92 89 22 11 14 13 12 15 2
CFT, Delay 91 88 24 13 12 10 14 16 ~ 5
CFT, Recog 75 80 4 15 5 10 13 16 18
SSPT 61 ~24 ~52 ~ 7 - 7 ~36 ~ 8 ~36 2
CFT, Match 78 3 80 4 13 7 25 9 20
Phoneme Matching 74 ~13 ~54 ~11 ~33 ~30 ~28 ~ 7 ~36
Shape Matching 74 ~16 ~55 ~16 ~32 ~32 ~16 ~38 ~13
CFT Copy 81 40 75 0 19 20 1 11 4
VB, Copy 74 31 57 26 27 33 9 ~ 1 23
Cowboy, Immed 91 18 7 89 11 2 20 13 - 6
Cowboy, Delay 92 13 8 89 15 15 20 17 ~ 2
Cowboy, Recog 76 5 9 75 - 5 ~12 9 ~14 37
DS, Backward 65 13 19 5 64 5 20 37 11
Letter Fluency 87 5 15 15 87 15 14 ~ 8 ~ 8
Category Fluency 71 20 21 3 70 22 10 25 10
Letter Canc. 91 ~15 ~23 2 ~25 ~82 ~17 ~15 ~23
Symbol Canc. 86 ~20 ~32 ~ 2 ~20 ~79 ~17 ~16 ~ 8
LM I 91 35 13 30 22 15 77 11 9
LM II 89 37 13 31 23 10 74 12 12
LM Recog 79 ~ 9 45 25 14 27 62 16 ~ 3
JOLO 75 36 48 17 12 18 2 56 3
VMS, Backward 74 37 11 14 14 30 30 61 ~ 2
VR, Recog 71 30 18 3 37 7 12 55 36
VR, Immed 81 42 36 2 21 15 34 36 44
VR, Match 77 3 30 23 ~ 3 44 0 8 66
X Variance Explained -- 45 9 7 5 4 3 3 3
Eigenvalue "- 1107 204 109 104 009 009 008 007

 

Note. Decimals were omitted.

explanations of variable abbreviations.

h2 = communalities.

Please see text for

122

and Logical Memory (all three trials), respectively. Factor 7 consisted
of JOLO, Visual Memory Span, and Visual Reproductions Immediate trial.
Visual Reproductions II and the corresponding matching task made up the
last factor. The final solution explained 80 percent of the total
variance. However, only one factor (Visual Memory) accounted for the
majority (56 percent) of the explained variance. Furthermore, 17 of the
variables had considerably large loadings on multiple factors. The
factor loadings, eigenvalues, and communalities are in Table 17.

Specifying two factors for the two modalities generated a solution
that explained 54 percent of the variance; however, the test did not
divide solely on the modality principle. One factor was characterized
by tests of Verbal Learning (Logical Memory I and II and the three
Cowboy Story trials); the other consisted of all of the other tests and
could therefore be called a Broad Cognition or General Ability factor.
Factor loadings and communalities for this outcome are in Table 18.

Finally, specifying four factors yielded the most interpretable
solution, although it was not the four traits that were hypothesized.
Factor 1 consisted of all of the measures of perception, production, and
attention in both modalities, except Digit Span and Visual Memory Span.
Factor 2 was the Visual Learning factor, consisting of the three CFT
memory trials and the three Visual Reproductions memory trials. Factor
3 contained the two memory span tasks (Digit Span and Visual Memory
Span) and the two fluency tasks. Factor 4 was the Verbal Learning
factor, consisting of two of the three Logical Memory trials and the
three Cowboy Story trials. Logical Memory Recognition was also grouped
with Factor 4 variables because of the conceptual associations among

those variables, even though empirically this test loaded slightly

123

Table 18

Test Loadings on a Two-Factor Solution Using SPSS Factor Analysi .

 

 

Variable hz Factor 1 Factor 2
SSPT 47 ~68 ~10
JOLO 54 70 23
CFT, Match 43 65 10
VR, Match 26 49 14
DS Backward 46 60 31
Letter Canc. 61 ~78 ~00
Phoneme Matching 60 ~75 ~20
VMS Backward 50 62 35
Symbol Canc. 61 ~78 ~04
Shape Matching 67 ~79 ~21
Letter Fluency - 26 46 23
Category Fluency 47 67 14
CFT, Copy 59 77 06
VR, Copy 57 70 28
VR, Immed 67 78 26
VR, Delay 67 77 27
VR, Recog 48 ’ 67 19
CFT, Immed 51 61 37
CFT, Delay 50 58 40
CFT, Recog 39 47 41
LM I 68 5O 65
LM II 69 49 67
LM Recog 42 52 39
Cowboy, Immed 82 09 90
Cowboy, Delay 78 18 87
Cowboy, Recog 50 ~06 70
X Variance Explained ~~ 45 9
Eigenvalue ~~ 12.2 2.4

 

Note. Decimals were omitted. h2 = communalities.

explanations of the variable abbreviations.

Please see text for

124
better on Factors 1 and 3. Strong correlation with multiple factors was
characteristic of this last solution, observed in no fewer than 14
'variables. This model explained 66 percent of the total variance;

oommunalities and factor loadings are tabulated below (Table 19).

 

Prediction 8d: As further evidence of construct validity, each item of
the CFT Recognition Task will load on the a priori
scale for which it was designed (Global, Left Detail, or

Right Detail), except items 5, 9, and 29.

 

LISREL was again preferred over SPSS for confirmatory factor
analysis, for the reasons given above. For this prediction,
correlations among all the items of the CFT Recognition Trial, except
Items 5, 9, and 29 (detail elements that are not purely lateralized),
were entered as a matrix and analyzed in a multitrait model using the
LISREL program for confirmatory factor analysis. (The descriptives and
correlations among items are presented in Tables 20 and 21,
respectively.) Table 12 provides a list of the items designed for the
Global Scale and the Left and Right Detail subscales. This organization
dictated the parameters for the Lambda X and Phi matrices; the initial
estimates for factor loadings were generated from the pilot study.
However, the Phi matrix not positive definite, resulting in premature
termination of the analysis.

The same correlation matrix was subjected to a LISREL 2~factor
confirmatory analysis in which each variable was specified to load onto
the Global element factor or onto the Detail element factor. A chi-

square analysis was conducted to compare the observed correlational

125
Table 19

Test Loadings onaa Four-Fagtor Solution Uaing SPSS Factor Analysis.

 

 

 

Factors

Variable h2 1 2 3 4

SSPT 50 ~60 ~32 ~19 ~ 5
JOLO 57 51 49 23 12
VR, Match 61 73 5 ~12 25
CFT, Match 57 72 10 17 13
Phoneme Matching 68 ~71 ~16 ~36 ~18
Shape Matching 70 ~66 ~26 ~41 ~15
Letter Canc. 66 ~71 ~16 ~37 4
Symbol Canc. 65 ~69 ~22 ~36 1-
CFT Copy 64 66 43 16 ~ 2
VR, Copy 67 70 29 16 27
VR, Immed 70 55 53 30 13
VR, Delay 70 45 55 43 9
VR, Recog 52 36 46 42 3
CFT, Immed 87 23 88 13 13
CFT, Delay 85 21 87 13 16
CFT, Recog 73 13 82 7 19
DS, Backward 64 27 22 70 17
VMS, Backward 55 32 39 50 20
Letter Fluency 65 14 1 79 10
Category Fluency 60 36 24 64 0
LM I 71 23 39 49 52
LM II 72 22 4O 47 53
LM Recog 61 51 ~ 4 44 40
Cowboy, Immed 84 5 19 18 87
Cowboy, Delay 83 15 15 23 85
Cowboy, Recog 64 12 3 ~13 78
X Variance Explained ~~ 45 9 7 5
Eigenvalue ~~ 12.2 2.4 1.9 1.5

 

Note. Decimals were omitted. h2 = communalities. Please see text for

 

explanations of variable abbreviations.

Table 20

Descriptive Data for the CFT Recognition Items.

 

 

 

Item No. Mean SD Item No. Mean SD
1 .85 .36 16 .56 .50
2 .61 .49 17 .39 .49
3 .65 .48 18 .16 .37
4 .71 .46 19 .40 .49
5 .22 .41 20 .24 .43
6 .66 .48 21 .43 .50
7 .51 .50 22 .63 .49
8 .75 .44 23 .55 .50
9 .59 .49 24 .63 .49

10 .43 .50 25 .50 .50
11 .69 .46 26 .59 .49
12 .83 .38 27 .15 .36
13 .35 .48 28 .51 .50
14 .38 .49 29 .78 .41
15 .34 .48 30 .97 .18

 

127
Table 21

Correlations Among the CFT Recognition Iteaa. byaItea Number.

 

 

No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14

 

30 29
15 00 ~01
08 21 28

11 07 17 38 20

17 02 28 11 40 40

06 19 23 43 24 30 12

31 10 21 12 27 28 34 16
10 17 13 26 11 21 05 12 13 17
11 07 ~07 13 33 35 15 14 18 35 13
12 15 26 24 ~03 24 ~01 10 23 18 15 22
13 17 15 ~10 16 19 13 20 04 18 03 13 08
14 19 23 13 04 16 11 10 07 26 27 16 29 12
15 03 03 13 10 26 ~04 22 08 06 00 22 13 02 19
16 34 ~14 06 27 30 23 36 12 23 13 20 20 23 27
17 26 ~14 15 10 15 13 12 03 23 02 22 ~01 10 16
18 09 03 26 01 ~08 05 ~01 04 05 12 ~05 11 00 18
l9 14 22 40 02 19 29 14 15 11 ~01 04 24 18 19
20 16 ~10 19 07 03 12 12 08 03 05 14 04 03 06
21 30 13 35 06 16 05 21 03 21 12 18 21 08 13
22 21 06 17 06 18 19 28 04 31 11 30 09 28 16
23 ~06 ~12 ~00 ~10 20 16 34 ~11 22 01 04 07 05 14
24 21 25 26 22 24 24 37 15 31 25 40 34 28 31
25 16 09 17 10 08 24 25 11 14 18 22 15 07 12
26 04 10 06 ~13 ~01 ~01 ll 05 20 07 10 36 08 21
27 ~01 ~13 11 ~08 09 03 09 09 02 03 ~00 02 ~04 07
28 30 20 28 06 35 16 18 07 44 35 38 28 20 24
29 09 21 36 02 14 03 21 14 24 12 25 28 10 18
30 10 ~15 ~14 02 10 ~14 07 04 10 16 15 ~09 01 15

madam-hawks
H
:-

 

Note. Decimals were omitted.

 

Table 21 (cont’d)

128

 

 

 

No. 15 16 17 18 19 20 21 22 23 24 25 26 27 28
16 21

17 12 24

18 15 14 10

19 _ 35 21 17 22

20 16 23 05 56 20

21 10 13 02 12 04 10

22 26 25 13 14 20 27 30

23 13 20 02 09 04 08 10 14

24 21 21 08 21 10 27 34 32 19

25 05 21 19 12 02 19 18 12 18 26

26 01 05 O9 17 11 03 31 17 12 26 18

27 ~10 ~08 06 17 ~08 ~01 22 ~01 06 06 16 22

28 03 13 17 11 10 12 30 32 25 37 07 16 09

29 20 14 13 15 14 23 35 28 13 34 19 18 06 26
30 14 08 15 08 ~23 11 16 24 ~05 ~02 06 10 08 19
Note. Decimals were omitted.

Table 21 (cont’d)

 

 

 

No. 29 30
30 05
Note. Decimals were omitted.

 

129
matrix to the hypothetical matrix that was generated from the
theoretical model that was specified. The Adjusted Goodness of Fit
index was modest (.71), and the root mean square residual was large
(.09). The chi-square was statistically significant (x2[404, fl=88] =
499.91, p = .001), indicating that the observed matrix differed from the
correlation matrix that was reproduced from estimates of the
theoretically relevant parameters. Reassignment of items with
significant modification indices failed to change the summary statistics
for the confirmatory factor analysis; that is, the indicated
modifications of the model did not improve the overall fit of the model
to the data. Factor loadings and communalities are listed in Table 22.
The correlation between the two factors was .795.

As an alternative, SPSS was used. Specification of three factors
resulted in an uninterpretable clustering of the test items; the
solution explained only 32 percent of the total variance. A subsequent
analysis designating two factors (Global and Detail) explained less
variance (25 percent) and produced an uninterpretable solution. Post
hoc exploratory factor analysis (leaving the number of factors
unspecified, to be determined by an eigenvalue criterion of 1.0)
resulted in an 11~factor solution that explained 66.8 percent of the
variance, but the factor structure did not conform to any identifiable

organizational scheme.

 

Prediction 8e: CFT Recognition Total Scores will distribute normally

with no ceiling or floor effects.

 

130
Table 22

Recognition Item Loadinga on Global and Detail Factors Using LISRE .

 

 

Item No. h2 Factor 1 Factor 2
1 18 42 00
2 7 26 00
3 27 52 00
4 12 34 00
6 24 49 00
7 31 56 00
8 15 39 00
5 23 00 48
9 31 00 55

10 - 12 00 34
11 26 00 51
12 19 00 43
13 10 00 31
14 19 00 43
15 10 00 32
16 20 ‘ 00 45
17 8 00 28
18 7 00 26
19 10 00 32
20 10 00 32
21 21 00 46
22 27 00 52
23 8 00 28
24 43 00 66
25 - 12 00 35
26 10 00 32
27 1 00 11
28 34 00 59
29 23 00 48
30 2 00 16

 

Note. Decimals were omitted. h2 = communalities.

 

131

Scores on the CFT Recognition Trial distributed normally, as
predicted. For the 88 subjects who completed this task, the mean and
median were 16.0, and the mode was 15. Scores ranged from 4 to 27 (out
of 30 possible), with a standard deviation of 5.74. Performance on this
task correlated significantly with age (2 = ~.41, p < .0005);
consequently, descriptives are tabulated for younger and older subjects
in Table 23. The relationship between sex and Recognition Total Score
(r = ~.23) failed to reach significance at the level adopted in this

study (p = .05, two-tailed), as did the Age x Sex interaction.

 

Prediction 8f: Most items on the CFT Recognition Task will have

significant, positive corrected item-total correlations.

 

Point-biserial correlations between each CFT Recognition item and
the CFT Recognition Total Score were corrected by partialling out the
variance in the Total Score that was due to the item itself (Table 24).
All coefficients were positive. Only two values failed to reach
statistical significance (gs = .09 and .12; p > .05); the others ranged
from .18 to .60, with 17 of the 30 exceeding .31 to meet the stringent
alpha criterion (p 5 .001). Item difficulty indices (percent of the
sample that responded incorrectly) ranged from 3% to 85% and distributed
roughly normally, with 20 items in the 30%~to~70% range and 4 items and
6 items in the upper and lower tails, respectively; the mean item

difficulty for the 30 items was .465.

132
Table 23

ancriptivea for CFT Recognition (Raw Scorea). by Scale

 

Group n Mean Median Mode SD SEM Min. Max.

 

Total Scale

Total 88 16.0 16.0 15.0 5.7 2.3 4 27
Younger 47 17.9 19.0 21.0 5.5 2.2 5 27
Older 41 14.0 15.0 15.0 5.3 2.1 4 25
Men 37 17.6 18.0 15.0 5.5 2.2 4 27
Women 51 14.9 15.0 ' 11.0 5.7 2.3 5 27

Younger
Men 21 20.0 21.0 24.0 4.5 1.8 11 27
Women 26 16.1 17.5 18.0 5 7 2.3 5 27

Older
Men 16 14.4 15.0 10.0 5.0 2.0 4 23
Women 25 13.7 14.0 11.0 5.6 2.2 6 25

 

Note. Only the age effect was significant at the alpha designated for

 

theoretical comparisons (p g .001). However, age and sex effects were
significant at the more liberal alpha (p g .05). Consequently, data for

all groups are presented.

133

Table 23 (cont’d)

 

Group n Mean Median Mode SD SEM Min. Max.

 

Global Scale ‘

Total 88 4.7 5.0 6.0 1.8 1.1 0 7

 

Detail Scale 9

Total 88 11.3 11.5 12.0 4.6 2.0 3 20
Younger 47 13.1 14.0 15.0 4.3 1.9 4 20
Older 41 9.2 9.0 12.0 4.1 1.8 3 13
Men 37 12.6 12.0 12.0 4.4 1.9 3 20
Women 51 10.3 10.0 5 O 4.6 2.0 3 20

Younger
Men 21 14.9 15.0 17.0 3.5 1.5 7 20
Women 26 11.6 12.0 15.0 4.4 1.9 4 20

Older
Men 16 9.6 10.0 8.0 3.6 1.6 3 16
Women 25 9.0 9.0 5.0 4.5 2.0 3 18

 

‘ There were no age or sex differences on the Global Scale (p > .05).
h Only the age effect was significant at p 5 .001. Age and sex effects

were significant at p 5 .05.

Table 23 (cont’d)

134

 

 

 

 

Group n Mean Median Mode SD SEM Min. Max.
Left Detail Subscale ‘ .

Total 88 3.6 4.0 4.0 2.0 1.3 0 8
Younger 47 4.3 4.0 5.0 1.9 1.2 1 8
Older 41 2.7 3.0 1.0 1.7 1.1 0 6
Men 37 4.3 4.0 5.0 1.9 1.2 0 8
Women 51 3.0 3.0 1.0 1.9 1.2 0 7

Younger
Men 21 5.1 5.0 5.0 1.7 1.1 1 8
Women 26 3.7 3.5 3.0 1.8 1.2 1 7

Older
Men 16 3.2 3.0 3.0 1.6 1.0 0 6
Women 25 2.4 2.0 1.0 1.8 1.2 0 6

Right Detail Subscale b

Total 88 6.1 6.0 9.0 2.4 1.4 1 11

Younger 47 7.0 7.0 9.0 2.3 1.3 2 11

Older 41 5.2 4.0 4.0 2.3 1.3 1 9

‘ Only the age effect was significant at p _ .001. Age and sex effects

were significant at p 5 .05.

h Only age differences were significant (p _

.001 and .05).

Table 24

135

Discrimination Ability of the CFTafiecognition Iteaa.

 

 

Diffic. Item- Diffic. Item-
Item No. Index'l Total”. Item No. Index‘ Totalb

’1 15 .38 '*** 16 44 .43 '*'*

2 39 .18 ' 17 61 .26 *'

3 35 .43 "" 18 84 .27 "'

4 29 .24 " 19 60 .33 ""

5 78 .45 "" 20 76 .29 "'

6 34 .35 “" 21 57 .41 ""

7 49 .47 "" 22 37 .46""'

8 25 .28 "' 23 45 .22 '

9 41 .49 "'* 24 37 .60 ""
10 57 .29 "' 25 50 .35 ""
11 31 .42 "" 26 41 .27 "'
12 17 .39 "" 27 85 .09
13 65 .26 " 28 49 .51 ""
14 62 .40 "" 29 22 .44 ""
15 66 .28 " 30 3 .12

 

‘ Difficulty Index is the percent of the sample answering incorrectly.

b Item-Total is the partial correlation between the item score and the

Total Score, controlling for the overlap between the two.

‘ p < .05.

" p < .01.

“‘ p < .005.

"" p < .001.

136

 

Prediction 9a: The CFT Matching Task will have homogeneous content, as
indicated by a strong Total Scale alpha reliability and

by moderate subscale alpha reliabilities.

 

On the CFT Matching Trial, Cronbach’s alpha was computed for the
Total Scale, as well as for the Left Detail and the Right Detail
subscales. These are presented in Table 12. The Total Scale had a
modest interitem reliability (alpha = .58), suppressed in part by three
items with zero variance (all subjects answered three items correctly).
None of the remaining items on the Total Scale detracted from the alpha
reliability coefficient, indicating that all items contribute to the
integrity of the scale.

The Right Detail subscale also had modest internal consistency
despite the low number of items to its design and one item with zero
variance (alpha = .40, 3 items). Item-total statistics on this scale
indicate that Item 5 substantially reduced the subscale’s reliability;
deleting this item would raise the subscale alpha by .07.

Like the Right Detail subscale, the Left Detail subscale started
with few items and was further limited by one item with zero variance;
unlike the Right Detail subscale, however, this subscale failed to
achieve predicted levels of homogeneity (alpha = .08, 3 items). In
addition, one item detracted from the reliability coefficient by .08.

Intercorrelations among scales and subscales can be found in Table
13; all were significant (p < .0005). Correlations between the Total
Scale and each of other scales are, of course, very large because each

subscale is nested within the Total Scale. The correlation between the

137
two unnested subscales, Left Detail and Right Detail, was r = .55 (p

(.0005).

 

Prediction 9b: The CFT Matching Total Score will have good convergent
validity, that is, will correlate positively with the
WMS-R Visual Reproductions matching score and with the

Judgment of Line Orientation test score.

 

Matching Total Score correlated modestly with the VR Matching Total
Score (5 = .39, p < .0005) and with JOLO (g = .46, p < .0005). The
modesty of the relationship was expected given the limited number of
items on the Visual Reproductions recognition task and given the
specific nature of the JOLO stimuli (line orientations) compared to the
complex spatial relationships involved on the CFT.

In addition, correlations between CFT Matching and the other CFT
trials were computed. Correlations between the Matching Total Score and
the CFT memory trials failed to reach significance at the designated
probability (p g .001): Immediate free recall, g = .27, p < .005; Delay
free recall, g = .29, p < .005; and Recognition, ; = .18, p = .05. The
correlations with the other nonmemory trial were much stronger (Copy, 5

= .65, p < .0005).

 

Prediction 90: The CFT Matching Total Score will have good construct
validity in that it will load on a factor with other
visual-spatial perception tests, and not with auditory-

verbal perception tests or with memory tests.

 

138

Correlations among all of the measures listed in Table 14 were
entered as a matrix and analyzed in a multitrait-multimodality model
using the LISREL program for confirmatory factor analysis, as described
in Prediction 8c. However, as reported in that same section, the Phi
matrix was not positive definite, thus resulting in premature
termination of the analysis.

Because the data did not conform to the assumptions of LISREL, they
were subjected to the SPSS Factor routine with varimax rotation, which
organizes the variables into the solution that maximizes the variance
explained. The results of those three analyses are reported under

Prediction 8c.

 

Prediction 9d: As further evidence of construct validity, each item of
the CFT Matching Task will load on the a priori scale
for which it was designed (Left Detail or Right Detail),

except items 1 and 6.

 

For this prediction, correlations were computed among all the
detail items of the CFT Matching Trial (thus excluding the global,
nonlateralized items, 1 and 6). However, two of the remaining eight
items had zero variance so that correlations could not be computed
between them and the other items. Consequently, neither LISREL nor SPSS

could analyze the data.

 

Prediction 9e: CFT Matching Total Scores will be negatively skewed

with a prominent ceiling effect.

 

139

Scores on the OFT Matching Trial were negatively skewed, as
predicted. For the 88 subjects who completed this task, the mean was
9.4 and the median and mode were 10.0. Scores ranged from 6 to 10 (out
of 10 possible), with a standard deviation of 1.0. -Regression of the
Matching Total Score on age and sex failed to identify an effect for age
(3 = ~.25, p = .01), age2 (3 = ~.27, p = .01), or sex (3 = ~.14, p >
.05) at the significance level specified; the interactions also failed

to enter the equation (p > .05). Descriptives can be found in Table 25.

 

Prediction 9f: Most items on the OFT Matching Task will have

significant, positive corrected item-total correlations.

 

Point-biserial correlations between each CFT Matching item and the
CFT Matching Total Score were corrected by partialling out the variance
in the Total Score that was due to the item itself (Table 26). Three
coefficients could not be computed due to zero variance. The other
seven coefficients were positive and moderate, given the ceiling on
point-biserial correlations at .80. One item correlated .24 (p (.01)
with the Total Score; the others ranged from .28 to .41 (p < .005).
Item difficulty indices (percent of the sample that responded

incorrectly) ranged from 0% to 15X.

140
Table 25

ancriptivea for CFT Matching Total Scale (Raw Scorea).

 

 

Group n Mean Median Mode SD SEM Min. Max.
Total 88 9.4 10.0 10.0 1.0 .65 6 10
Younger 47 9.6 10.0 10.0 0.8 .52 7 10
Older 41 9.3 10.0 10.0 1.2 .78 6 10

 

Note. Only age differences were significant (p g .05).

141
Table 26

Discrimination Ability of the CFTyMatchingflteaa.

 

 

Diffic. Item- Diffic. Item-
Item No. Index‘ Totalb Item No. Indexa Totalb
1 1 .28“* 6 0 ~~°
2 2 .24" 7 1 .28’*'
3 0 ~~° 8 18 .37....
4 6 .30*“ 9 14 .41‘*'*
5 15 .38'**' 10 0 ~~°

 

3 Difficulty Index is the percent of the sample answering incorrectly.

b Item-Total is the partial correlation between the item score and the
Total Score, controlling for the overlap between them.

° Correlation could not be computed due to zero variance.

‘ p ( .05. " p < .01. *** p < .005. "*‘ p < .001.

DISCUSSION

- This project (1) introduced recognition and matching trials to
supplement an existing (free-recall) measure of visual memory, the
Complex Figure Test (CFT); (2) provided psychometric data and
preliminary norms for the new instruments, using an age- and sex-
stratified sample of healthy, community-dwelling adults; (3) evaluated
the role of attention in secondary memory processes in healthy older
adults; (4) and compared men and women on CFT performance. The findings
are summarized and discussed relative to each prediction and then the
findings are integrated within each of three sections: age differences,

sex differences, and psychometric properties.

Age Differences

Prediction 4a: Eight of the 10 tests supported the prediction that
attentional skill and perceptual accuracy would correlate negatively
with age in both auditory-verbal and visual-spatial modalities. Two
measures did not correlate with age: Digit Span Backwards and CFT
Matching.

The absence of age differences on these two measures may reflect a
sampling bias. Older adults in the community who were experiencing more
decline than their peers perhaps were too self-conscious about their
cognitive difficulties to participate in this study. If this is so,

then these results could indicate that the eight measures that were

142

143
significantly related to age are sensitive enough to detect these
changes even in the healthiest of volunteers.

Prediction 4b: The prediction that production ability would
correlate negatively with age in the auditory-verbal modality but not in
the visual-spatial modality got mixed support. One measure in each
modality (Category Fluency and Visual Reproductions [VR] Copy)
correlated significantly, whereas another measure in each modality
(Letter Fluency and CFT Copy) did not. These differences are difficult
to reconcile. As in the case of the previous prediction, this
inconsistency with prior research findings may reflect the different
sensitivities of these measures to aging that are brought out in a
healthy sample.

Prediction 5: The frequency cf hypertension (HTN), coronary heart
disease (CHD), cerebrovascular disease (CVD), diabetes mellitus (DM),
and pulmonary disease will correlate positively with age; current
alcohol and tobacco use will correlate negatively with age; and
nonsignificant correlations are expected between age and hydrocephalus,
renal disease, hepatic disease, seizures, and loss of consciousness
(LOC) that follows traumatic head injury. In this sample, none of the
health variables were related to age at the stringent significance level
applied, although DM approached significance. The failure to find a
relationship between age and those variables that were expected to
increase with age may again be due to sample bias; that is, the older
subjects in this study may represent an especially healthy subset of
community dwelling older adults.

To investigate the possibility that this sample may not be very

representative of the population at large, frequencies of these five

144
medical conditions were compared to national statistics. Older adults
in this sample less frequently reported HTN (40% versus 60%) and CHD
(25% versus 50%); they reported DM at the same rate as the older
American population at large (18%, within the reported range of 15-20%);
and they reported a higher incidence of CVD (10% versus 1~5%). (The
estimated rate of pulmonary disease in the general population could not
be ascertained.) The expected age-related decline in alcohol use and
smoking was also absent in this sample. This may be due to lower rates
of alcohol and tobacco use in the younger subset of this actively
religious sample.

Because the high incidence of CVD in this sample could potentially
confound analyses of the effects of normal aging on cognition, analyses
that produced evidence of age effeCts were run again, covarying for
history of CVD. This procedure was required in only one situation
(Prediction 2, Logical Memory analyses), and CVD failed to supplant age
or dramatically reduce its explanatory potential.

The findings presented thus far warrant caution in generalizing
from the rest of the results: Some impressions gained from this study
may reflect very healthy aging only, especially given that the one
blemish in the older adults in this sample (the high incidence of CVD)
was statistically corrected in the aging analyses.

Prediction 6: The prediction that depression and trait anxiety
would not correlate with age in this community-dwelling sample but that
state anxiety would correlate positively was fully supported. Reactive
anxiety during testing was more evident among older adults than among
younger adults. Because the high levels of anxiety might have

confounded analyses of the effects of normal aging on cognition,

145
analyses that produced evidence of age effects were run again, covarying
for State Anxiety. This procedure was required in only one situation
(Prediction 2, Logical Memory analyses), and State Anxiety failed to
explain any variance and did not diminish the age effect.

Prediction 1: The prediction was confirmed that age would not
explain a significant portion of consolidation variance. Auditory-
verbal and visual-spatial secondary memory consolidation did not change
as a function of age, either linearly or quadratically. This suggests
that healthy aging does not affect the automated consolidation processes
of secondary memory. Therefore, the decline observed with aging is more
likely due to effortful encoding/storage and/or retrieval processes.
Encoding/storage could not be isolated by the paradigm used in this
study; retrieval, however, was examined in the next prediction.

Prediction 2: The prediction that age would explain a significant
portion of variance in retrieval operations of secondary memory was
firmly supported in the auditory-verbal domain. Age explained 14
percent of the retrieval variance in a linear, not quadratic,
relationship. Although a positive history of cerebrovascular disease
(CVD) was also associated with decline in retrieval operations, this was
not responsible for the observed age effect. Age continued to explain
10 percent of the retrieval variance even after controlling for CVD.
Consequently, age was found to play a small but significant role in the
decline of effortful, retrieval processes for auditory-verbal material.
For visual-spatial material, however, age was not significantly related
to retrieval efficiency. This is especially surprising given Craik’s
(1977) proposal that visual-spatial memory processing is even more

effortful than is auditory-verbal and given Winograd and Simon’s (1980)

146
empirical support for that proposition. This finding in the present
study is difficult to reconcile with the attention-deficit model of age-
related memory decline that is under investigation here.

Prediction 3: The data failed to support the prediction that age
would explain significantly less variance in retrieval operations of
secondary memory after controlling for the effects of attention.
Examination of the effects of attention on retrieval efficiency were
relevant in the auditory-verbal modality only because age differences
were not observed in the visual-spatial modality (Prediction 2).
Although there was a trend for Digit Span Backward, none of the
attention measures explained a significant portion of retrieval variance
at the specified alpha. Consequently, the age effects observed in
auditory-verbal retrieval were not attributable to attention.

Taken together, the results from the analyses in Prediction 2 and
in Prediction 3 indicate that decline in attention capacity is not
responsible for age-related declines in retrieval functions. This now
puts the "anomalous" visual-spatial results from Prediction 2 in a
clearer perspective, at least with regard to the attention-deficit
model: It seems that effortful, attention-consuming processes are not
required for retrieval from secondary memory; consequently, differences
between auditory-verbal and visual-spatial effortful attention would not
appear in analyses of these functions.

Earlier evidence with older adults had already implicated attention
in encoding/storage processes (Craik & Byrd, 1982; Rabinowitz, Craik, &
Ackerman, 1982). Some investigators had postulated effects of attention
on retrieval processes, as well (Craik & Byrd, 1982). Others, however,

presented substantial and convincing evidence that retrieval processes

147
are spared by attentional decline and that only encoding/storage
processes are affected (Perlmutter and Mitchell, 1982; Smith, 1977,
1980). The results of this study are consistent with the latter
conclusion.

Although this study did not permit systematic isolation of
encoding/storage processes, the results thus far allow us to make some
inferences regarding these processes now. That is, because attention is
not required in consolidation and retrieval, it seems, by default, to
exert its influence in the first stage of secondary memory,
encoding/storage. Evidence for this from other sources has already been
presented (and cited in the previous paragraph). The current study,
however, goes beyond previous studies in allowing us to determine the
proportion of variance in encoding/storage that is attributable to age,
apart from attention. Having concluded that any attention effects in
secondary memory performance will reflect encoding/storage processing,
and not retrieval or consolidation, regression of immediate or delay
free recall on attention prior to age provided an index of the role of
attention in encoding/storage processing.

In post hoc analysis of visual-spatial memory, Visual Memory Span
Backward explained 17 percent of the total variance that was not
explained by perception and production variables (p < .0005). Even when
using liberal significance criteria (p = .05), the age effect (age and
age2 together explaining six percent of the total variance) disappeared
after controlling for attention.

In post hoc analysis of auditory-verbal memory, age entered on the
last step to explain only 10 percent more variance in LM-I after

controlling for attention (p < .0005). This is identical to the

148
relationship observed when retrieval alone was regressed on age;
consequently, this appears to be the same age-retrieval effect observed
earlier and does not represent new age differences in encoding/storage.
Therefore, in the auditory—verbal domain, age effects persisted even
after controlling for attention, but the effects represented age
differences in retrieval only; age explained none of the
encoding/storage variance.

Similar findings emerged for LM-II, except that attention played
no role at all. Age explained only 14 percent more variance beyond that
accounted for by production skills. Again, this amount is no different
from the retrieval variance already attributed to aging; age explained
little, if any, of the encoding/storage variance.

Summary. The expanded tests of auditory-verbal and visual-spatial
memory, together with many measures of perception, construction, and
attention, helped determine whether reduced attention capacity explains
age-related secondary memory decline. On auditory-verbal memory, no age
effect was observed on consolidation; encoding/storage varied by age,
but the effect was due entirely to attention deficits; only retrieval
showed age differences that could not be explained by attention (10
percent of total variance). On visual-spatial memory, no age effects
were observed in encoding/storage, consolidation, or retrieval after
perception and construction were controlled.

These findings have implications for memory intervention with
older adults. First, efforts to train older adults in mnemonics will
likely have only short-term effects during the training workshop where
attention is focused solely on memorization of the stimuli. However,

mnemonic devices typically demand substantial attention capacity and,

149
therefore, are not likely to generalize outside of memory training
workshops. Consistent with this are reports that older adults do not
spontaneously use the mnemonics they learn (Perlmutter & Mitchell, 1982;
Smith, 1980).

Second, visual learning appears to be very resistant to aging
processes and does not require excessive attentional resources.
Therefore, visual encoding of verbal material, for example, by means of
reading, should reduce the age effects on memory performance. This was
supported in a recent study by Denburg, Fastenau, and Fertuck (1994),
which showed significantly weaker age effects for paragraphs that were
read compared to paragraphs that were aurally presented. Even more
impressive is the fact that the reading material was incidentally
learned (i.e., the subjects were tOld that it was a test of reading and
were not told to remember the material) whereas the aural material was
presented in an intentional paradigm (i.e., subjects were told to
memorize the paragraphs before hearing them). Consequently, memory in
older adults may be best sharpened by increasing attention capacity
rather than by increasing attention demand through mnemonics, and by
encouraging visual encoding (e.g., using maps to supplement verbal
directions and writing/reading auditory material that needs to be
remembered).

Two limitations to this research warrant mention. The first, as
already noted, is that the sample appears to be healthier and more
educated than their cohorts in the population at large; consequently,
these results may not generalize readily to the all adults.

Furthermore, this sample was almost exclusively Caucasian.

150
A second limitation is that this study is cross-sectional in
design. Consequently, age effects represent cohort differences and not
necessarily the aging process itself. Plans are underway to seek
funding for a five-year follow-up study of this sample for cohort~
sequential analysis, which will allow true aging effects to be

distinguished from cohort and time-of-measurement effects.

Sex Differences

Prediction 7: The prediction that men would outperform women on
the delay recall trial of the OFT was not supported. No sex differences-
were found for Copy, Delay Recall, Recognition, or Matching trials. An
unexpected finding was a sex difference on the Immediate Recall trial,
accounting for nine percent of the variance on this measure. Although
the effect was observed on Immediate recall rather than on Delay, the
direction of the relationship (men outperforming women) was preserved.
This adds to a collection of findings that show modest sex effect in
favor of males on at least one of the CFT trials. This lends support to
the proposition that males, as a group, possess some capacity for
processing complex visual stimuli more efficiently and/or more
effectively. However, there was substantial overlap on all trials
between men and women in this study, thereby further supporting Harris’
(1985) argument for describing sex differences in visual-spatial

functions as sex-related variability and not as sexual dimorphisms.

s chometric Pro erties of CFT Reco ition M tchin
Prediction 8a: It was predicted that the CFT Recognition trial

would have homogeneous content. As expected, the Total Scale had a

151
strong interitem reliability (alpha = .84). The subscales ranged from
moderate (.59, .61, and .66) to high (.81), in rough relation to the
number of items on each scale; this was also consistent with the
prediction. Each item contributed substantially to the integrity of its
host scale. The correlation between the Global and Detail scales was
only moderate, suggesting that each scale measures some unique variance.
Left and Right Detail subscales, on the other hand, appeared to be no
more than "alternate forms" for measuring the exact same construct.

Prediction 8b: The prediction that CFT Recognition Total Score
would have good convergent validity was supported. CFT Recognition
correlated modestly but positively with VR recognition. The modesty of
the relationship was expected given the limited number of items on the
VR recognition task and given the simplicity of the VR stimuli. As
further evidence of convergent validity, CFT Recognition correlated
moderately with VR~I and VR~II and highly with the other CFT memory
trials; as evidence of divergent validity, CFT Recognition correlated
weakly with nonmemory (Copy and Matching) trials. It is evident from
these data that the Recognition Trial measures secondary memory and that
it is not confounded by construction or perceptual demands.

Prediction 80: Several factor analyses lent support to the
construct validity of the OFT Recognition trial. On eight-factor and
four-factor solutions, the CFT Recognition Total Score loaded onto a
"Visual Memory" factor; on a two-factor model, the Recognition score
loaded onto a "Broad Cognition" factor in contrast to the "Verbal
Memory" factor. Thus, all three models provided strong and consistent

support that the CFT Recognition trial is a measure of visual-spatial

152
secondary memory and that it is different from auditory-verbal measures
of secondary memory.

Prediction 8d: Contrary to the other prediction of construct
validity for the Recognition Trial, the a priori three-factor structure
(Global, Left Detail, and Right Detail) was not validated with this
healthy sample. Even after combining the two Detail subscales, the
correlation matrix failed to match the theoretical two-factor model
(Global and Detail scales).

Prediction 8e: As expected, CFT Recognition Total Scores
distributed normally with no ceiling or floor effects. For the 88
subjects who completed this task, the mean and median were 16.0, and the
mode was 15. Scores ranged from 4 to 27 (out of 30 possible).
Performance on this task correlated significantly with age, so
descriptives were tabulated for younger and older subjects.

Prediction 8f: As predicted, most items on the OFT Recognition
Task had significant, positive corrected item-total correlations. All
coefficients were positive, and only two values failed to reach
statistical significance. Item difficulty indices reflected that the
test samples a broad range of ability levels so that there are items
within virtually everyone’s capability and items that challenge even
people with very good memory skills; yet, the set of items converges on
an item difficulty of .50 where discrimination is maximized. This test
discriminates very effectively at all levels of competence and should be
equally useful with healthy and impaired persons alike.

Prediction 9a: The prediction that the CFT Matching Task would
have homogeneous content, as indicated by a strong Total Scale alpha

reliability and by moderate subscale alpha reliabilities, was supported

153
for the most part. The Total Scale had a modest interitem reliability
(alpha = .58), suppressed in part by three items with zero variance (all
subjects answered three items correctly). All of the remaining items
contributed to the integrity of the scale.

The Right Detail subscale also had modest internal consistency
despite its low number of items and one item with zero variance (alpha =
.40, 3 items); one of the three items substantially reduced the
subscale’s reliability. Like the Right Detail subscale, the Left Detail
subscale started with few items and was further limited by one item with
zero variance; unlike the Right Detail subscale, however, this subscale -
failed to achieve predicted levels of homogeneity (alpha = .08, 3
items). In addition, one of the items detracted substantially from the
reliability coefficient.

The correlation between the two unnested subscales, Left Detail and
Right Detail was moderate (g = .55), which may indicate that the two
subscales measure unique variance. These data offer some support for
the retention of the subscales but the evidence is relatively weak;
these subscales will need to be administered to a clinical sample, where
more variability can be expected, in order to examine their
reliabilities more fully. The Total Scale reliability has been
demonstrated here and this scale can be considered adequate for further
use and appropriate for clinical trials.

Prediction 9b: There was modest support for convergent validity in
the CFT Matching trial. Matching Total Score correlated modestly with
the VR Matching Total Score (3 = .39) and with JOLO (g = 46). The
modesty of the relationship was expected given the limited number of

items on the VR recognition task and given the specific nature of the

154
JOLO stimuli (line orientations) compared to the complex spatial
relationships involved on the CFT.

In addition, correlations between CFT Matching and the other CFT
trials supported that the Matching Trial is distinct from the secondary
memory trials. A high relationship between visual-spatial construction
was observed, which is conceptually meaningful because of the
substantial contribution of visual-spatial perception to copying
performance.

Prediction 9c: The prediction that the OFT Matching Total Score
would have good construct validity received some support. On eight-
factor and four-factor solutions, the CFT Matching Total Score
consistently loaded apart from memory tests. Thus, there was some
support that the OFT Matching trial measures nonmemory cognitive
functioning; however, the evidence failed to confirm this as a pure
visual-spatial task or as a pure perceptual measure.

Prediction 9d: As further evidence of construct validity, it was
predicted that each Matching item would load on the a priori scale for
which it was designed (Left Detail or Right Detail). Two of the eight
items had zero variance, such that correlations could not be computed
between them and the other items. Consequently, neither LISREL nor SPSS
factor analysis could analyze the data.

Prediction 9e: As predicted, scores on the OFT Matching Trial were
negatively skewed. For the 88 subjects who completed this task, the
mean was 9.4 and the median and mode were 10.0. Scores ranged from 6 to
10 (out of 10 possible). The Matching Total Score failed to correlate
with age and with sex. Descriptives were provided on the total sample

and for relevant subgroups.

155

Prediction 9f: As expected, most items on the CFT Matching Task
had significant, positive corrected item-total correlations. Three
point-biserial correlation coefficients could not be computed due to
zero variance. The other seven coefficients were positive and moderate,
given the ceiling on point-biserial correlations at .80. One item
correlated .24 with the Total Score; the others ranged from .28 to .41.
Item difficulty indices ranged from 0% to 15%. Unlike the Recognition
Task but similar to other tests of its kind (e.g., JOLO), the Matching
Trial was very easy for most healthy adults and did not discriminate
among these individuals very well. Data comparing clinical samples with
healthy samples and comparing patients with different types or degrees
of impairment will be necessary to demonstrate the utility of this
measure for discriminating perceptual performance in lower ranges of
ability.

Summary. The new CFT Recognition Trial was found to be reliable
and valid with this healthy sample. Psychometric evidence on the
Matching Trial was much less impressive, mostly due to the small number
of items on the measure. The ceiling effect on the Matching Trial also
limited the psychometric characteristics, especially with 30 percent of
the items producing zero-variance. Nonetheless, there are some
indicators that the Matching Trial and its two subscales may perform
better among clinical samples. Clinical trials for both the Recognition
Task and for the Matching Task are indicated for their further
deve10pment. Data are currently being collected at Henry Ford Hospital
in Detroit, Michigan, with epilepsy patients undergoing unilateral
temporal lobectomies; this is the first step toward clinical validation

with the new instruments.

Appendix A

Appendix A

Recognition Item§:for the Coaplex Figure Tea;

156

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

158

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

160

 

 

 

 

 

 

 

 

 

 

 

 

 

162

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

163

 

 

 

 

 

 

 

 

F—N

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

164

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

165

 

 

 

 

 

 

 

 

 

 

 

 

 

 

166

(O

> > > > >

_I_ .7

 

 

168

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I2.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

170

H

 

172

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[6

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

173

I7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ha,—

 

 

 

 

 

 

 

 

 

 

174

IS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

176

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ZO

178

22

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

179

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

23

 

181

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

25

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

182

2.6

> \/ > > >

 

k \4 > \m \w

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

186

30

 

Appendix B

Appendix 8

Matching Iteagafor the Coaplex FigureaTeag

187

 

189

 

 

LJ

 

 

 

 

190

 

 

 

 

 

 

191

 

 

 

 

 

192

 

 

 

 

 

 

193

M-é

 

 

 

 

 

 

195

M-B

 

196

 

 

 

 

 

 

197

M-lO

 

 

 

 

 

 

Appendix C

Appendix C

Neuropsychological Survey

198

Neuropsychological Questionnaire Subject Number: (Rec #3)

[Directions: Ask these questions before testing & record all responses.
When recording "Year" values, values up through 5~months~31~days are
truncated; values 2 6~mos~0~days are rounded up to next yr. Anything
in parentheses should be communicated to the subject to clarify the
question. Bracketed info is for the examiner only.]

[Universal Codes: "9/99/999" for no respgnse, refusal to answer;
maximum value is "8/89/899"; "0/00/000" if "Np";
"1/01/001" is the min. value for affirmative answer.]

 

"BEFORE WE BEGIN TESTING, I HAVE SOME QUESTIONS TO ASK YOU. SOME OF
THESE MAY BE DIFFICULT TO ANSWER. ANSWER TO THE BEST OF YOUR ABILITY."

[CIRCLE SUMECT’S GENDER] ......COOOOOOOOOO......OCOOOOOOO”[1] F[2]

What is your date Of birth? 0900000000000... __ - -

How many years of school have you completed? .....................

When did you last complete a class for a

numerical grade (not S/U, Pass/Fail, or Audit)? ... __ __~01~__ __ __ __
Can you remember the WORST time that you hit your head or

suffered a blow to your head? [If no, "000" for next 2 qus]

 

Were you dizzy afterwards? How long? [days; "000" if "No"] __

Were you unconscious? ............................ Yes [1] No [0]

How long? [minutes; "00" if "No"]

How many times have you been knocked unconscious? ...................

Have you ever had a stroke or a transient ischemic attack
(sometimes they’re called TIA’s or "mini-strokes")? ... Yes [1] No [0]

How many times has this happened? ...........................

Following the (worst) stroke, were you hospitalized?
[If yes] How long? [in days; "00" if "No"] ..................

 

Following the (worst) stroke, did you have rehabilitation?
[If yes] How long? [in weeks; "00" if "No"] .................

 

Have you ever had hydrocephalus (too much fluid
in the brain)? ........................................ Yes [1] No [0]

How many times have you had this? [If "No," "0"] ...............

Neuropsychological Questionnaire 2 Subject Number: (Rec #3)

Have you ever "blacked out" or lost consciousness without
knOWing Why? coco-000000000000000000000000000.0000... Yes [1] NO [0]

How long did it last? [In minutes; code "000" if "No"]

Have you ever been treated for seizures or
seizure prevention? .................................. Yes [1] No [0]

How many years were you in treatment? [If "No," "00"] ....... __

Do you know what caused the seizures? [See if this revises others]

Have you ever taken anticonvulsant medications, like Dilantin,
Tegratol, Phenobarbital, Depakene, or Depakote? 00 [No]

[If Yes] How many years did you take that? ................. __ __

What was that for?

Have you ever had numbness or tingling in your arms, legs,
or face? (Can you tell me more about that? Do you know why?) 00 [No]

How many times has this happened? [Exclude "falling asleep"]

Have you ever had spells of sudden weakness or paralysis in
your arms, legs, or face? (Can you tell me more about that?) 00 [No]

How many times has this happened? [Exclude "falling asleep"] __
Have you ever seen a neurologist for a health problem? Yes [1] No [0]

What was that for? [See if this revises other responses.]

Have you ever been tested with EEG, CT of the head, or
MRI 0f the head? IO.......OOOIOOOOOOOOOOOI.0.0.0.000... Yes [1] No [0]

What was that for? [See if this revises other responses.]

Have you ever had neurosurgery? ....................... Yes [1] No [0]
What was that for? [See if this revises other responses.]

Have you ever had high blood pressure? ............... Yes [1] No [0]
How many years was it a problem? ............................ __ __

Are you currently on any medications for that? ... Yes [1] No [0]

Was it under control (back to normal) when you last
had it CheCked? ......COOOOOOOOOOOOOO00.000.00.00. Yes [1] No [0]

Neuropsychological Questionnaire 3 Subject Number: (Rec

Have you ever had angina (chest pains)? ............... Yes [1] No

Have you ever had a heart attack, heart disease, or
any other heart problems? ............................ Yes [1] No

Have you ever had a medical procedure on your heart or
arteries, such as angioplasty, endartarectomy, surgery,
orapacelaker? OOOOOOOOOOOOOOOOOO......OOOOOOOOOOOI. Yes [1] No
Have you ever had diabetes? .......................... Yes [1] No

How many years was it a problem? ............................

Are you currently on any medications for that? ... Yes [1] No

Was it under control (back to normal) when you last
11“ it CheCked? ....O..0.........OOOOOOOOOOOOOOOOO Yes [1] No

Have you ever smoked cigarettes? [If No, "00" for next 4 questions.]

At resent, how many packs do you smoke PER WEEK? ...........

How many years have you smoked at this level
(total no. of years, not necessarily in succession)? ........

In your heaviest year, how many packs did you smoke PER WEEK? __

How many years did you smoke at that level
(total no. of years, not necessarily in succession)? ........

On average, how many alcoholic drinks‘ do you have PER WEEK?

[0]
[0]

[0]

[0]

' Ask type/amt: 12 oz beer (verify size of "can” or ”bottle”)

1.5 oz liquor (a "shot" or a "mixed drink”)
oz wine
3 oz port

II II ll
9!

How many years have you been drinking at this level
(total no. of years, not necessarily in succession)?

During your heaviest year of drinking, how many
alcoholic drinks‘I did you have PER WEEK?

About how many years you drink at that level
(total no. of years, not necessarily in succession)?

Neuropsychological Questionnaire 4 Subject Number: (Rec #4)

Have you ever used "street drugs"? .................. Yes [1] No [0]
What substances have you used?
How many years did you use (have you used) street drugs? ....

Haye you ever had any problems with dependency or addiction
to prescription medication? ............................ Yes [1] No [0]

[If yes] What medications have been a problem for you?

How many years was this (has this been) a problem? .......... __
Have you ever had problems with your kidneys? ......... Yes [1] No [0]
Have you ever had problems with your liver? ......... Yes [1] No [0]

Have you ever had problems with your lungs? ......... Yes [1] No [0]

Can you list all the medications you have taken
in the last week: [Record number of prescribed medications] .....

 

Name of Medication Dosage Freguegcy [Rx Codesl

MEM COG ANX DEP MTR

 

MEM COG ANX DEP MTR

 

MEM COG ANX DEP MTR

 

MEM COG ANX DEP MTR

 

MEM COG ANX DEP MTR

 

MEM COG ANX DEP MTR

 

MEM COG ANX DEP MTR

 

MEM COG ANX DEP MTR

 

MEM COG ANX DEP MTR

 

SUMS . . . MEM: __
COG: __
ANX: __
DEP: __
MTR: __

Appendix D

Appendix D

Letter Cancellation Stiaali

203

204

LC-l

"Vacation"

In an antique store, we spotted a doll house that I
bought for Jaime. The saleswoman said she’d ship my other
purchases to Washington. I thanked her, and we got back
into the car.

FINISHED

205

LC-2

It was also Ian who introduced me to my first oak-
apples, pulling the tree branch low enough for me to touch
them; and Ian who guided my fingers around the beautifully
formed nest of a jenny wren.

FINISHED

206

LC-3

Whenever I have an appetite for fish, I go to a village on
the lakeshore where it is served fresh off of the whaleboat.
Lately, since my retirement, I have made the lakeward trip
every weekend. I drive straight to Whitefish Bay and walk
the wharf to an alehouse on the lakefront. My timetable has
become so noticeable that the tradesmen pull out their
timepieces and smile as I pass or even greet me with a
mateworthy wave. Though nameless among them, I am not
anonymous. When I arrive at the tavern, the native cooks
prepare a tasteful solefish to mark my "homecoming. " They
affectionately call me the "homeless paleface," yet they keep
a place for me there among them.

FINISHED

207

LC-4

What began as a harebrained idea became the Batemans’
vacation cottage outside of Stoneford. Taking some time off
to attend a free real estate venture, the hopeful couple spent
nine days in the Whiteer mountains. While they were
guests in Stoneford, the salesman took them on horseback
through timeless forests and ancient graveyards. At the end
of a ninety minute ride, two large redwoods formed a
gateway to a slopeside terrace on the edge of the gorge. In
the foreground was a lovely dome-topped rotunda, modeled
from dwellings in early frontier settlements. As a timeshare,
the saleprice was more-than-manageable, and the Batemans
beamed with excitement when they signed the contract.

FINISHED

208

LC-S

There have never been any parents happier than DWight
and Harriet Hodges. Dwight and Harriet had never been
able to have children, and the health specialists had told them
it was impossible. However, this couple in their forties
proved the experts wrong as they hurried to the hospital on
that very wonderful birthday. How happy Harriet was to
finally see this day come. And Dwight, overprotective of his
mother-to-be, helped her with as much as he could.
Obsessively, he attended to every detail of the hospital trip
much as they had rehearsed it, even to the point of hailing
three hacks!

FINISHED

209

LC-6

I hastily rolled up my rawhide this morning to get a
headstart in the cool of the day. But alresdy by mine, the fog
burned off, unharnessing the heat of a searing sun. Hiking
the mountain is hard enough in pleasant weather. But now,
in the thick air of a hot and humid day like today, it seems
I can’t inhale deeply enough to ever catch a healthy breath.
' Nonetheless, I embrace these hardships because of a passion
I hold for the wilderness. I find my home here with wild
horses, in otherwise uninhabited hillsides. Here are the
treasures that Nature withholds; she keeps them here in trust
for the brotherhood of cowhands and in h0pe for humankind.

FINISHED

Appendix E

Appendix E

Symbol Cancellation Stimuli

210

p 211

SC-l

KEY

l—Cl'z'll-Il'l—ll>l+l)l-A

 

 

l~'-|-ll) ||-[>l+lr'l-I|'-ll-l+|-'-T'-‘l)l(J

 

'L>L--l+I--lI—I>Ir-l-é-1cI+T~I-II>1)II'I

 

l‘-'I)l+l-'-ll-I+l)l-ll(l-'-|'-—|C Il'll-l>l

 

I-II--'IcI>JrI~IcI>IéI+n-l-Hr-1>la -
[c L—IJ-rl c1:—1>|~1rm>1~lc|>1c1~l

 

212

Rln-Il—er-IDHIDJM

 

 

IC I-lj'fl ([I-|>l'-'II"ICJ>I?ICI>| CL?!
lFl>JC l?l-ll>l|-7l"lc l'-7>l-'-Jl"ll-L)l

w—v—V"

 

 

[Fl-41+] D | C Il-lg+ll"[)l-i|?|-'-il-Jl"l+]

 

L-=-lI-I-+lc l_>lr'l_c__l-II>I+I-=-I DWI—INF]
Lil-H)H—l>1+ll‘l-Il~l+-I+I-=-l~l )lfl

 

 

'L>|-°-I+T-=-ll-I>Ir'l+|C|+l=-l-Il>l)lfl

 

bl) l+lél|-|+I)l'f'|(I‘LF'Ile-ll-l>l

 

l:H--lcl>|r'l'-|Cl>Jél+J+-l-I|r'l) In
m—n-rlclsl>l~|rlcl>1~lcl>lcl~7
Lr-I>IcI~I-II>I+-1r-Ic1~1>I-°-I-r-|H>I .

 

 

 

.[Fl-41+l)ICll-|+|l'|)|-ll'-l-'-ll-|l'l+7

 

LrlrI-ILC I>Iflc[-Il>l+l+f>lI-l>lr'—l
El-ll)ll-l>[+lr‘I-l|'-'H-|+|~I'-I DIG

 

 

L>J~'-[+['-"li-l>|r'l-'—|Cl+J"-'l-l|>l)lfl

 

LCl-flfi'lCll-|>|'-'|f'lCl>['-|Cl>l(|?7

. 213

SC~3
KEY

Fifi'll-ll'I-H>J+I)F°-J

 

 

lC l-fl?l (Il-|>J'-II‘RI>I'='ICI>LCI'-'l
[l']>LCL'T'|-ll>ll-ll"lC l"-'l>l-'-ll'll-L)]

w

 

 

ll'l-l|+|) [C ll-|_:l_-*Lr_'l)l-|[-.-T.-..[|—|rl+j

 

bll'l-IIC |>IFICl-ll>l+l-'-|)ll-[>ll'7
Lil-ll)II-l>l+|l'l-il'-'|l-l+l-'-l?l)lC7

 

 

L>l~l-l7['-'ll-I>Jf'[-'-IC|+I?J-ll>l)lf'7

 

bl) l+l-'-ll-J+l)I-ll([-'-I?ICIt'll-lg

 

bl~l Cl>ll‘|'-IC|>l-°—l+lI-l-I|r‘])[fl
u I-H~I cn-I>I--Ir-ICI>I--ICI>Ic1--7

 

 

Lrl>1< l--J-Il>l+-Jr'lc l~|>l-=-J'r'II-I fl

 

[£I-H+IDIC l|-l+[l"l)l-l|-'[-'—|l-ll"l-fl

 

L-rlr'l-llc I>IrIcJ-II>I+I+I>II-1>Ifl
L-l—IDlk—l>l+l|‘l-Ilrll-l+lé|--l)lc_l

 

 

L>J+l+l'-ll-l>|r‘|-'~IC|+J'-'l-H>l)lfl
LC l-l|'-l CJl-l>l'='lr'ICl>[?|(l>|<|-Tl

 

 

Appendix F

Appendix F

Scoring Criteria for the CFT Drawinga

214

215

SCORING CRITERIA FOR THE COMPLEX FIGURE TEST
(Fastenau, 1991; rev. April, 1993)

General Principles

1.

ANY ELEMENT CAN ONLY BE SOORED ONCE. That is, it can only serve as
part of one unit. In times of doubt, it should be scored as the
unit which brings the most points for the examinee.

For each scorable unit (numbered items below; 18 total) award 1
point for each criterion met (maximum of 3 points per unit).

For all measurements, because the lines are wide, measure from the
approx. centers of the lines. Give the benefit of the doubt to the
examinee-~credit points when the measurement is borderline.

A metric ruler will be the easiest standard for scoring linear
criteria because most criteria are expressed as "110%."

A protractor is needed. Use endpoints of lines to measure angles.

For placement scoring, if the reference line (e.g., #1’s right side
in the phrase "+ #l’s right side") is not present, then use the
imaginary line as a reference point. For form scoring, all the
lines within the scoring unit must be present for credit.

Criteria

RECOGNIZABLE: Some element can be identified as an attempt at the

scoring unit in question. (Be lenient and credit as
recognizable if there is any resemblance; however, each
element can only be scored as one unit.)

GOOD FORM: Angles are accurate, lines intersect, shape/spacing is

proportional, & the number of sides/lines/dots is exact.
Form angles (v. placement angles) are those between
parts of the same unit.

Nppber of SideaZLinealDots Is Exact.

Unit # 1 Four sides (including R side) must be present.

Unit # 4. Two diagonals are present, even if not corner-to-corner.
Unit # 5: There must be only 4 lines, not 5.

Unit # 6' Four sides (including L side) must be present.

Two diagonals are present, even if not corner-to-corner.
Unit # 8: Stem of the cross that attaches to #1 must be present.
Unit #11: There are 3 dots (no credit for circles or lines).
Unit #12: There must be exactly 5 hash marks.
Unit #17: One and only one diagonal must be present in box.
Unit #18: Stem of the cross that attaches to #1 must be present.

216

Angles Are Within 15° of the Angle in the Stimulus.

L Special Angles:

Unit # 9: 28-62-90
Unit #13: 48-48-84
Unit #16: 55-125-55-125

L Right angles: 90 degrees

Linea Interaagt.

Using the crossing line as the standard, no gap/overhang > 10%.
Linea Are Straight.

Using the imaginary line between endpoints as the standard, no
point on the actual line deviates by > 10%.

Shapes and Spacing Are Proportional.

Ratios are within the specified limits.

To compute the ratio, divide the 1st measurement by the 2nd one; if
the resulting number falls within the boundaries of the criterion,
give credit. For example, in #1, compute the ratio by dividing the
length of the longest horizontal line by the length of the shortest
vertical line; credit any dividend between 1.0 and 2.0.

R Critical Ratios:

Unit # 1 Longest (L) side to shortest (8) side, L:S < 2
Unit # 5: Widest (W) space to narrowest (N) space, W:N < 2.
Unit # 6: Longest (L) side to shortest (S) side, L:S < 2
Unit # 8: Long (L) bar to shorter (S) crossbar, 3 < L:S < 7.
Unit #11: Longest (L) diameter to shortest (S) diam., L:S < 1.5.
Unit #12: Widest (W) space to narrowest (N) space, W:N < 2.
Longest (L) hash to shortest (S) hash, L:S < 2.
Unit #16: R vertical of #1 (V) to length of diamond (d), V:d > 2.
Unit #17: Longest (L) side to shortest (S) side, L:S < 1.5.
Bottom of #1 (B) to shortest (S) side of square,
3 < B:S < 5.
Unit #18: Long (L) bar to shorter (S) crossbar, 3 < L:S < 7.

GOOD PLACEMENT: Element is within proper boundaries, is properly
oriented to at least one major referent, and
intersects/bisects all other elements at proper
locations. "Placement" angles (v. "form" angles) are
those between one unit and other units.

217

Parallel: Not skewed more than 15°.

Horizontal: Parallel to at least one of the following
horizontal referents: top or bottom of base
rectangle or horizontal midline.

Vertical: Parallel to any of the following vertical
referents: L or R side of base rectangle or the
vertical midline.

+ Bisecting: Using the line biaected as the standard,
intersection is within 10% of the midpoint.
Element must bisect at all relevant points (e.g.,
#2 must bisect both top and bottom of #1).

 

Intersecting: Using the crossing line as the standard, no
gap/overhang > 10%.

Unit # 8: Connecting piece must meet #1 above horizontal midline
and below the top of #1.

Unit #11: Circle must fall in correct octant, although some
leniency is permitted if diagonal is missing.

The codes down the left margin are also recorded on the scoring sheet as
prompts to the experienced examiner. Less-experienced examiners should
take special care to consult the descriptions above, and inexperienced
examiners should begin with the detailed, item-by-item description of
scoring.

Appendix G

Appendix G

EMS-R VR Recognition/Matching Stipuli

218

 

 

 

 

 

 

221

DE
DE

 

 

 

 

 

 

 

 

 

 

 

 

DE
DE

 

 

 

 

 

 

 

 

 

 

 

 

 

222

 

 

 

__.IU

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

wayww

 

Appendix H

Appendix H

Cowboy Story Stimuli and Scoring

224

225
SCORING FOR THE "COWBOY" STORY
(Fastenau, 1991; modeled after Wechsler, 1987)

Words expressed verbatim but in mixed up order are credited, as long as
each word is only credited to the phrase with which it belongs, and each
word is credited to only one phrase. Examples:

"who went from AZ to S.F." = 2 verb pts;

"He had Lia pet gpg with him" is credited 1 verb for "with his

dog"; however, "left his dog with a friend" does not receive credit

for "with his dog" because the word with is clearly belongs to the
unit "at a friend’s."

 

If there are extra words beyond the verbatim, the verbatim phrase still
receives verbatim credit. For example, "Dressed very finely (1 verbatim
pt), he went back there (1 V pt) pp get the dpg (1 V pt)."

Verbatim words must be exact: same tense, number, and person (e.g., no
verbatim credit for "whistles" or "howled").

For CONTENT credit, any form of the critical words/phrases (see scoring
criteria below) is credited. E.g., "howled" receives content credit
even though it does not receive verbatim credit.

' Special CONTENT Rule: "To the dog"

"Dog" can be part of any of the three phrases that follow it:
E.g., "whistled (1 content pt) to the dog (1 more content pt),"
"called (1 C pt) the dog (1 more C pt)," or "patted (1 C pt) the
dog (1 C pt)." Of course, "dog" receives credit only once even if
it is used in two or three of those phrases.

However, "dog" canppp be credited twice if used as part of "But the
dog would have nothing to do with him." E.g., "He went back (V and
C pt) but the dog (no credit) didn’t recognize him (1 C pt)."

 

 

 

 

Examples of 1~pt Examples of O-pt
Text General Rule Content Resppnses Content Resppnses
A cowboy "Cowboy" is ~ —— man, guy, person,
required cowhand
from AZ "AZ" and an who lived in AZ; visiting AZ; went
indication that was in AZ to AZ; Phoenix;
he lived there the desert/plains

went to S.F. "San Francisco" took a trip to S.F.; CA, San Diego, the
and indication went to the Bay city; from S.F.;
that he traveled City; visited S.F. who was/lived in SF

with his dog

which he
left

at a
friend’s

while he
purchased

a new suit
of clothes.

Dressed
finely,

he went
back

to the
dog!.

whistled to
him,

called him
by name,

and patted
him.

But the dog
would have
nothing to
do with

him

Indication that
the dog accom-
panied him on

a trip

Indication that
he voluntarily
parted w/ the dog

Indication of a
friend’s home

Indication that
he was buying
something

clothing

a no suggestion
that it’s a type
he already had

Indication that
he looked more
formal in his
new clothes

Indication of
a return trip

dog (in any
context)

"whistled" or
form of the word

Indication he
vocalized in a
beckoning gesture

Indication he
touched dog in
a friendly way

Indication dog
did not recog-
nize him; must
include "dog"
or a pronoun
refering to dog

226

with a puppy, taking

his dog, carrying
his mutt along

dropped off, gave,
tied up

with a buddy, where
his friends lived

bought, went to the

store for, picked up,
shopped for; got (if
coupled with "new" or
with idea of buying)

some clothes; a new
outfit; a new suit

Dressed up; Looking
fine; With a dapper
look; In his sharp
suit

returned, came back,

went to his friend’s

doggie; puppy; St-

Bernard, collie, etc.

whistling, whistle

yelled to him,
hollered his name,
called to him

petted, stroked,
rubbed

it ran away, he

backed away, the dog

ignored him, it did
not notice the cow-
boy, he snubbed his

pet, cat, horse;
to see his dog;
to be with his dog

which ran away,
which was taken,
which died

neighbor’s, ranch,
parent’s, a kennel

retrieved, ordered,
unpacked, looked
at, priced

In his new clothes,
Dressed differently

he went home, he
left the store

pet, cat, horse

yelled at him,
screamed at him,
grabbed his collar

grabbed, hit,
slapped

snapped at him,
bit him, jumped
on him, chased him

master, the dog wouldn’t come

in his new
hat

and coat,

but gave a
mournful

howl.

Coaxing was
of no
effect;

so the
cowboy
went away

and donned
his old
garments

whereupon
the dog

immediately

showed his
wild joy

on seeing
his master

as he
thought
he ought
to be

"hat" (in any
context)

"C0812"

Indication that
the dog was sad,

missed his master

"howl" (in any
context)

Indication that
the cowboy tried
to persuade dog

Indication that
the cowboy left
the dog’s sight

Indication that
he put on his
former clothes

dog (in any
context)

Indication of
dramatic change

Indication of
pleasure (either
by the dog or
by the cowboy)

[Indication that

the cowboy was
seen by the dog

Indication that
the dog recog-
nized the
cowboy again

jacket
pining, bewailing,

groaning, sorrowful

cry, whined

He couldn’t convince;

Nothing worked to
show the dog; urging

left, went inside,

took off

changed back,
got into his old
clothes

dossie; puppy; St-

Bernard, collie, etc.

suddenly, quickly,
as soon as, now

became ecstatic, was
excited, was happy

when the cowboy
showed up, when his
master appeared, how
his master looked

like he should be,
back to his old
self, in familiar
garb, recognized
him, knew him

outfit, clothes

outfit, clothes

angry, resentful,
threatening

bark, growl,
whimper, moan,
groan

took off his coat,
returned his
clothes; put on new
clothes; some other
clothes

pet, cat, horse

began to, started
to, eventually

came to him, was
friendly, was back
to normal, recog-
nized him, welcomed
his master warmly

on knowing him,
seeing the clothes
in the house

in old clothes,
looking better;
came to him

Appendix I

Appendix I

Confirmation Letter to Participating Organizations

228

229
DEVELOPMENTAL MEMORY STUDY

The MSU Psychological Clinic is looking for volunteers for memory
testing. Your contribution to this study would be invaluable to us. In
this project, we are trying to better understand normal memory processes
in adults of all ages. We are also trying to improve our current memory
tests, and the results that you provide will help us do that.

In this study, you will do many different things. Many of the
tests measure memory directly. There are also some non-memory tests
that will help us examine functions that are sometimes related to memory
test performance, such as the ability to read, write, hear, and draw.
The exam will take approximately two and a half hours; in addition, you
will be asked some questions over the phone, and we will have a brief
questionnaire at the beginning of testing.

All information that you provide us will be strictly confidential.
After your exam, your test results will be filed under a code, without
using your name or any biographical information. These records will be
maintained in a secured file. This exam is for research purposes, so no
written report will be issued. However, if your test results suggest a
possible problem, we may recommend that you see your primary physician.

Participation in this study is voluntary, and you will be free to
withdraw from the project at any time. In addition, you may refuse to
do a particular task, and we will continue with the next one.
Certainly, the more that you are able to complete, the more your
participation will assist us in our research.

In exchange for your participation, we will make a contribution of
$10 to your church or to a fund your church designates. IF ALL 20 SIGN-
UP SPACES ARE FILLED AND THESE VOLUNTEERS COMPLETE TESTING, A BONUS OF
$50 WILL BE PAID in addition to the $10 per person, FOR A TOTAL DONATION
OF $250.

Please indicate your interest by providing your name and phone
number on the attached sign-up form. You will be contacted by one of
the research coordinators who will answer any questions you might have
and who will schedule you for testing. We will try to make the
arrangements as convenient as possible.

WE REGRET THAT, FOR THIS PROJECT, WE ARE UNABLE TO ACCOMODATE
INDIVIDUALS WHO HAVE (1) UNCORRECTABLE VISUAL IMPAIRMENT, (2)
UNCORRECTABLE HEARING IMPAIRMENT, (3) OR SUBSTANTIAL SPEECH IMPAIRMENT.
IN ADDITION, WE MUST LIMIT PARTICIPATION TO PERSONS WHO ARE ABLE TO
WRITE/DRAW WITH THEIR PREFERRED HAND.

This research project has been reviewed and approved by the MSU
University Committee on Research Involving Human Subjects. It is
conducted under the supervision of Norman Abeles, PhD, ABPP. For more
information. contact the Project Director, Phil Fastenau, M.A., at 355-
9564.

Appendix J

Appendix J

Conaapt Fora

230

231

DEVELOPMENTAL MEMORY STUDY

Thank you for volunteering for memory testing at the MSU
Psychological Clinic! Your contribution to this study will be
invaluable to us. In this project, we are trying to better understand
normal memory processes in adults. We are also trying to improve our
current memory tests, and the results that you provide will help us do
that.

In this study, you will do many different things. Many of the
tests measure memory directly. There are also some non-memory tests
that will help us examine functions that are sometimes related to memory
test performance, such as the ability to read, write, hear, and draw.
The exam will take approximately two and a half hours.

All information that you provide us is strictly confidential.
After your exam, your test results will be filed under a code, without
using your name or any biographical information. These records will be
maintained in a secured file. This exam is for research purposes, so no
written report will be issued. However, if your test results suggest a
possible problem, we may recommend that you see your primary care
physician.

Participation in this study is voluntary, and you are free to
withdraw from the project at any time. In addition, you may refuse to
do a particular task, and we will continue with the next one.
Certainly, the more that you are able to complete, the more your
participation will assist us in our research.

At this time, please ask any questions you have. Then, indicate
your understanding and willingness to participate by signing this form
in the space provided below.

 

Signature

 

Date

Appendix K

Appendix K

Debriefing Forp

232

 

233

DEVELOPMENTAL MEMORY STUDY:

A FINAL NOTE TO OUR VOLUNTEERS

Thank you so much for your help with our project. Your
contribution will help us better understand normal changes that take
place in memory throughout adulthood. Also, your responses~~as a group-
~provide us with a baseline of normal functioning, a picture of how
individuals in the community perform on these tests. As soon as the
project is complete, we will use this information in our own work with
neurological patients here in Lansing; in addition, there are many other
health professionals worldwide who have expressed interest in our
materials and in your results.

PLEASE: It is very important that other volunteers do not learn
specifics about the testing materials or about the testing procedures,
except information that was disseminated by the Clinic staff prior to
testing. In other words, PLEASE DO NOT DISCUSS WITH ANYONE THE
SPECIFICS OF THE TESTING until the project is complete. Even though it
will be tempting to compare responses and to describe your experiences
to family members and friends, it is vital, to our research and to our
provision of effective clinical services, that you wait until the
project is complete.

 

Thank you again for your support and thank you in advance for your
cooperation with the above request. We apologize for any inconvenience
this may cause. When the project is complete, Phil Fastenau will
arrange to share the outcome of the study with you and to discuss it
with interested groups.

LIST OF REFERENCES

LIST OF REFERENCES

Albert, M. [S.] (1984). Assessment of memory loss. Psychosomatics, 25
(Suppl. 12), 18-20.

Albert, M. S. (1988). Assessment of cognitive dysfunction. In M. 8.
Albert & M. B. Moss (Eds.), Geriatric neuropsychology (pp. 57-81).
NY: Guilford Press.

Barbizet, J., & Cany, E. (1968). Clinical and psychometrical study of a
patient with memory disturbances. Interpational Journal of
Neurology. 7, 44-54.

Barr, M. L., & Kiernan, J. A. (1988). The human nervous system: An
apatomical viewpoint (5th ed.). Philadelphia: Lippincott.

Bauer, R. M. (1993). Agnosia. In K. M. Heilman & E. Valenstein (Eds.),
Clinical neurooaychology (3rd ed., pp. 215-278). NY: Oxford
University Press.

Beck, A. T. (1978). Beck Depression Inventory. San Antonio, TX:
Psychological Corporation.

Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric
properties of the Beck Depression Inventory: Twenty-five years of
evaluation. Clinical Ps cholo Review 8, 77-100.

Bennett, P. H. (1990). Epidemiology of diabetes mellitus. In H. Rifkin
& D. Porte, Jr. (Eds.), Ellenbergaand Rifkin’a diabetes pellitua;
Theory and practice (4th ed.l (PP. 357-377). NY: Elsevier.

Bennett-Levy, J. (1984). Determinants of performance on the Rey-
Osterrieth Complex Figure Test: An analysis and a new technique
for single-case assessment. British Journa of Clinical
Ps cholo 23, 109-119.

Benson, D. F. (1985). Language in the left hemisphere. In D. F.
Benson, & E. Zaidel (Eds), The dual bpain: Hemispheric
specialization in humans (pp. 193-203). NY: Guilford Press.

Benton, A. L. (1974). The Revised Visual Retention Tea; (4th ed.). NY:
Psychological Corporation.

 

Benton, A. L., & Tranel, D. (1993). Visuoperceptual, visuospatial and
visuoconstructive disorders. In K. M. Heilman & E. Valenstein

234

235

(Eds.), Clinicalineuropsychology (3rd ed., pp. 165-214). New York:
Oxford University Press.

Benton, A. L., Hamsher, K. deS., Varney, N. R., & Spreen, O. (1983).
Contributiona to neuropaychological assessment: A clinical manual.
NY: Oxford University Press..

Berry, De To Re, Allen, Re Se, & SChlitt’ Fe Ae (1991). The Rey-
Osterrieth Complex Figure: Psychometric characteristics in a
geriatric sample. The Clinical Neuropaychologiat. 5, 143-153.

Bienenfeld, D. (1990a). Physical changes with aging. In D. Bienenfeld
(Ed.), Clinical geropsychiatry (pp. 3-16). Baltimore: Williams &
Wilkins.

Bienenfeld, D. (1990b). Substance abuse. In D. Bienenfeld (Ed.),
Clinical geropsychiatry (pp. 164-177). Baltimore: Williams &
Wilkins.

Binder, L. M. (1982). Constructional strategies on Complex Figure
drawings after unilateral brain damage. Journal of Clinical

Neuropsychology, 4(1), 51-58.

Bornstein, R. A., & Kelly, M. P. (1991). Risk factors for stroke and
neuropsychological performance. In R. A. Bornstein & G. Brown

(Eds.), Neurobehayioral aspecta of cerebrovascular disease (pp.
182-201). NY: Oxford.

 

Brouwers, P., Cox, 0., Martin, A., Chase, T., & Fedio, P. (1984).
Differential perceptual-spatial impairment in Huntington’s and
Alzheimer’s dementias. Archives of Neurolo 41, 1073-6.

Chaplin, W. E. (1984). State-Trait Anxiety Inventory. In D. J. Keyser
& R. C. Sweetland (Eds.), Iaat Critiquea: Vol. I (pp. 626—632).
Kansas City, MO: Test Corporation of America.

Chiarello, C. (1988). Lateralization of lexical processes in the normal
brain: A review of visual half-field research. In H. A. Whitaker

(Ed.), Contapporary reviews in neuropsychology (pp. 36-73). NY:
Springer-Verlag.

Cohen, J., & Cohen, P. (1983). Applied multiple regressionlcorrelatipn
analysis for the behavioral sciences (2nd edition). Hillsdale, NJ:

Lawrence Erlbaum and Associates.

Coltheart, M., Hull, E., Slater, D. (1975). Sex differences in imagery
and reading. Nature, 253, 438-440.

Craik, F. I. M. (1977). Age differences in human memory. In J. E.

Birren & K. W. Schaie (Eds.), Hapdbook of the psychology of aging
(pp. 384-420). NY: Van Nostrand Reinhold Co.

Craik, F. I. M. (1979). Human memory. Annual Review of Ps cholo 30,
63-102.

236

Craik, F. I. M., & Byrd, M. (1982). Aging and cognitive deficits: The
role of attentional resources. In F. I. M. Craik & S. Trehub
(Eds.), Aging and cggpitive processes (pp. 191-211). NY: Plenum
Press.

 

Craik, F. I. M., & McDowd, J. M. (1987). Age differences in recall and

recognition. Journal of Experimental Paychology: Learning;
Mapory. and Cognition. 13(3), 474-479.

Crossley, M., Hiscock, M., & Beckie, J. (1991, February). Aging and
divided-attention: A four-year follow-up study. Paper presented
at the annual meeting of the International Neuropsychological
Society, San Antonio, Texas, USA.

Cummings, J. L. (1985). Hemispheric asymmetries in visual-perceptual
and visual-spatial function. In D. F. Benson, & E. Zaidel (Eds),
The dual brain: Healapheric apecialiaation in huaapa (pp. 233-
246). NY: Guilford Press.

Cummings, J. L., & Mahler, M. E. (1991). Cerebrovascular dementia. In -
R. A. Bornstein a G. G. Brown (Eds.), Neurobehavioral aspects of
cerebrovaacular disease (pp. 131-149). NY: Oxford.

 

Damasio, A. R., & Anderson, S. W. (1993). The frontal lobes. In K. M.

Heilman & E. Valenstein (Eds.), Clinical neuropsychology (3rd ed.,
pp. 409-460). New York: Oxford University Press.

Davidson, M. B. (1991). An overview of diabetes mellitus. In R. B.
Frykberg (Ed.), The high risk foot in diabetes mellitus (pp. 1~31).
NY: Churchill Livingstone.

Delaney, R. C., Rosen, A. J., Mattson, R. H., & Novelly, R. A. (1980).
Memory function in focal epilepsy: A comparison of non-surgical,
unilateral temporal lobe and frontal lobe samples. Cortex, 16,
103-117.

DeLuca, J., & Cicerone, K. D. (1991). Construct validity of tests of
verbal and visual mental imagery. Clinicaleeurops cholo ist 5,
176-182.

Denburg, N. L., Fastenau, P. S., & Fertuck, E. A. (1994, February).
Cowboy Story: Resistance to aging found in reading memory. Paper
presented at the annual meeting of the International
Neuropsychological Society, Cincinnati, Ohio, USA.

Epilepsy Foundation of America (1992). How to recognize and classify
seizurea_and epilepay. Landover, MD: Author.

Erickson, R. C., & Scott, M. L. (1977). Clinical memory testing: A
review. Psychological Bulletin, 84, 1130-1149.

Fastenau, P. S., & Denburg, N. L. (1994, February). Reliability and
validity of the Extended Complex Figure Test (ECFT). Paper

237

presented at the annual meeting of the International
Neuropsychological Society, Cincinnati, Ohio, USA.

Fastenau, P. S., & Manning, A. A. (1992, February). Development of a
recognition task for the Complex Figure Test. Paper presented at
the annual meeting of the International Neuropsychological
Society, Sna Diego, California, USA.

Fastenau, P. S., & Sloan, L. E. (1993, February). Copy trial for the Wechsler
Memory Scale-Revised Visual Reproductions subtest: Normative study.
Paper presented at the annual meeting of the International Neuropsycho-
logical Society, Galveston, TX, USA.

Fisher, D. 6., Sweet, J. J., & Pfaelzer-Smith, E. A. (1986). Influence
of depression on repeated neuropsychological testing. Ipa
International Journal of Clinical Neuro s cholo 8, 14-18.

Franz, S. I. (1919). Handbook of aental examination pethoda. NY:
Journal of Nervous and Mental Disease.

 

Fromm, D., a Schopflocher, D. (1984). Neuropsychological test
performance in depressed patients before and after drug therapy.
Biolo ical Ps chiatr 19, 55-72.

Gazzaniga, M. S. (1983a). Right hemisphere language following brain
bisection: A 20-year perspective. Aperican Paycholo ist 38, 525-
537.

Gazzaniga, M. S. (1983b). Reply to Levy and to Zaidel. American
Ps cholo ist 38, 547-549.

Gazzaniga, M. S. (1984). Right hemisphere language: Remaining
problems. American Ps cholo ist 39, 1494-1496.

Gazzaniga, M. S., & LeDoux, J. E. (1978). The integrated mind. NY:
Plenum Press.

Gorsuch, R. L. (1991). Things learned from another perspective (so
far). Apericangaycholo ist 46, 1089-1090.

Hanger, P. A., Montague, J. R., & Smith, M. (1991, February). A
recognition memory test for the Visual Reproduction Subtest of the
Wechsler Memory Scale-Revised. Poster presented at the Annual
Meeting of the International Neuropsychological Society, San
Antonio, Texas.

Harris, L. J. (1978). Sex differences in spatial ability: Possible
environmental, genetic, and neurological factors. In M. Kinsbourne

(Ed.), Asypmetrical function of the brain (pp. 405-522).
Cambridge: Cambridge University Press.

Harris, L. J. (1981). Sex-related variations in spatial skill. In L.
S. Liben, A. H. Patterson, & N. Newcombe (Eds.), Spatial

238

representation and behaviorpacroaa the life span: Theory gag
application (pp. 83-125). NY: Academic Press.

Harris, L. J. (1985). Delicacy of fibres in the brain: Early and
recent neuropsychological explanations of sex differences in
cognition and temperament. In J. Ghesquiere, R. D. Martin, & F.
Newcombe (Eds.), Human sexual dimorphism (pp. 283-321). London:
Taylor and Francis.

Hasher, L., & Zacks, R. T. (1979). Automatic and effortful processes in
memory. Journal of Ex erimental Ps cholo : General 108, 356-388.

Heilman, K. M., & Valenstein, E. (Eds. 1993). Clinical neuropsychology
(3rd ed.). NY: Oxford University Press.

Heilman, K. M., & Van Den Abell, T. (1979). Right hemispheric dominance
for mediating cerebral activation. Neuropaychologia, 17, 315-321.

Heilman, K. M., & Van Den Abell, T. (1980). Right hemispheric dominance
for attention: The mechanism underlying hemispheric asymmetries of
inattention (neglect). Neurolo 30, 327-330.

Heilman, K. M., Watson, R. T., & Valenstein, E. (1993). Neglect and
related disorders. In K. M. Heilman & E. Valenstein (Eds.),
Clinical neuropsychology (3rd ed., pp. 279-336). NY: Oxford
University Press.

Hemsley, D. (1974). Relationship between two tests of visual retention.
Perceptual and Motor Skills, 39, 1132-1134.

Howe, M. L., & Hunter, M. A. (1985). Adult age differences in storage-
retrieval processes: A stages-of-learning analysis of
developmental interactions in concreteness effects. Can ian

Journal of Psychology, 39(1), 130-150.

Howe, M. L., O’Sullivan, J. T., Brainerd, C. J., & Kingma, J. (1989).
Localizing the development of ability differences in organized
memory. Contem rar Educational Ps cholo 14, 336-356.

Hyde, J. S. (1990). Meta-analysis and the psychology of gender
differences. Si ns 16, 55-73.

Jacoby, L. L. (1984). Incidental versus intentional retrieval:
Remembering and awareness as separate issues. In L. R. Squire and

N. Butters (Eds.), Neuropsychology of memory (pp. 145-156). NY:
Guilford Press.

Joreskog, K. G., & Sorbom, D. (1989). LISREL 7.16: Estimation pf
linear structupal equation systems [Computer program].
Mooresville, IN: Scientific Software, Inc.

Kaplan, E. (1988). A process approach to neuropsychological assessment.
In T. Boll & B. K. Bryant (Eds.), Clinical neuropsychology and

239

ppain function: Research. measurement, and practice (pp. 125-167).
Washington, DC: American Psychological Association.

Kimura, D. (1963). Right temporal-lobe damage. Archivea of Neurologyy
8, 264-271.

King, G. D., Hannay, H. J., Masek, B. J., & Burns, J. W. (1978).
Effects of anxiety and sex on neuropsychological tests. Journal of
Copaulting and Clinical Ps cholo 46, 375-376.

King, M. C. (1981). Effects of non-focal brain dysfunction on visual
memory. Journal of Clinical Paychology, 37, 638-643.

Klicpera, C. (1983). Poor planning as a characteristic of problem-
solving behavior in dyslexic children: A study with the Rey-
Osterrieth Complex Figure Test. Acta Paedopsychiatrica, 49, 73-82.

Koppitz, E. M. (1964). The Bender Geapalt Teat for young children. NY:
Grune & Stratton. -

La Rue, A. (1992). Aging and neurgpaychological assessmen . NY: Plenum.

Levy, J. (1983). Language, cognition, and the right hemisphere: A
response to Gazzaniga. American Ps cholo ist 38, 538-541.

Lezak, M. D. (1983). Neuropsychological assessment (2nd ed.). NY:

Oxford University Press.

McGlone, J. (1978). Sex differences in functional brain asymmetry.
Cortex, 14, 122-128.

Mesulam, M-M. (1985a). Attention, confusional states, and neglect. In
M-M. Mesulam (Ed.), Principles of behavioral neurology (pp.
125-168). Philadelphia: F. A. Davis Co.

Mesulam, M-M. (1985b). Patterns in behavioral neuroanatomy:
Association areas, the limbic system, and hemispheric
specialization. In M-M. Mesulam (Ed.), Principles of behayioral
neurology (pp. 1-70). Philadelphia: F. A. Davis Co.

Milberg, W. P., Hebben, N., & Kaplan, E. (1986). The Boston Process
Approach to neuropsychological assessment. In 1. Grant & K. M.

Adams (Eds.), Neuropaycological assessment of neuropsychiatric
disorders (pp. 65-86). NY: Oxford University Press.

Milner, B. (1971). Interhemispheric differences in the localization of
psychological processes in man. British Medical Bulletin. 21(3),
272-277.

Milner, B. (1975). Psychological aspects of focal epilepsy and its
neurological management. In D. P. Purpura, J. K. Penry, & R. D.
Walter (Eds.), Advances in neurology (Vol. 8, pp. 299-321). NY:
Raven Press.

240

Milner, B. (1978). Clues to the cerebral organization of memory. In P.
A. Buser and A. Rougeul-Buser (Eds.), Cerebpal correlates of
conacious experience: INSERM aypposiup no. 6 (pp. 139-153).
Amsterdam: Elsevier/North-Holland Biomedical Press.

Moss, M. 8., Albert, M. S., Butters, N., & Payne, M. (1986).
Differential patterns of memory loss among patients with
Alzheimer’s disease, Huntington’s disease, and alcoholic
Korsakoff’s syndrome. Archives of Neurology. 43, 239-246.

Myers, J. J. (1984). Right hemisphere language: Science or fiction?
Aperican Paychologiat. 39, 315-320.

Nakra, B. R. S., & Grossberg, G. T. (1990). Mood disorders. In D.

Bienenfeld (Ed.), Clinical geropsychiatry (pp. 107-124).
Baltimore: Williams & Wilkins.

National High Blood Pressure Education Program (1992). The fifth report
of the Joint Natiopal Copaittee on Detection; Evaluation, and
Treatment of High Blood Preaaure. Bethesda, MD: Author; National
Heart, Lung, and Blood Institute; National Institutes of Health.

National Institute on Alcohol Abuse and Alcoholism (1990). Seventh
special report to the U. 8. Congress on alcohol and health
(Contract No. ADM-281-88-0002). Rockville, MD: Author; Alcohol,
Drug Abuse, and Mental Health Administration; Public Health
Service; U. S. Department of Health and Human Services.

National Institutes of Health Consensus Development Conference (1990).

Sprgery for Epilepsy: Consensus Statement. Bethesda, MD: Office
of Medical Applications of Research, National Institutes of Health.

Nelson, N., Rosenthal, R., & Rosnow, R. (1986). Interpretation of
significance levels and effect sizes by psychological research.
American Ps cholo ist 41, 1299-1301.

Ogden, J. A. (1987). The "neglected" left hemisphere and its
contribution to visuospatial neglect. In M. Jeannerod (Ed.),

Neurophysiological and neuropsychologipal aspects of spatial
neglect (pp. 215-233). Amsterdam: Elsevier Science Publishers
BIVO

Olton, D. S., Gamzu, E., a Corkin, S. (Eds.; 1985). Memory

dysfunctions: An integration of animal and human research from

preclinical and clinical perspectives. (Annals of the New York
Academy of Sciences, vol. 444)

Osterrieth, P. A. (1944). Le test du copie d’une figure complexe.
Archives of Psychology (Chicago), 30, 206-356.

Ozer, D. J. (1985). Correlation and the coefficient of determination.

Psychological Balletip, 97, 307-315.

241

Palmer, S. E. (1977). Hierarchical structure in perceptual
representation. Cognitive Paychology, 9, 441-474.

Perlmutter, M., & Mitchell, D. B. (1982). The appearance and
disappearance of age differences in adult memory. In F. I. M.
Craik & S. Trehub (Eds.), Aging and cognitive processes (pp. 127-
144). NY: Plenum Press.

 

Pillon, B. (1981). Troubles visuo-constructifs et méthodes de
compensation: Resultats de 85 patients atteints de lesions
cérébrales. Neuro s cholo ia 19, 375~383.

Poon, L. W. (1985). Differences in human memory with aging: Nature,
causes, and clinical implications. In J. E. Birren & K. W. Schaie
(Eds.), Handbook of the psychology of aging (2nd ed.; pp. 427-462).
NY: Van Nostrand Reinhold Co.

Posner, M. I., & Petersen, S. E. (1990). The attention system of the
human brain. Annual Review of Neuroscience, 13, 25-42.

Posner, M. I., & Presti, D. E. (1987). Selective attention and
cognitive control. Trenda in Neuroscience, 10, 13-16).

Rabinowitz, J. C., Craik, F. I. M., & Ackerman, B. P. (1982). A
processing resource account of age differences in recall. Canadian
Journal of Paychology, 36(2), 325-344.

Rao, S. M., Hammeke, T. A., Huang, J. Y. S., et al. (1982, February).
Memory disturbance in chronic, progressive multiple sclerosis.
Paper presented at the tenth annual meeting of the International
Neuropsychological Society, Pittsburgh.

Rausch, R. (1985). Differences in cognitive functions with left and
right temporal lobe dysfunction. In D. F. Benson & E. Zaidel

(Eds.), The dual brain: Hemispheric specialization in humans (pp.
247-261). NY: Guilford Press.

 

Read, D. E. (1987). Neuropsychological assessment of memory in the
elderly. Canadian Journal of Psychology, 41(2), 158-174.

Reed, S. K. (1974). Structural descriptions and the limitations of
visual images. Memory and Cognition, 2(2), 329-336.

Rehm, L. P. (1988). Assessment of depression. In A. S. Bellack & M.
Hersen (Eds.), Behavioral assessment: A praptical handbook (3rd
ed.; pp. 313-364). NY: Pergamon Press.

Reitan, R. M., & Wolfson, D. (1993). The Halstead-Reitan
Neuropsychological Teat Battery: Theory and clinical
interpretation (2nd ed.). Tuscon, AZ: Neuropsychology Press.

Rey, A. (1941). L’examen psychologique dans les cas d’encephalopathie
traumatique. Archives oflPaycholpgy (Chicago), 28, 286-340.

242

Rey, A. (1964). L’exaaen clinique paychologl_. Paris: Presses
Universitaires de France.

Richards, P. M., & Ruff, R. M. (1989). Motivational effects on
neuropsychological functioning: Comparison of depressed versus
nondepressed individuals. Journal of Consulting and Clinical

Psychology, 57, 396-402.

Rosnow, R. L., & Rosenthal, R. (1988). Focused tests of significance
and effect size estimation in counseling psychology. Journal of
Counselin Ps cholo 35, 203-208.

Ruskin, P. E. (1990). Anxiety and somatoform disorders. In D.
Bienenfeld (Ed.), Clinical geropsychiatry (pp. 137-150).
Baltimore: Williams & Wilkins.

Smith, A. D. (1977). Adult age differences in cued recall.
Developaantal Psychology, 13 (4), 326-331.

Smith, A. D. (1980). Age differences in encoding, storage, and
retrieval. In L. W. Poon, J. L. Fozard, L. S. Cermak, D. Arenberg,

a L. W. Thompson (Eds.), New directionalin memory and aging (pp.
23-45). Hillsdale, NJ: Lawrence Erlbaum Associates.

Snyder, P. J. (1993). Handedness and sex effects on visuoconstructive
and visuoperceptual abilities in college students (Doctoral
dissertation, Michigan State University, 1992). Dissertation
Abstracts International, 53(7), 38308.

Solomon, P. R., Goethals, G. R., Kelley, C. M., a Stephens, B. R. (Eds.,
1989). Memory: Interdisciplinary approache . NY: Springer-Verlag.

Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs,

G. A. (1983). Manual for the State-Trait Anxiety Inventory. Palo
Alto, CA: Consulting Psychologists Press, Inc.

 

Springer, S. P., & Deutsch, G. (1985). Left brain, right brain (rev.
ed.). NY: Freeman & Co.

SPSS, Inc. (1990). Statistical Package for the Social Sciences:
Release 4 [Computer program]. Chicago, IL: Author.

Squire, L. R. (1982). The neuropsychology of human memory. Annual
Review of Neuroagience, 5, 241-273.

 

Squire, L. R. (1986). The neuropsychology of memory dysfunction and its
assessment. In 1. Grant & K. M. Adams (Eds.), Neuropsychological

assessment of neuropsychiatric disorders (pp. 268-299). NY:
Oxford University Press.

Squire, L. R., & Butters, N. (Eds., 1985). Neuropsychology of memory.
NY: Guilford Press.

243

Stankov, L. (1988). Aging, attention, and intelligence. Psychology and
A in 3, 59-74.

Talland, G. A. (1965). Deranged memory: A paychonopic atudy of the
amnesic sypdrome. NY: Academic Press.

Taylor, L. B. (1959). The appraisal of children with cerebral
deficits. Cambridge, MA: Harvard University Press.

Taylor, L. B. (1969). Localisation of cerebral lesions by psychological
testing. Clinical Neurosur er 16, 269-287.

Tombaugh, T. N., Hubley, A. M., Faulkner, P., & Schmidt, J. P. (1990,
February). The Rey-Osterrieth and Taylor complex figures:
Comparative studies, modified figures and normative data for the
Taylor Figure. Paper presented at the 18th Annual Meeting of the
International Neuropsychological Society, Kissimmee, FL.

U. S. Department of Health and Human Services (1993). Health United

States 1992 and Healthy Beople ZOOOlHeview (DHHS Pub. No. PHS 93-
1232). Hyattsville, MD: Author.

 

Waber, D. P., Bernstein, J. H., & Merola, J. (1989). Remembering the Rey-
Osterrieth Complex Figure: A dual-code, cognitive neuro-
psychological model. Developaental Neuropsychology, 5(1), 1-15.

Waber, D. P., & Holmes, J. M. (1985). Assessing children’s copy
productions of the Rey-Osterrieth Complex Figure. Journal of
Clinical and Experimental Neuropsychology, 7(3), 264-280.

Waber, D. P., a Holmes, J. M. (1986). Assessing children’s memory
productions of the Rey-Osterrieth Complex Figure. Joprnal of
Clinical and Experimental Neuropsychology, 8(5), 563-580.

Waldstein, S. R., Manuck, S. B., Ryan, C. M., & Muldoon, M. F. (1991).
Neuropsychological correlates of hypertension: Review and

methodologic considerations. Psychological Bulletin, 11 , 451-468.

Waugh, N. C., & Barr, R. A. (1982). Encoding deficits in aging. In F.

I. M. Craik & S. Trehub (Eds.), Aging and cognitive processes (pp.
183-190). NY: Plenum Press.

Wechsler, D. (1945). A standardized memory scale for clinical use.
Journalaof Ps cholo 19, 87-95.

Wechsler, D. (1981). Wechsler Agult Intelligenca Scale-Revised manual.
NY: Psychological Corporation. ‘

Wechsler, D. (1987). Wechsler Memory Scale-Revised manual. San
Antonio, TX: Psychological Corporation.

Weintraub, S., & Mesulam, M-M. (1985). Mental state assessment of young
and elderly adults in behavioral neurology. In M-M. Mesulam (Ed.),

 

244

Principlea of behavioral neurology (pp.71~123). Philadelphia:
F. A. Davis Co.

Wilson, R. S., Rosenbaum, G., & Brown, G. (1979). The problem of
premorbid intelligence in neuropsychological assessment. Journal
of Clinical Neuropsychology, 1(1), 49-53.

Winograd, E., A Simon, E. W. (1980). Visual memory and imagery in the
aged. In L. W. Poon, J. L. Fozard, L. S. Cermak, D. Arenberg, & L.
W. Thompson (Eds.), New directiona in memory and aging (pp. 485-
506). Hillsdale, NJ: Lawrence Erlbaum Associates.

Witelson, S. F. (1976). Sex and the single hemisphere: Specialization of
the right hemisphere for spatial processing. Science. 193, 425-427.

Woodcock, R. W., & Johnson, M. B. (1989). Woodcock-Johnaon Paycho-
educational Teat Battery. Heviaag. Chicago: Riverside Publishing.

Wrightsman, L. S. (1962). The effects of anxiety, achievement
motivation, and task importance to intelligence test performance.
Jourpal of Educational Ps cholo 53, 150-156.

Zaidel, E. (1977). Unilateral auditory language comprehension on the
Token Test following cerebral commissurotomy and hemispherectomy.
Neuro s cholo ia 15, 1~18.

Zaidel, E. (1983). A response to Gazzaniga: Language in the right
hemisphere, convergent perspectives. American Ps cholo ist 38,
542-546.

Zaidel, E. (1985). Language in the right hemisphere. In D. F. Benson,

& E. Zaidel (Eds), The dual brain: Hemispheric specialization in
humans (pp. 193-203). NY: Guilford Press.