 

5?...1

x...

 

.
I. x
LIABLE)".
: a

.1. 4 rr
; . Lazy
.32.
2...

ran

:1!

. 03...... .t;«(
1

. 1331)..

Z. \ .
.3; 11.
«311‘
.1 #1..
.1...

‘ m} U
£3mA329uw...
: a... as;
I .t

{4.1.} I
:2..,ah...3\n..
.41! a t
3 351. mu.
2»... .5 .... mu
(1". k.
5, t z.
.. aurauﬁisi
Err“. ..m
a... ..

tilt: .!

it: .

9.!

Val

51:3.
3
15A.
:

373,!
9a 72

.45.... . .«rf.

l
t 5..

r.
3.

M33»

.u

:
as... «I;

. (II! 7.2
.2 is. _.r 1:!
k 1.1.1 1.31"

A'§.W\_
, in.

.1 at
uni)
u4o:

3.1..

11.41,
. . . , -o )va.!
St... A. :51»... 3&3?» . . iv 1:55.. 31:. 3'.

 

 

y... ...u.,....&e.n..a.§ut.,..,.
,..L. 91.. innit, ‘ ‘
4.9.15: ...hy.éh..ﬁ.nlm.mwwﬁ n

1 was":

,2
2009—

llBRARY
Michigan State
University

This is to certify that the
dissertation entitled

An Investigation of the Prefrontal Cortex
Function Theory Of Cognitive Aging

presented by

Julie Daugherty Simensky

has been accepted towards fulﬁllment
of the requirements for

Ph-D- degreein Clinical Psychology

ﬁlm, We

Major professor

 

Datew

MSU is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

DEC 0 3 2002

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDuep65oo15

AN INVESTIGATION OF THE PREFRONTAL CORTEX FUNCTION
THEORY OF COGNITIVE AGING

By

Julie Daugherty Simensky

A DISSERTATION
Submitted to
Michigan State University

in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Psychology

2001

ABSTRACT

AN INVESTIGATION OF THE PREFRONTAL CORTEX FUNCTION THEORY OF
COGNITIVE AGING

By

Julie Daugherty Simensky

The particular vulnerability of the prefrontal cortex to age-related deterioration
has been frequently posited as an explanation for functional decline in aging. West’s
(1996) prefrontal cortex ﬁrnction theory of cognitive aging maintains that age-related
decline in cognitive processes supported by the prefrontal cortex emerge at an earlier
age and are of greater magnitude than age-related declines observed in cognitive
processes supported by nonfrontal regions. In this study, four speciﬁc processes
underlying prefrontal cortex ﬁinction (prospective memory, retrospective memory,
interference control, and inhibitory control) were theorized to account for age-related
declines in a number of cognitive functions, including verbal memory. This study
investigated to what extent these four prefrontal cortex processes were discernable and
the mediating role they played between age and verbal memory. One hundred nine
healthy older adults, aged 55-88 years, completed multiple performance measures
thought to evaluate prefrontal cortex function, verbal memory, and depression.
Exploratory factor analyses indicated a two-factor model of prefrontal cortex function
that was only marginally consistent with West’s proposed model. These two factors
were: 1) a factor containing measures of interference control, sustained attention,

processing speed, and inhibitory control; and 2) a working memory factor. Structural

equation modeling analyses indicated that this two-factor measurement model strongly
mediated the relationship between age and verbal memory performance. Verbal
memory performance decreased as a result of poorer prefrontal cortex function
performance with older age. Results suggested that despite difﬁculties in the speciﬁc
and selective measurement of prefrontal cortex function, there is strong support for the
hypothesis that cognitive aging is highly dependent on the integrity of prefrontal cortex

functions.

In memory of my father, Richard A. Daugherty

iv

ACKNOWLEDGEMENTS

The assistance and support of many people contributed to my ability to
successfully ﬁnish this project. I want to ﬁrst thank members of my committee
for their unique contributions and support. In particular, I want to sincerely thank
my committee chairperson and advisor, Norman Abeles, Ph.D., for his guidance,
wisdom, and unwavering support throughout my graduate studies. I consider it an
honor to have had the privilege to work with him. I would also like to thank Alex
VonEye, PhD. for his particular expertise and guidance in completing my
statistical analyses. I want to especially thank Mindy Firland-Whitsett, PhD. for
her empathetic support, technical knowledge, and friendship — because of her
assistance structural equation modeling is no longer a mystery to me. I must also

extend my sincere gratitude to the participants who made this study possible.

Most importantly, I want to thank my family. You have loved, supported,
and encouraged me throughout all of my accomplishments and hardships.
Because of you, I am proud of who I am today. To my husband, Steve - there are
no words strong or beautiful enough to convey how much your support, patience,
humor, and love have meant to me throughout the years and during the

completion of this project. Thanks to you and to my family for helping me to

keep my focus on the fact that the true joy of life is the trip, rather than the ﬁnal

destination.

TABLE OF CONTENTS

LIST OF TABLES .................................................................. ix
LIST OF FIGURES .................................................................. x
INTRODUCTION ................................................................... 1
Neurobiology of the Aging Prefrontal Cortex ........................... 1
Structural Changes in the Aging Brain .......................... 1
Functional Changes in the Aging Brain ......................... 3
Cognitive Functions of the Prefrontal Cortex ............................ 7
The Role of Preﬁontal Cortex Function in Memory ......... 10
Hierarchical Model of Prefrontal Cortex Function ........... ll
Specialization of Function Within the Prefrontal Cortex. . .. 13
The Preﬁ'ontal Cortex Function Theory of Cognitive Aging ......... l6
Prospective Memory and Frequency Estimation .............. 19
Interference Control and Sustained Attention ................. 23
Inhibition of Prepotent Response ................................. 25
Working Memory ................................................. 26
Measurement and Alternative Theoretical Considerations ........... 30
HYPOTHESES ...................................................................... 32
METHOD ............................................................................ 35
Participants .................................................................. 35
Procedure .................................................................... 36
Measures ..................................................................... 36
Wisconsin Card Sorting Test ..................................... 36
Stroop Color and Word Test ..................................... 37
Paced Auditory Serial Addition Test ........................... 38
Ruff Figural Fluency Test ........................................ 39
Trailmaking Test A & B .......................................... 4O
Symbol-Digit Modalities Test .................................... 40
Controlled Oral Word Association Test ......................... 4O
Wechsler Memory Scale-Revised Digit Span .................. 41
Wechsler Memory Scale-Revised Visual Memory Span ..... 41
California Verbal Learning Test ................................. 42
Wechsler Memory Scale-Revised Logical Memory ........... 43
Measure of Mental Status ......................................... 43

vii

Measures of Depression .......................................... 44

RESULTS ............................................................................. 44
Missing Data ................................................................. 44
Preliminary Data Examination ........................................... 45
Hypothesis 1 ................................................................ 47

Correlations of Manifest Variables to Latent Variable ....... 48
Exploratory Analysis of 3-Factor Model ...................... 49
Identiﬁcation of a New Measurement Model ................ 52
Selection of Additional Variables ............................... 53
Results of Exploratory Factor Analyses ....................... 54
Hypothesis 2 - Structural Equation Modeling .......................... 58
Model 1 ............................................................. 59
Model 2 ............................................................. 60
Model 3 ............................................................. 61

DISCUSSION ....................................................................... 63

The Model of Prefrontal Cortex Function ............................... 64
Alternative Models ................................................ 67
Measurement Issues ............................................... 72
Relationship Between Measurement and Neuroanatomy. . . .74
Sample Characteristics ............................................ 77

Prefrontal Cortex Function as a Mediator ............................... 80

Theoretical Considerations and Future Directions ...................... 86

APPENDIX A ......................................................................... 89

APPENDIX B ......................................................................... 91

APPENDIX C ........................................................................ 93

APPENDIX D ......................................................................... 95

APPENDIX E ......................................................................... 97

REFERENCES ...................................................................... ‘. 99

viii

LIST OF TABLES

Table 1 - Four Secondary Processes of West’s (1996) Prefrontal Cortex Function
Theory of Cognitive Aging

Table 2 — Correlations of Manifest Variables to each Latent Variable
Table 3 — Communalities of the 12 Original Prefrontal Cortex Function Variables
Table 4 — Varimax 3 - Factor Structure of Retained Original Variables

Table 5 — Communalities of Retained Original Prefrontal Cortex Function
Variables and Added Relevant Variables

Table 6 — Varimax 3 — Factor Structure of Final Variables

Table 7 — Varimax 2 — Factor Structure of Final Variables

Lx

LIST OF FIGURES

Figure 1 — Three Factor Model of Prefrontal Cortex Function

Figure 2 — Two Factor Model of Prefrontal Cortex Function

Figure 3 — Causal Model of Prefrontal Cortex Function as a Mediator
Figure 4 — Model 1 Standardized Path Coefﬁcients - Direct Eﬁ‘ect's Model
Figure 5 — Model 2 Standardized Path Coefﬁcients — Indirect Effects Model

Figure 6 — Model 3 Standardized Path Coefﬁcients - Mediating Model

INTRODUCTION

A substantial body of research has identiﬁed measurable declines in some cognitive
abilities with normal aging. There is growing evidence that the process of normal aging
may be linked to a selective decline in cognitive functioning associated with the frontal
lobes, particularly the prefrontal cortex. Substantial neural changes occur with advancing
age and evidence demonstrates that the prefrontal cortex is more sensitive to the effects of
aging than other cortical regions (Haug & Eggers, 1991). Furthermore, persons with age-
associated memory impairment demonstrate disproportionate decline in frontal lobe
functioning as assessed by both neuropsycholgoical tests and neuroirnaging studies
(Hanninen et al., 1997). West (1996) proposed a prefrontal cortex function theory of
cognitive aging which incorporated existing theory and research. This theory states that
the prefrontal cortex is responsible for the integration, formation, and execution of
complex, novel behavioral structures or temporal gestalts, which support the direction of
behavior in an orderly, purposeful manner. The normal aging process interferes with the
secondary processes which contribute to this integrative function. The overall purpose of
this study is to investigate the applicability of this prefrontal cortex ﬁrnction theory of
cognitive aging to a sample of healthy older adults. The speciﬁc questions were: 1) to
what extent do the data ﬁt the model of prefrontal cortex function decline with age; and 2)
are the secondary processes which contribute to the overall integrative function of the

preﬁ'ontal cortex discernible in older adults?

Neurobiology of the Aging Preﬁontal Cortex
Structural Changes in the Aging Brain
The aging brain loses weight, volume, and neurons. Morphological changes in the

aging brain have been examined by a wide range of techniques, from measurement of the

gross anatomical volume of various regions to evaluation of synaptic density and integrity
and neurochemistry (West, 1996). Morphological study evidence of a general decrease in
gross brain volume reaching signiﬁcance in the 7th decade of life has been well
documented (Haug & Eggers, 1991; Raz, 1996). The greatest regional differences in the
degree of volume reduction, estimated to be 10-17%, are found in the prefrontal cortex.
This is compared to volume reductions in the temporal, parietal, and occipital cortices, all
estimated at 1% (Giaquinto, 1988; Guttman et al., 1998; Haug et al., 1983; Haug &
Eggers, 1991; Raz et al., 1997). These volume reductions are the result of a reduction in
neuron size rather than ﬁ'om an actual loss of neurons (Haug & Eggers, 1991). With
advancing age, all areas of the cortex evidence reduced neuronal size; however, earlier and
more severe reductions in neuron size are observed in the frontal cortex, when compared
to the parietal, temporal, or primary visual cortex.

Scheibel et a1. (1975) have described how neurons proceed through a series of
degenerative changes with advancing age which may explain the origin of reductions in
neural size with advancing age. Neurons appear to undergo subtle alterations with
progressive loss of horizontally oriented dendritic systems in areas that are known to
receive speciﬁc synaptic terminals. At the same time, this loss is accompanied or preceded
by the irregular swelling of the soma and major dendrites (Scheibel et al., 1975). This loss
of dendritic arborization and the cellular decrease are believed to impair the
communication capabilities of the nervous tissue. Reviewing studies on neural depletion
in aging, Squire (1987) suggested that "losses from the prefrontal cortex may be sizable,
even in terms of a daily estimate" (p.119). West (1996) concluded that the reduction in
neuronal size with advancing age could result from the loss of connective processes of the
neuron.

With advancing age, an increased number of pathological structures in the brain
accompanies the reduction in cortical volume. Senile plaques and tangles increase in

number throughout the life span in both humans and non-human primates (West, 1996).

In post-mortum studies, Struble et al. (1985) measured the origin and regional distribution
of senile plaques in non-human primates revealing greater numbers in the prefrontal and
temporal cortices compared with the posterior, frontal-parietal, or hippocampal regions.
Furthermore, these researchers suggested that the most mature plaques were located
within the prefrontal and temporal regions. Similar results were reported by Heilbroner
and Kemper (1990) who noted that highest concentrations of senile plaques were
observed in the frontal areas in non-human primates. Thus, in contrast to Alzheimer's
disease where plaques are predominantly found in the hippocampus, locus coeruleus,
nucleus basalis, and association cortex, healthy non-human primates demonstrate greater
concentrations in the frontal and temporal cortices (Giaquinto, 1988; West, 1996).

Other white matter changes in normal aging have been documented. Studying
MRI ﬁndings of periventricular high-signal intensity patterns in a sample of healthy, older
adults over age 60, Gerard and Weisberg (1986) found that 10% had periventricular
lesions. These lesions were localized to the anterior frontal regions in all cases. This
percentage increased to 31% when patients having risk factors but no cerebrovascular
symptoms were included. Furthermore, neuroirnaging has indicated demyelination in the
periventricular white matter connected to the anterior frontal horns (Gerard & Weisberg,
1986; Giaquinto, 1988; Guttman et al., 1998).

A number of region-speciﬁc declines in the synthesis, concentration, and number of
receptor sites for some neurotransmitters associated with cognitive functioning occur with
advancing age. Some have demonstrated greater changes in neurotransmitter systems
with advancing age within the frontal cortex in both humans and non-human primates
(DeKosky et al., 1985; Fuster, 1998; Goldman-Rakic & Brown, 1981; Marcusson et al.,
1984; Suhara, 1990; Wong, et al., 1984). However, others have failed to conﬁrm such
effects (Giaquinto, 1988).

Function_al Chan es in the ' Brain

Two neuroirnaging methods have been primarily used to study the ﬁmctional
integrity of the aging brain. The Xenon133 inhalation (Xe) method measures the rate of
regional cerebral blood ﬂow (rCBF) or oxygen use in the cortex. A newer and less
utilized method, positron emission tomography (PET) measures both cerebral blood ﬂow
and glucose metabolism. Utilizing Xe and the PET oxygen monitoring methods,
consistent declines in rCBF with advancing age have been documented (Gur et al., 1987;
Meyer et al., 1994; Shaw et al., 1984; Warren, Butler, Katholi, & Halsey, 1985). This
age-related reduction in rCBF during a resting state has been estimated to exceed 27% in
some cortical regions (Shaw et al., 1984). Speciﬁcally, in young and early middle-aged
participants (under age 45), a pattern of hyperﬁontality, characterized by greater blood
ﬂow in the anterior versus the posterior cortical regions is frequently observed.

However, in older participants, this pattern is reversed, referred to as hypofrontality. This
ﬁnding has been demonstrated in both cross-sectional and longitudinal studies (Gur et al.,
1987; Shaw et al., 1984). The greatest reduction in rCBF has been observed in the
prefrontal and temporal cortices, or selectively in the left-anterior cortex in cross-sectional
studies (Gur et al, 1987; Warren, Butler, Katholi, & Halsey, 1985). Shaw et al.'s (1984)
longitudinal study demonstrated that over a 4-year period, during which time participants
remained healthy, signiﬁcant reductions in rCBF were observed only in the prefrontal
cortex. Furthermore, this study also showed an interaction between the effects of normal
aging and the degree of cerebrovascular disease in relation to rCBF. For participants
younger than 70 years, the presence of increasingly severe cerebrovascular disease
resulted in increased declines in rCBF. For participants older than 70 years, this
interaction was not observed, suggesting that the effect of cerebrovascular disease may be
more deleterious during the middle and early-late adult years (West, 1996). Together,
these results demonstrate that with advancing age, rCBF at rest is selectively reduced,

especially in the frontal cortex.

Changes in rCBF with age have also been documented during the performance of
cognitive tasks. For all participants, an increase in rCBF is evidenced while performing
cognitive tasks such as mathematical calculations, verbal analogies, and verbal memory
tasks (Gur et al., 1987; Raglund et al., 1997; Warren et al., 1985). Age-related differences
in the magnitude or pattern of activation have not been observed in response to cognitive
demands, despite consistent declines in resting rate of rCBF (Gur et al., 1987 ; Warren et
al., 1985). However, neither of these studies considered the impact of behavioral
performance. Raglund et al. (1997) examined PET rCBF change during working and
declarative memory in relation to task performance and found that activational patterns
diverged when performance was considered. High performers on tasks of working
memory activated frontal dorsolateral and inferior regions whereas high performers on the
declarative memory task activated only the occipitotemporal region. Notably, older adults
were not included in this study; however, based on previously reviewed ﬁndings, it could
be hypothesized that different activational patterns may be evidenced by older adults.
Speciﬁcally, because of the well-documented prefrontal neural decline, less divergence in
activational patterns according to task performance would be expected. Grady et al.
(1995) obtained results from their rCBF study of young and old adults during encoding
and recognition of faces which partially supported this hypothesis. Young adults (M = 29
years) showed increased rCBF in the right hippocampus and left prefrontal and temporal
cortices during encoding and in the right prefrontal and parietal cortex during recognition.
Older adults (M = 69 years) demonstrated no signiﬁcant activation in areas activated
during encoding in young adults but did show right prefrontal activation during
recognition.

As previously mentioned, evidence from PET glucose metabolism studies is less
deﬁnitive. Some studies demonstrate trends towards reduced rates of cerebral metabolism
with age when young and old participants are compared, with increased reduction being
evidenced in the frontal regions (de Leon et al., 1983; Kuhl, Metler, Rieger, & Hawkins,

1984). Other research has shown preserved glucose metabolism with advancing age in
healthy adults (de Leon et al., 1984).

In summary, studies utilizing both Xe and PET methods have generally found age-
related decreases in oxygen use; however, PET studies have not typically demonstrated
signiﬁcant age-related changes in glucose metabolism (West, 1996). Both theoretical and
methodological reasons have been offered to explain this inconsistency. Roland (1984)
proposed that although oxygen utilization and glucose metabolism are closely related in
young— and middle-age, they become "decoupled" in older adults. Methodologically,
differences in sample characteristics could explain differences in study ﬁndings. West
(1996) suggests that participants in the Xe studies may not have been screened as strictly
as those in the PET glucose metabolism studies. Furthermore, because of the economic
and ethical costs of PET studies, smaller numbers of participants are used, which may
result in insufﬁcient power to detect meaningﬁil differences between young and old
participants.

There has been a recent proliferation of functional magnetic resonance imaging
(MRI) studies that have demonstrated strong support for the ﬁ'ontal aging hypothesis.
Grady et al. (1994) have demonstrated that older adults show greater cortical activation
outside of the prefrontal cortex when compared to younger adults during a spatial
working memory task. Greater left prefrontal activity during a verbal encoding task and
greater right prefrontal activity during a retrieval task demonstrated by younger adults was
not seen in a sample of older adults (Cabeza et al., 1997; Tulving, et al, 1994).
Comparitively, older adults demonstrated little prefrontal activity during encoding but
engaged bilateral prefrontal activity during retrieval. Madden et al. (1999) found that
during a recognition verbal memory task, younger adults demonstrated primarily right
prefrontal activation when compared to a more bilateral pattern of prefrontal activation
demonstrated by older adults. Renter-Lorenz (2000) studied PET activation patterns of

younger and older adults during both letter identity and spatial location memory tasks.

Younger adults showed left-sided frontal activation for the letter identity task and right-
sided frontal activation for the spatial location task, whereas older adults showed bilateral
frontal activation during both tasks. A possible explanation for these ﬁndings is that
preﬁ'ontal cortex functioning is selectively impaired in older adults and that the effort to
recruit all available prefrontal resources results in increased activation of prefrontal and
some non-prefrontal areas in older but not younger adults.

Taken together, these data suggest substantial structural and probable functional
changes in the cerebral cortex of healthy older adults. This provides considerable
morphological and functional support for a prefrontal cortex theory of cognitive aging.

Cognitive Functions of the Preﬁ'ontal Cortex

Harlow (1868) was one of the ﬁrst to describe the changes in behavior and
personality resulting from prefrontal damage in reports of his patient, Phineas Gage. The
nature of ﬁmction(s) of the prefrontal cortex have been the center of much focus and
debate. Some have suggested that few if any cognitive functions are controlled by the
preﬁ'ontal cortex, based on frequent observations that little 'or no change occurs in some
humans and animals after prefrontal insult (Stuss & Benson, 1986). Others, such as Luria
(1980), have theorized that the prefrontal cortex supports the highest level of cognitive
organization, serving as the central executive of the brain. Early theory and description of
prefrontal cortex ﬁmction was based on observations of survivors of frontal lobe damage
or animal studies. More recent advances such as neuroirnaging have allowed for more
precise examination and subsequent elaboration of knowledge of the prefrontal cortex.

Clinically, descriptions of changes as a result of prefrontal damage include
disruption of: initiation, planning and goal selection, attention, memory, anticipation,
abstraction, use of feedback, and self-awareness (Fuster, 1998; Luria, 1980; Stuss &
Benson, 1984; Perecman, 1987; Shimarnura, Janowsky, & Squire, 1991; West, 1996).

Luria (1973) described one of the more prevalent deﬁcits related to prefrontal damage as

the fragmentation of complex behavior and action sequences. While individual units that
comprise an action sequence are able to be performed perfectly, there is a lack of structure
and goal directedness. Furthermore, complex behavior is fragmented by a disruption of
the temporal gestalt (Fuster, 1998); a temporal guideline cannot be formed to direct
individual behavioral units required for task completion. For example, a person with a
prefrontal injury may be able to complete any individual house cleaning task in isolation
but when given a list of house cleaning tasks to perform, would have great difﬁculty doing
so because of an inability to organize the requirements of the list eﬁiciently.

When behavior is not guided by temporal gestalts, a number of behavioral deﬁcits
that are commonly observed in persons with prefrontal injury may occur. Stuss and
Benson (1987) have described a dissociation between self-consciousness and self-
knowledge; when a person is ﬁrlly aware of making performance errors based on some
external criteria but is unable to utilize this knowledge to change her/his behavior. This
dissociation has been clearly demonstrated in studies of persons with prefrontal damage on
two common neuropsychological tests purported to measure frontal lobe functioning. The
“Wisconsin Card Sorting Task (WCST; Heaton, 1981) requires an individual to sort a
series of cards based on a continually changing rule. When a category shift occurs,
persons with prefrontal damage frequently continue to sort the cards to the now incorrect
category, despite being able to express that their behavior is incorrect (Heaton, 1981;
Stuss & Benson, 1986). In the Stroop Color-Word Naming Task (Stroop; 193 5), one
trial requires the naming of the color of ink in which a word is printed rather than reading
the word. Persons with preﬁ'ontal damage will frequently continue to read the word
rather than naming the color despite being able to explain the rule they are to follow
(Duncan, 1995; West, 1996). Taken together, these performances suggest a type of goal
neglect; inability to utilize knowledge of "what to do" to guide behavior. West (1996)
theorizes that this goal neglect can lead to two other types of behavior that are often

observed in persons with prefrontal damage: perseveration and distractibility.

Perseveration is the tendency to persist in elements of some previous task or
behavior after that response has become inappropriate. In other words, an action
sequence gets "captured" by one of its constituent parts. Distractibility is the tendency to
respond to external stimuli that are not related to achieving a goal. When this occurs, an
action sequence is "captured" by an irrelevant element in the external environment. Both
result from the disturbance and breakdown of a temporal gestalt or sequence of actions
(F uster, 1998). Perseveration is a particularly common and noticeable consequence of
prefrontal pathology (Goldberg & Bilder, 1987; Luria, 1980; Stuss & Benson, 1984).
The Role of the Prefrontal Cortex in Inhibition

An inhibitory control model of prefrontal cortex function suggests that preﬁ'ontal
ﬁrnction is represented by processes which inhibit prepotent responses and sustain
attention. Inhibition of prepotent responses keeps highly salient but unsuited responses
from gaining control of a behavior or action sequence (Dempster, 1991). Inhibitory
processes also allow attention to be sustained, even while there is the risk of distraction
from external/environmental sources. Stuss and Benson (1986) highlighted the
importance of the prefrontal cortex in maintaining attention over time and preventing
distraction, both of which allow information to be organized into workable chunks.
Furthermore, inhibitory control allows attention to be focused and sustained upon novelty
within the environment. Detection of novelty permits the ability to encode other
behaviorally relevant aspects of an event, contributing to the formation of a temporal
gestalt (West, 1996). A deﬁcit in novelty detection not only hampers the ability to code
the beginning and end of discrete events and tag them with appropriate temporal-spatial
information (Knight & Grabowecky, 1995), it creates an experience of the environment in
which even common events and actions are unpredictable and unforeseen (F uster, 1998).

Both inhibitory control and sustained attention have been demonstrated to be
related to the functioning of the anterior cingulate within the midprefrontal region (Posner

& Raichle, 1996). The anterior cingulate functions to assist in preparing an individual for

some anticipated event. Speciﬁcally, it allows for inhibition of an automatic response so
that a less automatic or less compatible response can be reported, such as in the Stroop
color-word naming trial. Additionally, human neuropsychological studies have shown that
damage to the right dorsolateral region has been shown to impair performance on tasks of
sustained attention (Vendrell et al., 1995) and inhibition of prepotent responses (Wilkins,
Shallice, & McCarthy, 1987). Incisa della Rocchetta and Milner (1993) demonstrated that
left dorsal prefrontal incisions impaired the ability to suppress potentially interfering items
during a verbal memory task. However, some have questioned the frontal-lobe speciﬁcity
of these processes because of the interconnections the frontal lobes have to midbrain and
brain stem attention and arousal systems (Stuss & Benson, 1986).
The Role of Prefrorﬂ Cortex Function in Memory

The signiﬁcance of the frontal lobes in memory ﬁinction has been subject to much
research and debate. Some have emphasized the primary role of the frontal cortex on
memory tests such as delayed response, conditioned associative learning, and memory for
temporal order (Levine, Stuss, & Milberg, 1997; Mangels, 1997 ; Parkin, Walter, &
Hunkin, 1995; Petrides, 1985; Schacter, 1987). Others, however, have suggested that
memory ﬁrnctions per se are not impaired in patients with frontal damage and that
decrements in performance on memory tests are secondary to disorders of attention,
problem-solving, and inhibitory control (Hecaen & Albert, 1978; Luria, 1973). Stuss and
Benson (1986) have suggested that the accumulated evidence on the role of the frontal
lobes in memory implies that lesions and neuronal loss do not cause a primary disturbance
of memory, but that they do interfere with mnestic activity. In contrast, Shimamura,
Janowsky, and Squire (1991) have concluded that speciﬁc memory disorders are the result
of frontal lobe lesions, including impaired short-term memory, free recall, metarnemory,
memory for temporal order, and source memory.

A memory-based model of prefrontal function has been proposed largely on the

basis of evidence from non-human primate studies utilizing variations of the delayed-

10

response task (West, 1996). A typical delayed-response task would involve a monkey
watching while a reward was hidden at one of a number of locations. After a delay, the
monkey is allowed to retrieve the reward. Successﬁrl performance requires the
maintenance of an internal representation of the hidden location of the reward over a
delay. This representational memory, also referred to as working memory, allows the
animal to bridge the temporal gap between hiding and responding (Goldman-Rakic, 1987).
Disruption of representational memory can contribute to secondary deﬁcits, including
perseveration and distractibility (West, 1996). If an animal cannot maintain on-line,
context-speciﬁc information to guide behavior, then its behavior is inﬂuenced by habitual,
prepotent responses (perseveration) and/or external cues (distractibility).

In addition to the non-human primate literature supporting a representational
memory deﬁcit of prefrontal dysﬁmction, a computational model has been designed
(Kimberg & Farah, 1993) which has been shown to simulate performance accurately on a
number of tasks sensitive of the effects of prefrontal injury, including representational
memory. In this mathematical model, the strength of various associations in
representational memory is affected by preﬁ'ontal insult, resulting in a decreased ability to
determine the relevance of speciﬁc goals and declarative and environmental information
for task performance. Using this model, the authors successﬁilly simulated performance
on the Stroop Color-Word Task, WCST, and a motor sequencing task which suggested
that the weakening of associations as a result of prefrontal damage provides an accurate
explanation of many of the performance deﬁcits observed following prefrontal injury,
including representational memory.

A number of ﬁndings question the adequacy of the representational memory model
of prefrontal function. Delayed-response deﬁcits are not observed in the absence of
external distracting stimuli or competing response alternatives (Fuster, 1998). West
(1996) suggests that if the simple passage of time and loss of access to representational

memory were suﬂicient to elicit the delayed-response deﬁcit, then it should be observed

11

under all conditions. Thus, a deﬁcit in representational memory does not sufficiently
account for all of the behavioral deﬁcits which arise after prefrontal insult. Some have
suggested that it an additional function of the prefrontal cortex which was previously
discussed, inhibition of prepotent responses, accounts for these additional performance
deﬁcits (Posner & Rothbart, 1991; Stuss & Benson, 1987).
Hierarchical Model of Prefrontal Cortex Function

Fuster (1998) suggested that the most characteristic ﬁmction of the preﬁ'ontal
cortex is temporal structuring of behavior. Accordingly, the prefrontal cortex forms a
temporal framework for any behavior that has a unifying goal or purpose in novel
situations. The most critical factor demanding the function of the prefrontal cortex is
time; to bridge temporal gaps. This temporal structuring of behavior is subserved by three
subordinate functions: provisional memory, prospective memory, and interference control
(Fuster, 1998). The two memory processes within this model represent a division of
Goldman-Rakic's (1987) representational memory aspect of preﬁontal cortex function.
The ﬁrst, provisional memory, is similar to the idea of short-term or primary memory
(West, 1996). Provisional memory allows referencing of some event/action to past
experiences or to some ﬁxture goal/action sequence. Information, which may be either
newly acquired or a reactivation of some existing memory trace, is held until a goal is
attained, providing a foundation on which the action sequence is constructed.

Prospective memory includes anticipation, foresight, or preparatory set. It
functions to operate on information derived from the provisional process to prepare the
individual for some impending response (West, 1996). The prefrontal cortex, using data
from the provisional memory process, prepares for anticipated events on the way to
achievement of goals.

These two types of prefrontal memory processes are distinguishable in reports of
the frequently observed dissociation of self-consciousness and self-knowledge with

prefrontal damage. The ability to verbalize the requirements of a task is preserved

12

(provisional memory), however the task-relevant knowledge cannot be applied in the
direction of behavior (prospective memory) (F uster, 1998). The prefrontal cortex is not
assumed to store the actual representation used to guide behavior (the outcome of the
provisional memory process) but rather supports the maintenance of this representation
stored in temporal and parietal cortices through an interaction with sub'cortical structures
(West, 1996). However, prospective memory is theorized to be highly dependent on the
prefrontal cortex, making it more susceptible to prefrontal insult.

The ﬁnal component of prefrontal ﬁmctioning according to this model is
interference control. Fuster (1998) described this as the suppression of inﬂuence of
potentially interfering stimuli existing in the external enviromnent or arising from internal
sources. Impaired interference control leads to a distortion of the temporal gestalt.
Notably, he theorized that this process was localized within the orbitomedial prefrontal
cortex.

In summary, Fuster’s model suggests that temporal structuring of behavior is the
primary role of the prefrontal cortex. It is supported by three subordinate functions which
allow for the formation and execution of temporal gestalts or complex behavioral
sequences that are novel to the organism. This theory was one of the ﬁrst to address the
issue of unity and diversity of the prefrontal cortex and to stress the importance of the
temporal integration of behavior (Stuss & Benson, 1986).

Specialization of Function Within the Preﬁ'ongl Cortex

The prefrontal cortex is generally divided into three distinct areas: dorsolateral,
orbitomedial, and anterior cingulate. Many have speculated that various functions are
localizable to these given regions. For instance, Goldman-Rakic's (1987) work with non-
human primates has suggested that representational memory is localized to the principle
sulcus and inferior convexity. Much debate exists about the localization of inhibitory
control processes. Some have suggested that inhibitory control, as a memory-related

process, is localizable within the dorsolateral region (Diamond, 1990) whereas others have

13

suggested it is a function of the orbitomedial area (F uster, 1998). Goldman-Rakic (1987)
has proposed that both inhibitory control and memory processes are distributed across the
substructures within both of the divisions of the prefrontal cortex. Most recently, the
anterior cingulate has been shown to be involved in preparation for an anticipated event by
inhibiting automatic responses to allow for less automatic responses to be selected
(Cabeza and Nyberg, 1997; Posner & Raichle, 1996). This effect was demonstrated most
strongly in reviewed studies that employed the Stroop task.

These ﬁndings are supported by neuroirnaging studies conducted with healthy
adults. Speciﬁcally, performance on tests of working memory has been associated with
disproportionately increased activity (as measures by rCBF, MRI, and PET methods) in
the dorsolateral prefrontal cortex (Smith, Jonides, Marshuetz, & Koeppe, 1998; Barch et
al., 1997; Cabeza & Nyberg, 1997; Raglund et al., 1997; Posner & Raichle, 1996;
Schumacher et al., 1996). Increased activity in the orbitomedial preﬁ'ontal cortex has been
associated with performance on measures of sustained attention and interference control
(Cabeza & Nyberg, 1997; Malloy, Bihrle, Duffy, & Cirnino, 1993). Performance on tasks
of selective attention and ability to inhibit prepotent responses has been associated with
increased activity in the anterior cingulate (Cabeza & Nyberg, 1997 ; Posner & Raichle,
1996). However, much of the research investigating specialization of function within the
prefrontal cortex has been conducted with individuals and animals that have focal lesions
or with healthy, young adults. Are these same functional declines evidenced in more
common cases of global or generalized prefrontal damage, such as in the case of head
injury or aging processes?

The Relationship Between PrefrontaL and Non-Prefrongl Brﬁt m

Because of the interconnectedness of the prefrontal region to other brain regions,
it is important to ask how preﬁontal regions interact with the rest of the brain in
accomplishing the task of guiding purposeﬁil behavior. Any theory of prefrontal function

must address the separation of function between prefrontal and other brain regions. This is

14

a daunting task, as the frontal lobes are one of the most highly interconnected areas of the
brain to both cortical and subcortical brain regions. However, evidence exists which
suggests a dissociation of some types of memory functions that are ascribed to damage of
the prefrontal cortex versus damage to other brain regions, primarily the medial temporal
cortex and hippocampus (West, 1996). The pattern of errors made on memory tasks
following prefrontal and hippocampal lesions is markedly different in both humans and
non-human primates (Diamond, 1990). Those with prefrontal lesions demonstrate a
characteristic pattern of returning to a previously rewarded, prepotent response. In
contrast, those with hippocampal lesions make errors that are unrelated to previous
rewarded responses. A number of studies have demonstrated that persons with preﬁ'ontal
damage outperform those with medial temporal damage on tasks of paired associate
learning, story recall, and diagram recall. In contrast, on memory tasks in which structure
is not inherent in the material and must be supplied by the person, such as the California
Verbal Learning Test (CVLT; Delis, Kramer, Kaplan, & Ober, 1984), individuals with
preﬁ'ontal damage performed signiﬁcantly worse (Schacter, 1987; Shimamura, 1995;
Stuss, 1991). Moscovitch (1994) investigated cognitive resources and dual-task
interference effects at retrieval in healthy older adults, concluding that strategic memory
functions which are more resource demanding were mediated by the prefrontal cortex and
dissociated from episodic memory functions involving the medial temporal lobes and
hippocampus. More recently, Winocur, Moscovitch, and Stuss (1996) demonstrated that
explicit memory was reliant upon medial temporal functioning and was dissociated from
implicit memory, which was reliant on prefrontal cortex ﬁanctioning.

Although evidence exists which supports the notion that clear dissociations can be
made between prefrontal cortex functioning and cognitive process supported by other
brain regions, a number of ﬁndings suggest that localization may not be as clearly deﬁned.
Work with both animals and humans suggests that damage to the medial temporal area can

result in ﬁmctional impairments similar to those arising from prefrontal insult (West,

15

1996). Incisa della Rocchetta and Milner (1993) demonstrated that persons with
hippocampal excisions demonstrated similar impairment of free recall in unstructured
situations to individuals with left-frontal excisions. Similarly, Corcoran and Upton (1993)
demonstrated that patients with hippocampal damage performed more poorly on the
WCST than those with frontal damage. Winocur (1988) suggested that removal of the
hippocampus can result in an increased susceptibility to interference in learning paradigms.
Kopelrnan and Stanhope (1997) found that patients with temporal lobe or diencephalic
lesions demonstrated major deﬁcits initially acquiring information and had faster long-term
forgetting rates when compared to those with frontal lobe lesions. Finally, West's (1996)
review of the literature on source memory led him to conclude that medial temporal insult
also results in disruptions of memory for context.

Furthermore, the three major divisions of the prefrontal cortex are highly
connected with subcortical regions; thus executive function is supported by an integrated
and synergistic system arising from parallel prefrontal-subcortical circuits. These three
parallel frontostriatal circuits underlie executive function and emotion (Denckla & Reiss,
1997). Particularly, lesions in the dorsolateral - dorsolateral caudate - globus pallidus -
thalamic circuit result in executive dysfunction as evidenced by impaired working memory.
Lesions in the orbitofrontal - ventromedial caudate - globus pallidus - thalamic circuit
result in irritability and disinhibition. Finally, lesions to the anterior cingulate - nucleus
accumbens - globus pallidus - thalamic circuit result in inertia and apathy. As such, it is
difﬁcult to speciﬁcally ascribe any differential neuropsychological functions speciﬁcally to
neuroanatomical substrates.

The Prefrontal Cortex Theory of Cognitive Aging

As previously described, a distinct set of cognitive deﬁcits is associated with
prefrontal cortex damage. These deﬁcits must be accounted for by any theory of
prefrontal cortex function. Although previously reviewed theories accounted for portions

of the deﬁcits observed after damage to the prefrontal cortex, none were able to account

16

for all of the following: fragmentation of complex sequential behavior, increased
distractibility and perseveration, and insensitivity to external stimuli which leads to an
inability to modify behavior based on verbal instructions, to differentiate novelty from
familiar, and to detect temporal characteristics of information retained in the environment.
West (1996) proposed an integrative model of prefrontal cortex ﬁinction, incorporating
ideas from previously reviewed theories. His theory states that the function of the
prefrontal cortex is organized in to a hierarchy of interactive processes. Based upon
Fuster's work (1998), this theory describes that at a general level, the prefrontal cortex
supports the integration, formation, and execution of complex, novel behavioral structures
or temporal gestalts. This allows behavior to be directed in an orderly, purposeﬁrl manner
(West, 1996). This integrative ﬁmction is supported by four secondary processes as
outlined in Table 1: provisional/retrospective memory, prospective memory, interference
control, and inhibition of prepotent responses (Diamond, 1990; Fuster, 1989; Goldman-
Rakic, 1987). The retrospective memory process maintains on-line, task-relevant
information on which the temporal gestalt is constructed. The prospective memory
process serves to prepare the person for the execution of the temporal gestalt in response
to appropriate external or internal cues. The interference control process serves to clear
current task inappropriate information from the retrospective memory store. The
inhibition of prepotent responses process keeps behavior from being captured by dominant
internal response patterns or highly salient environmental stimuli. The action and
interaction of these four processes support a wide range of cognitive activities, including

the selection of task-relevant information, the integration of semantic and contextual

l7

Table 1

Four Secondary Processes of West’s (1996) Prefrontal Cortex Theory of Cognitive Aging

 

Term

Retrospective memory

Prospective memory

Interference control

Inhibition of prepotent
responses

m

maintains on-line, task relevant
information on which a time-line
for action is constructed

prepares person to execute time-
line action sequence in response
to some intemal/external cue

clears action-irrelevant information

from the retrospective memory
store

prevents behavior from being
“captured” by a dominant internal
response pattern or a highly salient
environmental stimuli

Example

Recalling ingredients and steps for
a recipe.

Remembering what ingredients to
add when the timer you previously
set goes oﬂ".

“Clearing” the steps in the recipe
you have completed from your
mind, as well as information not
related to the task (i.e.,
ingredients/steps from the previous
recipe you completed).

Not becoming distracted by the
television or thoughts about work
that might lead to adding the same
ingredient twice or adding the
wrong ingredient.

 

information into discrete memory traces, and the planning and execution of complex,

purposeﬁil behavior (West, 1996).

In order for a model or theory to be of value, it must accurately depict ﬁndings

from a wide range of empirical work. It should accurately predict when age-related

deﬁcits should be observed both at general and speciﬁc levels. A theory of cognitive aging

should be able to account for the types of tasks that should and should not demonstrate

evidence of age-related decline, as well as be able to distinguish those task conditions that

do and do not produce age-related deﬁcits (West, 1996). Furthermore, it should be able

to differentiate age-related declines which are the result of ﬁrnctional impairment of the

18

prefrontal cortex versus other brain structures or some general process, such as age-
related reduced processing speed (Salthouse, F ristoe, & Rhee, 1996).

With this in mind, under which conditions would age-related deﬁcits be expected
to arise? Mth respect to provisional or retrospective memory, deﬁcits would be expected
when performance on a task becomes dependent on the retrieval of contextual information
about an event and less related to the semantic content of memory. Prospective memory
deﬁcits would be expected to occur when task performance cannot be supported by
external cues or reminders. Finally, as interference control and inhibition of prepotent
response processes are compromised by advancing age, this model suggests that observed
behavior will become increasingly characterized by a tendency to respond to current task-
inappropriate information or based on a dominant behavioral pattern (West, 1996). In the
following sections, evidence from the following paradigms is reviewed in light of the
proposed model of prefrontal cortex function: prospective memory, frequency estimation,
source memory, interference control and sustained attention, inhibition of prepotent
responses, and working memory.

P_r_o_spective Memclv and Frequency EﬂimLﬁon

The effects of advancing age on memory for activities to perform, or prospective
memory, are somewhat limited (West, 1996). Prospective memory differs from the kind
of memory that is dependent on the integrity of the medial temporal lobe or diencephalic
structures in that the former is needed to manipulate and organize memory, whereas the
latter is needed to store and consolidate new memories (Shimamura, Janowsky, & Squire,
1991 ). Cues to elicit a prospective action can be either event- or time-based. The
prefrontal cortex model of aging predicts that time-based memory tasks should
demonstrate greater age-related decline than event-based tasks because time-based tasks
are presumably devoid of external cues that support prospective action. McDaniel and
Einstein (1992) found that older adults were less accurate than younger adults on tasks

involving time- but not event-based cues to elicit the prospective action. However, in a

19

study examining the effects of increasing the delay between receiving the prospective
instructions and the cue to perform the prospective actions, Einstein Holland, McDaniel,
and Guynn (1992) found that performance was not affected and that this increase in delay
did not interact with age.

The complexity of the prospective cue or action has also been investigated. Older
adults demonstrated impaired performance relative to younger adults when a task required
monitoring of four prospective cues, which was in contrast to the similar performance of
both groups when a task required the monitoring of only one cue (Einstein et al., 1992).

A ﬁnal way in which prospective memory performance has been examined is in
studies of the use of external memory aids. Overall, these studies have demonstrated that
performance on a variety of memory tasks, including prospective tasks, is enhanced by the
use of an external memory aid (West, 1996). What has yet to be determined is the role
that advancing age plays in the use and effects of external aids during prospective memory
tasks. Older adults may rely more on memory aids as task difﬁculty increases or it may be
that older adults would not use memory aids as effectively as task demands increase. The
preﬁ'ontal cortex theory of cognitive aging would predict that prospective memory
performance would be enhanced by using external aids because part of the role of the
prefrontal cortex would be replaced by the external aid.

Frequency estimation has received considerable attention in ﬁeld of cognitive
aging. A typical frequency estimation task requires the studying of a list of items over a
period of time and then providing a judgment about the frequency of the studied stimuli
during a test phase. Hasher and lacks (1979) proposed that encoding of frequency
information was an automatic process and thus would not be inﬂuenced by the eﬂ‘ects of
advancing age. Aging research has both supported and questioned the age insensitivity of
the encoding of frequency information. Older and younger adults have been found to
make comparable estimates at low frequencies of occurrence; however at high frequencies

of occurrence, the performance of older adults was impaired (Kausler, Salthouse, &

20

Saults, 1987; Hasher & Zacks, 1979). The current model of prefrontal ﬁinction would
suggest that this deﬁcit results from an inability to form discrete context-dependent
memory traces of the study stimuli with increasing frequency of occurrence or to update
existing representations (West, 1996).

Target familiarity has also been shown to differentially effect frequency estimation
performance in younger and older adults. Wiggs (1993) found that the performance of
younger and older adults was comparable in their accuracy of estimation for English
words but that older adults were signiﬁcantly impaired in estimation when Japanese
ideograrns were used as stimuli. This ﬁnding is in agreement with existing ideas of
prefrontal function, as older adults, like persons with frontal lobe lesions, may experience
difficulty when novel stimuli must be incorporated into memory. Overall, West's (1996)
review of the frequency estimation literature led him to conclude that existing instances of
age invariance in frequency estimation appear either to be predicted by the preﬁ'ontal
cortex model or to have resulted from methodological problems.

Source Memory

Numerous studies of episodic memory and source amnesia have suggested that
older adults are impaired relative to younger adults in their ability to remember the source
of newly acquired information. Reduced source memory has been attributable to
decreased effectiveness in the ability to integrate events with their temporal context in
memory (McIntyre & Craik, 1987). Evaluating source memory requires an individual to
recall the context in which some fact was learned or action performed. Variables that
appear to inﬂuence performance on source memory tasks include the length of time
between initial exposure to the material and recall of source information and the
discriminable salience of informational sources (West, 1996).

McIntyre and Craik (1987) demonstrated that older adults (M_= 69.2 years old)
performed more poorly on a source memory task which required learning of both real and

ﬁctitious facts about famous, non-famous, and ﬁctitious people when compared to

2]

younger adults (M = 19.4 years old). This effect was demonstrated after 10 minutes and
even more strongly after one week. Notably, the decrease in source memory with
increasing delay was accompanied by a decrease in item recognition memory for older
adults at the one-week follow-up; thus, poor performance of the older adults at this time
point may have been partly attributable to reduced memory overall. Craik et al. (1990)
found that the degree of source amnesia in a sample of healthy adults was signiﬁcantly
related to age, verbal ﬂuency, and performance on the WC ST, leading to the conclusion
that impairments resulting ﬁ'om the normal aging process are related to functioning of the
preﬁ'ontal cortex. Janowsky, Shimamura, and Squire (1989) found that the ﬁ'equency of
source errors committed by healthy older adults (M = 63.9 years old) and patients with
frontal lobe lesions were similar and exceeded the number committed by healthy younger
adults (M = 43.9 years old). In these latter two studies, decreased source memory was
not accompanied by or related to reduced item recognition. Parkin and Walter (1992)
found that explicit verbal recognition, deﬁned by having participants indicate that their
answer was based on "remembering" the source of the information, declined with age and
familiarity-based recognition, deﬁned by having participants indicate that their answer was
based on information that they "knew", increased. Furthermore, the extent to which older
participants relied on familiarity-based recognition correlated with reduced performance
on neuropsychological indices of frontal dysfunction.

Evidence for the effect of increasing delay on source memory appears to be mixed.
While some studies have demonstrated greater source amnesia with increased delays
(Spencer & Raz, 1994; McIntyre & Craik, 1987), other have found no such effects
(Schacter et al., 1991). West (1996) concluded that the effect of increasing delay on
source memory appears to be tied to a concurrent reduction in overall recall but that
source memory difference between younger and older adults persist even when no item

recall deﬁcit is observed.

22

Another factor proposed to inﬂuence the effect of advancing age on source
memory is the focus of attention within the encoding phase of the task. Schacter et al.
(1994) found that directing attention to the source or fact information during the encoding
phase of an experiment did not produce difference in source memory for younger or older
adults. Thus, age-related differences in source memory do not appear to be the result of
differences in the ability of younger and older adults to attend to information.

In conclusion, the current model of prefrontal ﬁmction would suggest that the
source memory deﬁcits observed in older adults result from the impoverished integration
of content and context information at encoding and possibly recall (West, 1996). Thus, it
could be hypothesized that on tasks of verbal memory recall and recognition, which
involves the presentation of two separate lists of related items to remember, such as the
CVLT, advancing age should result in decreased recognition discrirninability and increased
incidence of intrusion errors speciﬁc to the related lists (i.e., items presented on the ﬁrst
list being reported during recall of the second list and items from the second list being
reported during a delayed recall of the ﬁrst list).

Interference Con_trol and Sustained Atten_tio_n

The study of interference control has most commonly employed the Brown-
Peterson Consonant Trigram Task or tasks on which performance involves release from
proactive interference. Researchers examining age-related performance declines on the
Brown-Peterson task have generally indicated that older adults tend to demonstrate poorer
recall than younger adults and that this age-related deﬁcit does not interact with the length
of the retention interval (West, 1996). This is consistent with ﬁndings ﬁ'om the study of
the effects of advancing age on working memory, which suggest that increasing age is
associated with a decline in the number of items maintained on-line but that this
information is relatively stable across time (Salthouse, 1994).

Older adults generally demonstrate greater buildup of interference over trials
(Dobbs, Aubrey, & Rule, 1989) which can be attributable to a decreased ability to clear

23

information from working memory, a deﬁcit of the interference control process.
Shimarnura and Jurica (1994) demonstrated that advancing age was associated with
heightened proactive interference due to previously presented stimuli during an
investigation using a self-ordered pointing task of prefrontal cortex ﬁmction. Detection
and control of buildup of interference during an implicit stem completion task was found
to be related to performance on tasks of prefrontal cortex function and declined with
advancing age (Nyberg, Winocur, & Moscovitch, 1997). Dobbs et al.'s (1989) review of
existing literature in the area of release from proactive interference led to the conclusion
that when an appropriate measure of release (the Wickens ratio) is used, older adults
demonstrated reduced release from proactive interference in each of the ﬁve studies
reviewed.

Investigation of the evidence of declines in sustained attentional ability or vigilance
with age yields mixed results. Chao and Knight (1997) used event-related potentials and
behavioral measures to compare the performance of healthy young and old adults on two
different auditory delayed match-to-sample tasks. Older adults had reduced attention-
related activity over frontal regions and distracting stimuli elicited an enhanced primary
auditory evoked response in older adults. The authors concluded that increased
distractibility and impaired sustained atttention in older adults may be due to altered
prefrontal cortex functioning. Optimal performance on a sustained attention task is a
ﬁmction of the ability to focus over time, discriminate between target and nontarget
stimuli, maintain preparatory set, and respond quickly and accurately (West, 1996). The
prefrontal cortex has been implicated in the support of each of these skills (Fuster, 1980;
Posner & Rothbart, 1991).

Age-related decline in target hit rate was observed by Parasurarnan and Giarnbra
(1991), who concluded that older adults demonstrated faster rates of attention decrease
than younger adults while completing assessed tasks. This age-related decline was not

attenuated with extensive practice and was exacerbated by increased event rate. Age-

24

related decline on the Paced Auditory Serial Addition Test (PASAT; Gronwell &
Sampson, 1974; Gronwell & Wrightson, 1974), a task assessing sustained attention and
rate of information processing, has been consistently reported (O'Donnell et al., 1994;
Brittain et al., 1991; Stuss, Stethem, & Poirier, 1987). On auditory delayed match—to-
sample tasks, older adults were signiﬁcantly more impaired than younger adults by
distractors during long delays (Chao & Knight, 1997). Furthermore, during this task,
older adults demonstrated signiﬁcantly reduced attention-related activity over frontal
regions, as measured by event-related potentials. However, preserved sustained attention
capacity in older adults has been reported in studies using tasks that place fewer demands
on a potentially limited capacity attentional mechanism (West, 1996). Overall, the data
seem to suggest that age-related deﬁcits in sustained attention will arise under task
conditions which place a high demand on a limited capacity attentional system.

In conclusion, the effects of age on the performance on tasks sensitive to the
effects of interference in the short-term retention of material and the ability to sustain
attention over time can all be interpreted within the current theoretical framework of
prefrontal cortex ﬁinction (West, 1996). Age-related decline on measures of interference
control and sustained attention represent difﬁculties with controlling the interfering effects
of internal and external stimuli, contributing to difﬁculties in the formation of a temporal
gestalt and action sequence.

Inﬁtaition of Prepotent Response

Research utilizing the Stroop Color-Word Naming Task has provided numerous
ﬁndings related to the effects of advancing age on the ability to suppress a prepotent
response. Older adults consistently demonstrate slowed performance during the color-
word trial and increased interference scores (Wecker et al., 2000; Hanninen et al., 1997;
Uttl & Graf, 1997; West, 1996; Uchiyama et al., 1994; LaRue, 1992; Stuss & Benson,
1986). Houx et al.'s (1993) cross-sectional lifespan study indicated that age-related

increases in these eﬂ'ects begin to appear in the 6th and 7th decades of life and continue to

25

increase with age. Additionally, health status has been shown to interact with the effects
of advancing age, demonstrating an accelerated increase with age in the Stroop effect for
persons with even relatively minor health problems (West, 1996).

A number of explanations have been offered to explain the Stroop effect with
advancing age. Dulaney and Rogers (1994) have suggested that the inability on the part
of older adults to use a reading suppression response is the reason for age-related
increases in the Stroop effect. Panek, Rush, and Slade (1983) theorized that older adults
demonstrate greater Stroop effects due to an increase in response dominance for word
reading over color naming with advancing age. Notably, age-related diﬂerences in the
Stroop effect persist after controlling for the effect of decreased processing speed (Spieler,
Balota, & Faust, 1996). West (1996) concluded that none of these theories has been able
to provide a ﬁrll account of age—related increased in the Stroop effect, leaving open the
possibility that at least part of this effect is attributable to a reduced efficiency of the
inhibition of prepotent response processes supported by the prefrontal cortex.

The ability to inhibit a prepotent response keeps highly salient but unsuited
responses from gaining control of a behavior or action sequence (Dempster, 1991). As
such, perseverative errors can be thought of as a break down in this inhibitory process.
Perseverative errors represent an action sequence which is "captured” by one of its
constituent parts which may have been previously rewarded, as is the case with
perseverative errors on the WC ST. Thus, measures of perseverative errors provide
another way of analyzing the effects of increasing age on the ability to suppress a
prepotent response. It could be hypothesized that advancing age would be associated with
increased frequency of perseverative errors on tasks requiring the production of responses
to various task demands such as free and cued recall, design ﬂuency, and the WCST, due
to decreased ability to suppress prepotent responses.

Working Memog

26

West (1996) used the term retrospective memory to describe the process which
serves to maintain on-line, task-relevant information on which the temporal gestalt is
constructed. When combined with his deﬁnition of prospective memory, these two
constructs are more commonly known as working memory. Working memory has been
described as knowing what to do, when to do, and how to do it (Pribram, 1997). These
three factors are heavily dependent on the integrity of prefrontal cortex function. The
"what" is related to familiarity, a particular kind of memory. The "when" is a result of
sustained attention and the ability to discern when automatic processes become ineffective.
The "how" involves determining what is appropriate; when something works and does not
work. Thus, decreased working memory performance as a result of preﬁ'ontal disturbance
means that under certain conditions intentional and attentional processes needed for the
completion of learning tasks are disrupted (Pribram, 1997).

Decline in working memory performance with advancing age has been consistently
demonstrated. Libon et al. (1994) demonstrated that working memory performance, as
measured by overall performance on the WCST, as well as performance on other tasks of
executive functioning, declined with advancing age, with an old-old (75 years and older)
group of healthy volunteers demonstrating poorest performance. Hanninen et al. (1997)
demonstrated that even when adjusting for the effects of education, older adults with age-
associated memory impairment demonstrated signﬁcantly worse performance on the
WC ST and Trail Making Tests when compared to age-matched healthy controls. In a
study of short-term memory and frontal dysfunction, Parkin and Walter (1991) found that
performance on a working memory task (W C ST) declined with age and was related to
performance on the Brown-Peterson task. WC ST performance by a group of healthy 80
to 87 year old adults versus a group of healthy 64 to 69 year old adults revealed signiﬁcant
differences for the number of categories attained and total number of errors committed.
Hultsch et al. (1992) examined changes in mean performance on working memory,

information processing, and intellectual ability tasks over a three year period in a sample of

27

healthy community-dwelling older adults. Results showed signiﬁcant average decline on
working memory, verbal ﬂuency, and world knowledge. A step-down analysis revealed
that covarying decline in speed of information processing did not eliminate these
signiﬁcant declines. Similarly, age-related declines in WCST performance, described as
being reminiscent of frontal lobe and/or basal ganglia damage, were demonstrated in a
sample of healthy volunteers even after the effects of speed of information processing
were controlled for (Robbins et al., 1998). In contrast, Salthouse (1994) summarized a
number of his collaborative studies conducted on age effects of working memory. Results
consistently demonstrated that age was associated with signiﬁcant amounts of the variance
in computation and reading span performance. However, many of the age-related
inﬂuences were mediated by slower speed of information processing. More recent studies
reviewed demonstrated that slower processing primarily inﬂuenced the time required to
achieve a stable encoding of the material rather than the rate at which material was lost
across time or subsequent processing. Similarly, F ristoe, Salthouse, and Woodard (1997)
demonstrated that age-related reductions in processing speed mediated age diﬂ‘erences in
working memory performance and feedback-usage on the WCST. Hartley et al. (2001)
compared the performances of younger and older (mean age = 72.9 years) on a series of
working memory tests. They concluded that there is generalized decline in working
memory systems with advancing age, with the age differences being mediated, to a
moderate extent, by age-related differences in processing speed. Thus, age-related decline
in working memory performance has been strongly demonstrated. However, evidence is
mixed as to the role of speed of information processing in this decline. The current model
of prefrontal cortex function would suggest that the working memory deﬁcits observed in
older adults result from difficulty maintaining on-line, task-relevant information and
impaired ability to prepare for execution of the temporal gestalt in response to appropriate
cues, all of which are inﬂuenced the integrity of interference control mechanisms. It could

be that decreased processing speed is a result of ineffective, inefficient interference control

28

mechanisms; if inhibitory control processes cannot serve to limit the access of currently
task-irrelevant information to current processing efforts, new behavioral sequences cannot
be planned and executed as quickly. Since inhibitory control ﬁrnctions serve a dual
purpose of preparing an individual to respond to current task demands based on event-
speciﬁc information and to separate events into distinct temporal units, break down of
these ﬁrnctions can be expected to result in decreased speed of information processing, as
well as an increased frequency of errors in information processing.

Many studies have examined the relationship between frontal lobe functioning and
various aspects of verbal memory. A number of these studies provide ﬁrrther evidence
demonstrating decline in working memory and other prefrontal abilities with advancing
age. Parkin and Walter's (1992) investigation of recollective experience and age
demonstrated that familiarity-based recognition increased with age (in contrast to source-
based recognition, which declined with age) and was correlated with neuropsychological
indices of frontal lobe dysfunction. These authors used the WCST and a verbal ﬂuency
test as two of their measures of frontal lobe dysfunction and found that performance on
both declined with advancing age. In their investigation of changes in cerebral ﬁrnctioning
with normal aging, Mittenberg et al. (1989) concluded that predominant age-related
neuropsychological changes occur in frontal rather than right-hemisphere abilities.
Neuropsychological measures of frontal ﬁrnction included verbal and design ﬂuency, as
well as verbal and visual recency discrimination. Levine, Stuss, and Milberg (1997) used
conditional associative learning tasks to evaluate the hypothesis that age-related cognitive
decline is related to ﬁ'ontal dysﬁrnction. Results demonstrated that deﬁcits in conditional
associative learning were attributable to strategic rather than basic associative processes
and error analysis revealed that past information failed to guide behavior in both persons
with frontal lobe lesions and in a group of healthy, older adults (M = 71 years old).
Furthermore, these authors concluded that congruence between older adults and the

participants with dorsolateral prefrontal lesions on a measure of defective inhibitory

29

control suggested that age-related decline in inhibitory processes is due to prefrontal
dysfunction. A rCBF study examining the relationship between age and recognition
memory ability revealed that older adults (M = 69.4 years old) showed no signiﬁcant
activation in areas (right hippocampus and left prefrontal and temporal cortices) activated
during encoding in young adults (M = 25.2 years old) but did show right prefrontal
activation similar to younger adults during recognition (Grady et al., 1995).

In summary, all of these studies suggest decline in working memory, as well as
potentially other memory abilities, with advancing age. The prefrontal cortex theory of
cognitive aging posits that advancing age results in a breakdown of many of the processes
which support the integration, formation, and execution of complex, novel behavioral
structures. These processes include those related to working memory, as well as to
interference control. The former involves maintaining on-line, task-relevant information on
which the temporal gestalt is constructed, as well as preparation for execution of the
temporal gestalt in response to appropriate external or internal cues. The latter involves
clearing current task inappropriate information from memory stores and keeping behavior
from being captured by dominant internal response patterns or highly salient
environmental stimuli. Data from this proposed investigation of the prefrontal cortex
theory of cognitive aging are expected to demonstrate the existence of discernible
neuropsychological functions which are associated with prefrontal cortex function.
Performance on neuropsychological measures of prefrontal lobe ﬁrnctioning is expected to
decline with advancing age. Furthermore, this decline performance on prefrontal cortex
ﬁrnction measures with advancing age is expected to signiﬁcantly inﬂuence the
relationship between age and verbal memory performance.

Measurement and Alternative Theoretical Considerations

A number of methodological challenges exist when attempting to substantiate any

theory of cognitive functioning, as few investigative procedures are available which allow

for deﬁnitive localization of function and evaluation of speciﬁc cognitive processes

30

underlying neuropsychological test performance. It is often assumed that levels of
performance on different cognitive measures and the age-related effects on those measures
are determined by separate and distinct mechanisms. With many neuropsychological tests,
the discovery of selective impairments by individuals with damage in particular brain
regions has led to the inference that various brain structures are specialized for different
types of processing. When combined with results indicating age-related deﬁcits on those
measures, these ﬁndings have led to speculation that certain brain regions are more
sensitive to age-related decline than others (Salthouse, F ristoe, & Rhee, 1996). Although
certain measures may be sensitive to damage in particular brain regions, the measures may
not be speciﬁc because damage to other regions could lead to similar patterns of
impairment. Seidman et al. (1995) raised this issue in their investigation of experimental
and clinical neuropsychological measures of prefrontal dysﬁrnction in schizophrenia,
stating that frontal deﬁcits may be due to diffuse brain damage, circumscribed prefrontal
damage, and/or damage in other brain regions having prefrontal connections.
Furthermore, many of the neuroirnaging studies used to identify what brain regions are
activated during performance on various cognitive tasks contribute to these
methodological concerns. Often times " activation" of a speciﬁc brain region is equated
with excitatory activity, leading to the conclusion that the activated brain region is
responsible for the a certain ﬁrnction. However, alternative explanations must be
considered: 1) "activation" may represent inhibitory activity; and/or 2) "activation" may
represent a system of excitatory or inhibitory processes in which the end result is
activation of a speciﬁc brain region.

Additionally, neuropsychological measures used to associate cognitive
performances with certain brain regions often represent more than one ﬁrnctional process.
Salthouse, F ristoe, and Rhee (1996) investigated the localizability of age-related effects on
neuropsychological measures and found that they were not independent. Across all

variables examined, an average of 58% of the age-related variance in a given variable was

31

shared with that in other variables. The authors concluded that only a portion of the age-
related inﬂuences on many commonly used neuropsychological measures is speciﬁc and
potentially localizable. Delis, Squire, Bihrele, and Massman (1992) performed a
componential analysis of measures of problem-solving ability, determining that a wide
spectrum of deﬁcits in abstract thinking, cognitive ﬂexibility, and use of knowledge
contributed to problem-solving performance. This has been conﬁrmed by a number of
factor analytic studies of commonly used measures of frontal lobe ﬁrnctioning which have
demonstrated similar results (Greve et al., 1996; Sullivan et al., 1993).

An alternative theoretical consideration to the decline of cognitive performance
with advancing age is that decreased speed of information processing underlies the poorer
frontal lobe and memory performance observed at older ages (Salthouse, 1985). Slower
speed of processing has been shown to moderate the effects of decreased memory,
reasoning, and spatial, and ﬂuency performance with advancing age (F ristoe, Salthouse, &
Woodard, 1997; Fisk & Warr, 1996; Salthouse, Fristoe, & Rhee, 1996; Salthouse, 1994;
Lindenberger, Mayr, & Kliegl, 1993; Tun et al., 1992). However, Robbins et al. (1998)
demonstrated more support for a frontal lobe functioning versus processing speed decline
hypothesis of cognitive aging. Notably, measurement and "localization" of processing
speed is plagued by many of the same difﬁculties as the measurement of prefrontal cortex
ﬁrnctioning. Furthermore, although it has been assumed that processing speed indicates
the underlying rate at which information is activated within areas of cognitive functioning,
this interpretation is open to debate and may not constitute exclusive support of an

account of cognitive age deﬁcits in terms of information activation (Fisk & Warr, 1996).

Hypotheses
The purpose of this study is to investigate the applicability of the prefrontal cortex
theory to cognitive aging in a sample of healthy, older adults. The main premise of this

study is that the prefrontal cortex is critically important for the integration, formation, and

32

execution of complex, novel behavioral structures and temporal gestalts, which supports
the direction of behavior in an orderly, purposeﬁrl manner. This integrative function is
subserved by four secondary processes, identiﬁed by West (1996):
provisional/retrospective memory, prospective memory, interference control, and
inhibition of prepotent responses. Table 1 provides a summarized description of these
processes. For the purpose of this investigation, the concepts of provisional memory and
prospective memory will be considered together as the concept of "working memory".
The following hypotheses will be tested:

1) The three-factor model of prefrontal cortex function (working memory,
interference control, and inhibition of prepotent responses) will be analyzed using
conﬁrmatory factor analytic techniques (see Figure 1). The three-factor measurement
model of prefrontal cortex function will ﬁt better than a one- or two— factor model (see
Figure 2). In the case of a poor ﬁt, an exploratory factor analytic representation of

prefrontal cortex ﬁrnction variables will be generated and used for subsequent analyses.

 

A.

< Working Memory
V/
//\ Prefrontal Cortex

Interference Control L‘ Function

 

\

   
 

 

Figure 1. Three — Factor Model of Prefrontal Cortex Function

33

The factor representing the ability to inhibit prepotent responses is hypothesized to be
comprised of the Stroop color-word naming trial score, and perseverative errors scores
from the WCST, Ruff Figural Fluency Test, and California Verbal Learning Test (CVLT).
The factor representing working memory is hypothesized to be comprised of total scores
from the Digit Span (DSp) and Visual Memory Span (VMSp) subtests from the Wechsler
Memory Scale-Revised (WMS-R; Wechsler, 1987) and the WCST total categories
achieved score. The ﬁnal factor representing interference control is hypothesized to be
comprised of scores from the Trailmaking Test Parts A and B, the total score from the
Symbol Digit Modalities Test , and the CVLT proactive interference score.

 

l

k working Memory T\‘/
v/
, Prefrontal Cortex

/
( Interference Control

1

l
\ /
v/

 

Figure 2. Two — Factor Model of Prefrontal Cortex Function

2) Causal models will be tested in which prefrontal cortex ﬁrnction will mediate the
association between age and verbal memory measures (see Figure 3), with decline in
preﬁ'ontal cortex ﬁrnctioning mediating the decline in verbal memory performance with

advancing age. Because verbal memory performance is often chosen as a key measure in

34

evaluating changes with advancing age because of its direct, pragmatic application to daily

ﬁrnctioning, it was utilized in this study to investigate the hypothesized mediating role of

prefrontal cortex ﬁrnction on cognitive aging.

 

     
 

Prefrontal Cortex
Funcﬁon

Age r—ﬁ '/

Verbal Memory

 

Figure 3.

Participants

Causal Model of Prefrontal Cortex Function as a Mediator

METHOD

One hundred nine community dwelling, independently-living older adults were

drawn from a larger intervention study that recruited participants from the greater

Lansing, Michigan area through newspaper advertisements. Those who scored at 24 or

better on the Mini Mental Status Examination and reported no signiﬁcant history of severe

neurological or medical problems likely to adversely affect their cognitive abilities (e.g.,

stroke, major ischemia, terminal illness, signiﬁcant traumatic brain injury) on a self-report

medical history questionnaire were included. The 60 women and 49 men ranged in age

35

from 55 to 88 years old (M = 70.2; _S_12 = 6.9) and had a mean education of 16.4 years (SD
= 2.9).
Procedure

Participants were assessed at MSU‘s Psychological Clinic. The entire assessment
for each participant, including measures unrelated to this study, required from two to two-
and-one-half hours. Participants were unaware of the hypotheses of this study when they
participated in the assessment. Assessments were conducted and scored by the author and
graduate level students enrolled in MSU‘s Clinical Psychology program who have
experience in administering and scoring these tests. The author and two other graduate
students specializing in neuropsychology re-checked the scoring on all tests. Upon
completing the assessment, participants were enrolled in a bi-weekly support group that
focused on instruction in mnemonic memory training strategies and either attention
retraining or relaxation techniques. Groups lasted four weeks, after which time
participants were re-administered the assessment battery and then received feedback about
their performance as compared to others their age. One test that required correct color
vision (Stroop Color Word Test) could not be completed by four male participants who
were determined to be color vision-deﬁcient.

Measures
M_easures of Prefrontal Cortex Functioniag

Msconsjn Card Sorting Task (WCST)

The WCST assesses the ability to generate and test hypotheses and to shift and
maintain cognitive sets. Described by many as a measure of working memory, it contains
128 cards which the participant attempts to discover three diﬂ°erent rules/categories for
sorting multidimensional stimuli by matching each card to one of four key cards.

Feedback is provided when cards are matched correctly or incorrectly. After the
participant matches 10 cards of a given category, the examiner changes the matching

category, without indicating this to the participant. Each category has to be discovered

36

twice. This test yields four measures: total categories achieved (WCSTcat); perseverative
errors involving continuation of sorting based on an immediately preceding rule
(W CSTpe); and nonperseverative errors (W CSTnp).

The WCST is widely used in clinical settings to detect cognitive impairments
associated with frontal lesions (Stuss & Benson, 1986). Spreen and Strauss (1991)
critiqued the WCST, reporting that studies have generally conﬁrmed that this test is
sensitive to frontal functioning, although some have reported more perseveration in
patients with right as compared to left prefrontal damage. Drewe's (1974) investigation of
patient's with schizophrenia who had undergone frontal leucotomy demonstrated
signiﬁcant difficulties with perseveration on the WC ST. Milner (1963) found that patients
with dorsolateral frontal excisions demonstrated signiﬁcantly poorer performance on this
task when compared to those with orbitomedial and posterior lesions. Speciﬁcally, those
with dorsolateral lesions showed an inability to shift from one sorting principal to another.
The work of both Taylor (1979) and Hermann et al. (1988) supported these ﬁndings and
suggested that the WCST is sensitive to ﬁrnction in dorsolateral areas of both ﬁ'ontal
lobes, but more to the left than to the right. Most recently, Ragland et al. (1997) found
that WC ST performance was associated with increased PET rCBF activity in the
dorsolateral preﬁontal cortex. Notably, participants with the highest WC ST scores
activated dorsolateral prefrontal and inferior frontal regions, exclusively.

Stroop Color and Word Test (Stroop)

The Stroop measures the ease with which a person can shift perceptual set to
conform to changing demands and suppress a habitual or prepotent response in favor of an
unusual one. This test yields three basic scores: word naming (Stroopw) determined by
the number of words ( red, green, or blue) read correctly in 45 seconds; color naming
(Stroopc) determined by the number of colors (red, green, or blue) named correctly in 45
seconds; color-word naming (Stroopcw) determined by the number of correct responses in

45 seconds. The Stroop-cw trial consists of color words printed in incongruous colored

37

ink (e. g., the word "blue" printed in red ink). The participant is asked to name the color of
ink in which the word is printed, requiring them to suppress a highly prepotent response
(reading the word). The extent to which the tendency to read the word intrudes upon the
color naming assignment, referred to as the interference effect, is calculated from the three
basic scores.

Regard (1981) reported that patients with frontal lobe damage performed
signiﬁcantly worse on the Stroop-cw when compared to patients with lesions in other
areas of the cortex. Uchiyarna et al. (1994) found Stroop-c and Stroop-cw score to be
two of the most sensitive measures for predicting frontal lobe ﬁrnctioning in both geriatric
and non-geriatric samples. Vendrell et al. (1995) demonstrated that the brain region most
related to errors on the Stroop-cw was the right prefrontal lateral cortex utilizing magnetic
resonance imaging (MRI). In their review of PET studies, Cabeza and Nyberg (1997)
found that the ability to select between competing processing alternatives on the basis of
some pre-existing internal, conscious plan, as evidence by Stroop performance, was
consistently related to increased activity in the anterior cingulate. Spreen and Strauss
(1991) reported a one month test-retest reliability for each page of the Stroop to be .90,
.83, and .91, respectively.

Paced Auditory Serial Addition Test (PASAT)

The PASAT (Brittain, LaMarche, Reeder, Roth, & Boll) is designed to measure
sustained attention and information processing abilities. A pre-recorded tape delivers a
random series of 61 numbers from 1 through 9. The respondent is instructed to add pairs
of numbers such that each number is added to the one immediately preceding it: the
second number is added to the ﬁrst, the third to the second, the fourth to the third, and so
on. The same 60 numbers are given in four different trials at differing presentation rates
(2.4, 2.0, 1.6, and 1.2 seconds). This test requires an individual to comprehend the

auditory input, respond verbally, inhibit encoding of one's own response while attending to

38

the next stimulus in a series, and perform at an externally determined pace (Spreen &
Strauss, 1991).

The PASAT has good internal consistency, with a split-half reliability of .96 (Egan,
1988). O'Donnell, et al.'s (1994) factor analytic study of a numerous neuropsychological
tests demonstrated that the PASAT was one of the measures that deﬁned an attentional
factor. Brittain et al. (1991) examined the effect of age, gender, race, IQ, and education
on PASAT performance in healthy adults aged 17-88. Age and IQ were found to
signiﬁcantly affect PASAT results. A mildly signiﬁcant effect was observed for gender as
well, with males performing slightly better than females. Schmidt, Trueblood, Merwin,
and Durham's (1994) factor analytic study of attention tests included the PASAT as one of
the measures of a global construct of attention.

Ruff Figaral Fluency Test (RUFF)

A measure of design ﬂuency, the RUFF (Ruff, Allen, Farrow, Niemann, & Wylie,
1994) assesses spontaneous ﬂexibility, the ready ﬂow of ideas or answers in a response to
a single questions. Requiring divergent thought and production, successful performance
on the RUFF also relies on self-monitoring, remembering, and creative imagination.
Participants are required to generate as many different designs as they can by connecting
dots in a speciﬁc conﬁguration within a one minute time limit. Five trials, each with a
diﬂ‘erent dot conﬁguration pattern or different background on which the dots are to be
connected, are given. Performances are scored for number of unique patterns
(productivity - RUFFdes) and for number of repetitions of patterns (perseverations -
RUFFper).

Age and education have been shown to affect RUFF productivity scores in
normative studies (Ruff et al., 1994). Productivity scores have also been shown to
discriminate mild from severe head trauma patients and both groups from normal control
participants (Lezak, 1995). Furthermore, perseverative responses increased from control

(5.8) to mildly injured (8.8) to severely injured (10.1).

39

Trailmaking Test A & B (TM'D

TMT (Reitan & Wolfson, 1985) assesses speed of information processing,
attention, mental ﬂexibility, and visuomotor tracking. It requires participants to join
randomly located numbers in numerical sequence (TMTa) and in alternate numerical-
alphabetical sequence (TMTb) as quickly as possible. Resulting scores are based on time
taken to successfully complete each task.

Normative studies have shown that performance times increase signiﬁcantly in
each succeeding decade (Stuss, Stethem, & Poirier, 1987). Drebing et al. (1994) found
the TMT to be one of the most sensitive and speciﬁc neuropsychological tests to detect
early cognitive decline in higher cognitively functioning older adults. Factor analytic
studies have demonstrated that TMT performance consistently loads on a factor generally
described as "attention" (O'Donnell et al., 1994; Schmidt et al., 1994; Uchiyama et al.,
1994). Scores are strongly inﬂuenced by level of education (Spreen & Strauss, 1991).
Retest reliability over a six month interval was reported as .98 for TMTa and .67 for
TMTb (Lezak, 1995).

gambol-Diiit Modalities Test (SDMT)

SDMT (Smith, 1973) is a measure of processing speed, sustained attention, and
visual-motor scanning. It requires the matching of numbers to symbols within a 90 second
time constraint. The score is based on the number of successfully matched symbol-number
pairs. This test is similar to the Wechsler Digit Symbol subtest except the matching is
inverse.

Smith (1973) found the SDMT to possess good predictive validityin dissociating
normal adults from those with brain-damage. A number of factor analytic studies have
included the SDMT in factors described as information processing and visuo-motor
scanning (Schmidt et al., 1994).

Coptrolled OralQWord Association Test LCOWAT)

40

A measure of verbal ﬂuency, the COWAT (Benton & Hamsher, 1983) also
assesses spontaneous ﬂexibility. Participants are asked to produce as many words as
possible that begin with a designated letter of the alphabet. Three trials are conducted
using the letters ”C", "F", and "L", with a one-minute time limit imposed for each trial.
The score is the sum of all admissible words for the three letters.

Frontal lesions, regardless of side, tend to depress ﬂuency scores, with left frontal
lesions resulting in lower word production than right frontal ones (Miceli et al., 1981,
Benton, 1968). Eslinger and Grattan (1993) demonstrated that COWAT performance, as
well as performance on other measures of spontaneous ﬂexibility, was markedly decreased
in persons with frontal lobe lesions. Martin, Wiggs, LaLonde, and Mack (1994)
demonstrated that verbal ﬂuency places greater demands on frontal lobe mediated
strategic search processes than on temporal lobe mediated semantic knowledge. A PET-
scan study (Parks et al., 1988) with normal volunteers indicated that verbal ﬂuency
activated frontal and bilateral temporal lobes. A rCBF study (Gourovitch et al., 2000)
with normal volunteers found that during a verbal ﬂuency task, the anterior cingulate, left
prefrontal regions, thalamus, and cerebellum were activated. Education and age have been
shown to inﬂuence COWAT performance. Retest reliability with adults after 19-24 days
was reported as .88. (Spreen & Strauss, 1991).

Wechgﬂer Memory Soﬁe-Revised Digit Span (DSp)

DSp (W echsler, 1987) assesses verbal/auditory attention, short-term memory,
and/or working memory. Aural presentation of digits made to participants must be
repeated back to the examiner in either a forward or backward sequence. DSp
performance has been shown to decrease with age and in some studies has been positively
associated with higher levels of education ( Lezak, 1995; Ryan, Lopez, & Paolo, 1996).
Shimamura, Janowsky, and Squire (1991) found that patients with frontal lobe lesions
exhibited signiﬁcant impairment on the DSp when compared to controls.

Wechsler Memory Scale-Revised Visual Memory Span (VMSp)

41

VMSp (Wechsler, 1987) is the visual-spatial analog to DSp. It assesses visual
attention, short-term memory, and/or working memory. Participants are required tap a
sequence of colored blocks in either the same forward or backward order as the examiner
has tapped. VMSp performance has demonstrated the same relationship to age and
education at DSp (Lezak, 1995).

Mgures of Verbal Learning and Memory

mom Verbal Learning Test (CVLT)

The California Verbal Learning Test (CVLT; Delis, Kramer, Kaplan, & Ober,
1984) involves learning lists of words presented in successive trials in the same sequence.
Its scoring system provides information on the quantity of material learned, as well as for:
rate of learning over several trials, the encoding strategy employed (semantic versus serial
clustering), the types of recall errors made, vulnerability of memory to time and
interference conditions, and the degree to which memory performance improves with
assisted recall cues.

Rosenbaum's (1984) reported that the CVLT is an extremely sensitive instrument
for detecting subtle memory deﬁcits; however, this high sensitivity comes at the expense
of more false-positive errors. Over a one-year interval, CVLT test-retest reliability
coefficients ranged from .26 (percent middle recall) to .79 (long-delay cued recall). Total
immediate recall of List A across the ﬁve trials correlated .66 with the WMS-R General
Memory Index and in the .60s with several other WMS-R variables (Delis, Kramer,
Kaplan, & Ober, 198 7). These authors also factor analyzed all CVLT intercorrelations
and found that its multiple indices identiﬁed theoretically meaningﬁrl factors consonant
with the constructs they were designed to measure. An additional factor analytic study
demonstrated that verbal memory, as measured by the CVLT, consisted of a number of
theoretically meaningﬁrl component factors reﬂecting learning strategy, acquisition rate,
serial position effect, discriminability, and learning interference (Delis, Freelan, Kramer, &
Kaplan, 1988). Multiple CVLT studies of selected neurological groups (chronic

42

alcoholics, Parkinson's, Huntington's, and Alzheimer’s disease) have yielded results
commensurate with each group's previously determined memory problems. Four
measures from the CVLT were used in analyses as verbal memory variables: CVLTtot —
free recall total from trials 1 through 5; CVLTsf — short-delay free recall; CVLTlf — long-
delay free recall; CVLTrec — recognition hits. Total perseverative errors (CVLTper) and
the proactive interference score (CVLTpi) were used as measures of prefrontal cortex
ﬁmctioning.

Logical Memog (LM)

Immediate recall total of stories A and B from the Wechsler Memory Scale-
Revised (WMS-R; Wechsler, 1987) represents the recall of two verbal passages
immediately after each is presented. Total recall scores for LM can range ﬁom 0 to 46,
with higher scores reﬂecting more information units recalled. It has an average test-retest
stability of .7 9 over 4—6 weeks (Bowden & Bell, 1992) and has consistently shown to load
on the verbal memory factor in numerous factor analyses (Prigatano, 1978). It is strongly
correlated with both CVLTtot (.79) and CVLTlf (.93; Spreen & Strauss, 1998). Clinical
investigations using LM found it sensitive to memory disturbances in patient groups
including older adults with age-associated memory impairment, dementia, multiple
sclerosis, neurotoxin exposure, and depression (Spreen & Strauss, 1998).

Miasure of Mental Status

 

The Mini-Mental Status Exam (MMSE) is a 30-item test created to brieﬂy screen
for level of cognitive ﬁrnctioning. It assess orientation, immediate recall, attention,
calculation, language abilities, and constructional dyspraxia. Scores can range from 0,
indicating severe impairment to 30, indicating unimpaired mental status. It was
standardized on a sample of healthy older controls M = 74 years) whose scores ranges
ﬁ'om 24.6 to 27.6 (Folstein, F olstein, & McHugh, 1975). Demented persons scored
between 9.6 to 12.2, with no overlap between these two groups. A cut-off score of 24 is

often used to identify scores more suggestive of impairment (Lezak, 1995).

43

Me_asure of Depression

The 30 yes-no Geriatric Depression Scale (GDS) items assesses mood and
psychological symptoms of depression. Respondents answer each question according to
how they feel at the time of administration. GDS scores range from 0, indicating no
depressive statements were endorsed, to 30, indicating severe depression. The item-total
correlations ranged ﬁ'om .32 to .83 (M = .56), while internal consistency and split-half
reliability were both .94. The test-retest reliability over one month was .85 (Koenig,
Meador, Cohen, & Blazer, 1988).

Factor analysis of the GDS identiﬁed a major factor of general depression and
minor factors of worry/obsessive thoughts, and apathy/withdrawal (Parmalee et al., 1989).
Criterion validity with the Research Diagnostic Criteria was found to be .82. Concurrent
validity with the Hamilton and lung Depression Scales appear good (.83 and .86,
respectively; Yesavage et al., 1986). Stebbins and Hopp (1990) reported that the GDS
successﬁrlly discriminated between mildly demented depressed, and non-depressed

individuals.
RESULTS

Missing Data

Precautionary steps during data collection were taken to ensure a complete data
set. This included making alternative arrangements and multiple attempts to ﬁnish data
collection if a participant indicated they wished to discontinue with the testing. Following
data collection, if data were found missing, information was obtained ﬁom the clinician
who conducted the testing to determine the reason(s) for the missing data. Data from
seven participants were eliminated from analyses. These included three participants who
discontinued testing before the delayed recall trials could be administered on both verbal

memory measures and four male participants who were color vision-deﬁcient and unable

44

to complete the Stroop. Thus, the total number of participants used for all analyses was
102. No data were missing for any of these 102 participants.
Preliminary Data Examination

In order to examine the relationship between variables, Pearson product-moment
coefficients were calculated for all measures used in the study. In examining the
relationship between prefrontal cortex ﬁrnction measures, age, and education (see
Appendix A), the correlations between age and prefrontal cortex function variables were
signiﬁcant, consistent with West’s theory, and demonstrated declining performance with
advancing age, with a few exceptions. The magnitude of the signiﬁcant correlations
between prefrontal cortex ﬁrnctioning measures and age ranged ﬁom .24 to .53.
Perseverative errors on the RFF T and the CVLT, performance on the Stroop word-
naming trial, Digit Span total score, the proactive interference score from the CVLT, and
verbal ﬂuency were not signiﬁcantly related to age. Regarding the relationship between
education and performance on prefrontal cortex function measures, only Digit Span total
score, perseverative errors on the CVLT, and RUFF total designs were signiﬁcantly
related to education. The magnitude of these signiﬁcant relationships ranged from .20 to
.27. The remaining relationships between education and prefrontal cortex function
variables were not signiﬁcant.

Analysis of variance (AN OVA) was used to examine the relationship between
gender and performance on prefrontal cortex ﬁrnction measures. Only four prefrontal
cortex function variables were signiﬁcantly related to gender (see Appendix B). Males

performed signiﬁcantly better on the PASAT, completed more designs on the RUFF,

45

made fewer perseverative errors on the CVLT, but made more non-perseverative errors
on the WCST than females.

Depression was signiﬁcantly negatively related to seven prefrontal cortex function
measures (see Appendix C). Performance on Trailmaking Test Parts A and B, Visual
Memory Span total, the color-word naming trial of the Stroop, PASAT total score, and
RUFF total designs all were lower with higher levels of depressive symptomatology,
whereas the number of perseverative errors on the WC ST was higher with higher levels of
depressive symptomatology. The magnitude of signiﬁcant correlations ranged from .20 to
.33. Notably, the average GDS score for this sample was well below the accepted cut-off
score of 10 (M = 6.77, SD = 6.66), indicating a very low frequency of reported depressive
symptomatology in this sample.

Appendix D illustrates the relationship between measures of preﬁ'ontal cortex
function. The majority of prefrontal cortex function variables were signiﬁcantly related to
each other, with a few exceptions. The magnitude of signiﬁcant correlations ranged from
.21 to .76. Total score on the SDMT was the only measure that was signiﬁcantly related
to all other prefrontal cortex ﬁrnction variables. Perseverative errors on the RUFF was
not signiﬁcantly related to any other prefrontal cortex ﬁrnction variable. The proactive
interference score on the CVLT was signiﬁcantly related only to the verbal ﬂuency score,
with higher proactive interference scores being associated with poorer verbal ﬂuency
scores. Perseverative errors on the CVLT was signiﬁcantly related only to total categories
achieved on the WC ST and to WC ST non-perseverative errors, with greater perseverative
CVLT errors associated with fewer categories achieved on the WCST and greater WCST

non-perseverative errors.

The relationship between prefrontal cortex ﬁrnction and verbal memory measures
is illustrated in Appendix E. Only two prefrontal cortex function measures were
signiﬁcantly related to all verbal memory measures: fewer non-perseverative errors on the
WC ST and stronger performance on the color-word naming trial of the Stroop were
associated with stronger performance on all verbal memory measures. The total
immediate recall score from trials 1-5 on the CVLT was signiﬁcantly related, in the
expected direction (i.e., stronger performance on the prefrontal cortex ﬁrnction measure
was associated with stronger verbal memory performance), to the majority of prefrontal
cortex ﬁrnction variables. The magnitude of signiﬁcant relationships between verbal
memory and prefrontal cortex function measures ranged ﬁ'om .21 to .45. In contrast, total
recognition hits on the CVLT was signiﬁcantly related to only two prefrontal cortex
ﬁrnction measures: stronger performance on the color-word naming trial of the Stroop
and fewer non-perseverative errors on the WC ST were associated with greater recognition
hits.

Hypothesis 1: Examination of the 3-factor Model of Prefrontal Cortex Function

The three-factor model of prefrontal cortex function (working memory,
interference control, and inhibition of prepotent responses) was analyzed using the
following steps:

1. Correlations of manifest variables (speciﬁc test scores) for each corresponding latent
variable (each proposed factor) were examined.
2. Principal components analysis (PCA), constraining the data to three factors, was

employed to determine the ﬁt of the three-factor model.

47

As proposed, if the three-factor model of prefrontal cortex ﬁrnction was determined to not
offer the best ﬁt to the data, then a new measurement model would be determined through
exploratory factor analysis.

Conﬁtions of manifest vai'ables to their corresmnding latent variable

The relationship between speciﬁc test scores hypothesized to comprise each of the
three factors representing preﬁ'ontal cortex function was examined using Pearson product
moment correlations. Speciﬁc test scores/measures or manifest variables proposed to
comprise each latent variable were chosen on the basis of previous studies that had
indicated that these variables were related and theorized to be measuring the same or
similar cognitive processes. Table 2 presents the Pearson product moment correlations
between manifest variables comprising each latent variable. Only two of the measures
hypothesized to comprise the factor representing ability to inhibit prepotent responses

(Stroopcw and WCSTper) were signiﬁcantly related. Similarly, only two of the three
measures (Dsptot and VMSptot) hypothesized to comprise the working memory factor
were signiﬁcantly related. The ﬁnal measure, WC STcat, was unrelated to either measure.
In contrast, all measures hypothesized to comprise the interference control factor were
signiﬁcantly related.

In conclusion, two of the three factors proposed to represent prefrontal cortex
function were comprised of measures that were not highly homogeneous. While Cattell
(1982) has demonstrated that the best validity of a factor may be reached with a certain
moderate calculable homogeneity versus the highest homogeneity that can be achieved,

the low level of homogenity between some of the above measures strongly suggested that

48

another possible combination might provide a better representation of the three factors

proposed to comprise prefrontal cortex ﬁrnction.

Table 2

Correlations of Manifest Variables for Each Latent Variable

 

Ability to Inhibit Prepotent Responses

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Stroopcw WC SDCI' RFFT per C VLTper
Stroopcw 1.0 -.44** .01 -.03
WCSTper 1.0 -.08 -.11
RFFTJer 1.0 —.04
Working Memory

Dsptot WC STcat VMSptot

Dsptot 1.0 .07 .38"
WCSTcat 1.0 .02
Interference Control

PASAT TMTa TMTb SDMT
PASAT 1.0 -.41** -.41** .58"
TMTa 1.0 .53" -.60**
TMTb 1.0 -. 56"
CVLTpi -.02 .03 -. 17 .08

 

 

 

 

 

 

 

Note: * p 5 .05; " p 5 .01. PASAT = Paced Auditory Serial Addition Test; STROOPcw = Stroop Color-
Word Test — color-word naming trial; TMTa = Trailmaking Test Part A; TMTb = Trailmaking Test Part
B; Dsptot = Digit Span total score; VMSptot = Visual Memory Span total score; WCSTcat = Wisconsin
Card Sorting Test total categories achieved; WCSTper = Wisconsin Card Sorting Test perseverative
errors; RFFTper = Ruff Figural Fluency Test total perseverative errors; CVLTper = California Verbal
Learning Test total perseverative errors; CVLTpi = California Verbal Learning Test proactive
interference score; SDMT = Symbol Digit Modalities Test.

Principal components a_n_alysis (PCA) of threeﬁctor model of prefront_al cortex ﬁinction
A PCA was conducted on the 12 prefrontal cortex ﬁrnction measures. The

purpose of PCA is to reduce large data items into a few meaningful components based on

49

the intercorrlations among the items. Because PCA explains the variation that is unique to
an item and its error variance, as well (Pedhazur & Schmelkin, 1991), it was chosen as the
best data reduction method. The rule of eigenvalues greater than one (Johnson &
Mchern, 1992) was utilized to determine which components to retain. The Varimax
rotation method using generalized least squares was used to determine the best factor
solution. It was chosen in order to maximize the variances of the factors without changing
the mathematical properties of the solution (Tabachnick & Fidel], 1996).

Sample size restricted the number of measures that could be subjected to the PCA.
Using the rule of a minimum of 10 participants for each measure subjected to the
components analysis (Bryant & Yamold, 1995), the current data set allowed for only 10
measures to be used in the PCA. Measures to be eliminated from the PCA were
determined by examining loading coeffecient scores prior to and after rotation. As seen in
Table 3, three measures had coeffecient loading scores below 0.2 (RFFT per, CVLTper,
and CVLTpi). Comrey and Lee (1992) suggest that loadings below .32 (10% overlapping
variance) are “poor” measures of a component/factor. These three measures were

eliminated and the remaining 9 measures were subjected to PCA.

50

Table 3

Communalities of 12 Original Prefrontal Cortex Function Variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Shrunken r2 Final
PASAT .472 .583
STROOPcw .518 .601
TMTa .457 .850
TMTb .504 .613
Dsptot .319 .457
VMSptot .349 .427
WCSTcat .573 .893
WC STper .583 .669
SDMT .595 .683
RFFTper .007 . 1 16
CVLTper .131 .216
CVLTpi . 1 19 .217

 

 

 

 

 

Note: Extraction Method: Generalized Least Squares; PASAT = Paced Auditory Serial Addition Test;
STROOPcw = Stroop Color-Word Test - color-word naming trial; TMTa = Trailmaking Test Part A;
TMTb = Trailmaking Test Part B; Dsptot = Digit Span total score; VMSptot = Visual Memory Span total
score; WCSTcat = Wisconsin Card Sorting Test total categories achieved; WCSTper = Wisconsin Card
Sorting Test perseverative errors; SDMT = Symbol Digit Modalities Test; RFFTper = Ruﬂ‘Figural
Fluency Test total perseverations; CVLTper = California Verbal Learning Test total perseverative errors;
CVLTpi = California Verbal Learning Test proactive interference score.

The resulting three-factor structure is shown in Table 4. Eigenvalues for the three-
factor solution were between 1.1 to 2.6 and accounted for 60.4% of the variance. All
items loaded on at least one of the three factors with loading coefficients ranging from .42
to .81. However, measures comprising the three factors did not ﬁt the model of prefrontal
cortex ﬁrnctioning that was originally proposed. The ﬁrst factor, accounting for 29.1% of
the variance, was comprised of PASAT, Stroopcw, TMTa, TMTb, VMSptot, and SDMT.
The measures comprising this factor appear to represent a combination of attention,
concentration, set-shifting, and interference control. The second factor was comprised of

both WCST measures and accounted for 18.7% of the variance. This factor is consistent

51

Table 4

Varimax 3-Factor Structure (F actor-Variable Correlations) of Retained Original

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variables

Commuaalities
Variable Factor 1 Factor 2 Factor 3 Shrunken r2 Final
PASAT .61 .13 .34 .46 .53
Stroopcw .60 .3 8 .24 .5 1 .57
TMTa -.69 -. 13 .03 .40 .52
TMTb -.63 -.20 -.27 .46 .54
Dsptot .23 .04 .87 .3 1 .81
VMSmrt .54 -.09 .33 .36 .42
WCSTcat -.20 .85 -.01 .54 .73
WCSTper -.20 -.83 -.06 .57 .74
SDMT .79 .21 .17 .57 .70
% Variance 29.1 18.7 12.6

 

 

 

 

 

 

 

 

Note: Extraction Method — Generalized Least Squares; Rotation Method - Varimax with Kaiser
Normalization; PASAT = Paced Auditory Serial Addition Test; STROOPcw = Stroop Color-Word Test —
color-word naming trial; TMTa = Trailmaking Test Part A; TMTb = Trailmaking Test Part B; Dsptot =
Digit Span total score; VMSptot = Visual Memory Span total score; WCSTcat = Wisconsin Card Sorting
Test total categories achieved; WCSTper = Wisconsin Card Sorting Test perseverative errors; SDMT =
Symbol Digit Modalities Test.

with one of the three originally proposed factors, working memory. The ﬁnal factor was
comprised of one measure, DSptot, and accounted for 12.6% of the variance. In
conclusion, the nine originally proposed variables, when constrained to three factors, did
not adequately capture or represent the three identiﬁed constructs of prefrontal cortex
functioning.
Identiﬁcation ﬁr new measurement model of preﬁ'oatal cortex ﬁrnctioning

Since results of the PCA when the data was constrained to three factors

demonstrated that the proposed measurement model was not the best representation of the

52

factors underlying preﬁ'ontal cortex ﬁrnction, as proposed by West (1996), a series of

steps was undertaken to determine a new, more robust representative measurement model

based upon the data. These steps were as follows:

1.

Five additional measures of prefrontal cortex ﬁrnction that have been
previously identiﬁed by numerous other studies as measures of prefrontal
cortex function were added for subsequent analysis to determine potential
contribution to the measurement model: Wisconsin Card Sorting Test, non-
perseverative errors (W CSTnp), Stroop word naming trial score (Stroopw),
Stroop color naming trial score (Stroopc), Ruff Figural Fluency Test total
designs (RFFTdes), and the total score from the Controlled Oral Word
Association Test (Vﬂuency).

A series of exploratory factor analyses were used to determine two- and three-

factor models of prefrontal cortex ﬁrnction.

. Chi-square test was used to evaluate the goodness-of-ﬁt of the two- and three-

factor models in order to determine which factor model best ﬁt the data to be
used as the ﬁnal measurement model testing the last hypothesis investigating
the mediating role of prefrontal cortex function between age and verbal

memory performance.

Selection of additional variables

While difficulties with the measurement of ﬁrnctions associated with the prefrontal

cortex have been previously addressed, signiﬁcant evidence exists to suggest that there are

a number of neuropsychological tests that appear to consistently relate to ﬁrnctions

subserved by the prefrontal cortex. Ideally, those measures that have consistently

demonstrated, through neuroirnaging studies, to be associated with prefrontal activation

would have been utilized to represent one of the three underlying factors comprising

preﬁontal cortex ﬁrnction. These would have included: 1) a version of the n-back test

during which participants observe or listen for sequences of letters presented one at a time

53

and respond to an identiﬁed letter only when it follows another identiﬁed letter (Smith et
al., 1998; Barch et al., 1997; Braver et al., 1997; & Jonides et al., 1993); 2) an additional
measure of sustained attention, such as the Conners’ Continuous Performance Test or the
Attention Capacity Test (Chao & Knight, 1997; Parasurarnan & Giambra, 1991; Wilkings,
Shallice, & McCarthy, 1987); 3) a prospective memory measure (Einstein et al., 1992;
McDaniel & Einstein, 1992; and West, 1988); and 4) a measure of source memory
(Spencer & Raz, 1994; Schacter et al., 1994; Schacter et al., 1991; and Craik et al., 1990).

Studies demonstrating the reliability and validity of the ﬁve additional measures of
prefrontal cortex function selected have been previously reviewed. An additional score
from the WC ST was chosen because it was strongly correlated with the two other scores
from the WC ST, as well as with other measures of prefrontal cortex function. It was
added to provide another potential measure of working memory, since the current
working memory factor was comprised of only two variables. The word- and color-
naming trials from the Stroop were added because both were strongly correlated with a
number of other prefrontal cortex ﬁrnction measures related to attention, concentration,
and interference control, as well as to the color-word naming trial score from the Stroop
and because they provide additional measures of processing speed. Total scores ﬁom
both the Ruff Figural Fluency Test and the Controlled Oral Word Association Test were
added because of their strong correlational relationship to numerous other prefrontal
cortex ﬁmction measures and because numerous neuroirnaging studies have consistently
identiﬁed both as measures of spontaneous ﬂexibility that are associated with areas within
the prefrontal cortex (Gourovitch et al., 2000; Parks et al., 1988; Jason, 1985).

Results of exploratory factor a_r_ralyae_s

An EF A was conducted with the nine retained original preﬁ'ontal cortex function
measures and the ﬁve added prefrontal cortex function measures. Again, because of
restrictions due to sample size, measures were eliminated from the EPA based upon factor

loading coefficient scores prior to rotation. The ﬁve variables with the smallest factor

54

loading coefﬁcients were eliminated prior to constraining the data to three factors. As
seen in Table 5, ﬁve measures had coefﬁcient loading scores below 0.5 (TMTa, DSptot,
VMSptot, Vﬂuency, and RFFTdes) and were eliminated from subsequent analyses. EF A
was conducted with the ﬁnal, nine retained prefrontal cortex ﬁrnction measures,

constraining the data into both two— and three-factor solutions.

Table 5

Communalities of Retained Original Prefrontal Cortex Function Variables and

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Added Relevant Variables

Variable Shrunken r2 Final
PASAT .498 .588
STROOPw .516 .666
STROOPc .626 .794
STROOPcw .667 .757
TMTa .443 .577
TMTb .504 .682
Dsptot .342 .474
VMSptot .407 .566
WCSTcat .667 .769
WC STper .626 .710
WCSTnp .712 .871
SDMT .628 .702
Vﬂuency .322 .473
RFFTdes .398 .420

 

 

 

 

 

Note: Extraction Method: Generalized Least Squares; PASAT = Paced Auditory Serial Addition Test;
STROOPw = Stroop Color-Word Test — word naming trial; STROOPc = Stroop Color-Word Test — color
naming trial; STROOPcw = Stroop Color-Word Test — color-word naming trial; TMTa = Trailmaking
Test Part A; TMTb = Trailmaking Test Part B; Dsptot = Digit Span total score; VMSptot = Visual
Memory Span total score; WCSTcat = Wisconsin Card Sorting Test total categories achieved; WCSTper =
Wisconsin Card Sorting Test perseverative errors; WCSTnp = Wisconsin Card Sorting Test non-
perseverative errors; SDMT = Symbol Digit Modalities Test; Vﬂuency = Verbal Fluency; RFFTdes = Ruﬂ‘
Figural Fluency Test total designs.

The resulting three-factor structure is shown in Table 6. Eigenvalues for the three-

factor solution ranged from 1.6 to 2.4 and accounted for 66.4% of the variance. All items

55

loaded on at least one of the three factors with loading coefficients ranging from .53 to

.87. The ﬁrst factor, which accounted for 26.2% of the variance, was comprised of the

three WC ST measures and is consistent with the originally proposed working memory

factor. The second factor, accounting for 22.5 % of the variance, was comprised of the

three Stroop measures. This factor appears to represent a combination of processing

speed and inhibitory control. The ﬁnal factor was comprised of PASAT, TMTb, and

SDMT, accounted for 17.5% of the variance. This factor is comprised of variables

measuring sustained attention, mental ﬂexibility, and interference control.

Table 6

Varimax 3-Factor Structure (F actor-Variable Correlations) of Final Variables

 

 

 

 

 

 

 

 

 

 

 

 

 

Communalities

Variable Factor 1 Factor 2 Factor 3 Shrunken r2 Final
PASAT -.15 .40 .53 .43 .53
TMTb .13 -.18 -.75 .4i .63
SDMT -.20 .50 .60 .57 .67
WCSTcat -.84 .07 .09 .66 .76
WCSTper .77 .10 -.26 .61 .68
WCSTnp .91 -.14 -.08 .70 .87
Stroopw .07 .67 .25 .46 .57
Stroopc -.18 .87 .18 .59 .82
Stroopcw -.36 .60 .44 .63 .69

% Variance 26.2 22.5 17.5

 

 

 

 

 

 

 

 

Note: Extraction Method — Generalized Least Squares; Rotation Method — Varimax with Kaiser
Normalization; PASAT = Paced Auditory Serial Addition Test; STROOPcw = Stroop Color-Word Test -
color-word naming trial; '1be = Trailmaking Test Part B; Dsptot = Digit Span total score; VMSptot =
Visual Memory Span total score; WCSTcat = Wisconsin Card Sorting Test total categories achieved;
WCSTper = Wisconsin Card Sorting Test perseverative errors; SDMT = Symbol Digit Modalities Test.

The results of the two-factor structure are shown in Table 7. Eigenvalues for the

two-factor solution were 3.0 and 2.4 and accounted for 60.5% of the variance. Loading

56

coefﬁcients ranged from .52 to .87. The ﬁrst factor was comprised of PASAT, TMTb,
SDMT, and all three Stroop measures and accounted for 33.6% of the variance. It is
comprised of variables measuring sustained attention, interference control, processing
speed, and inhibitory control. While this factor is not a pure measure of interference
control, it was named “interference control” for subsequent analyses because interference
control is one of the major constructs that underlies performance on all of the tests
comprising this measure and to remain consistent with the language of the proposed
theory of preﬁontal cortex ﬁrnctioning. The ﬁnal factor was comprised of the three
WCST measures and accounted for 26.9% of the variance. This factor was named
“working memory” for subsequent analyses.

The goodness-of-ﬁt test was used to determine the best measurement model to use
for subsequent analyses. The 3-factor model yielded a chi-square value of 16.1 (df = 12)
and was signiﬁcant (p = .03). In contrast, the goodness-of-ﬁt test for the 2-factor model
was not signiﬁcant (p = .19) and yielded a chi-square value of 32.6 (df = 19). Thus, the
two-factor model of prefrontal cortex ﬁrnctioning, comprised of the interference control

and working memory factors, was used as the measurement model in subsequent analyses.

57

Table 7

Varimax 2~Factor Structure (Factor-Variable Correlations) of Final Variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Communalities

Variable Factor 1 Factor 2 Shrunken r2 Final
PASAT .64 -.17 .43 .52
TMTb -.56 . 17 .41 .53
SDMT .76 -.22 .57 .67
WC STcat . 12 -.85 .66 .77
WCSTper -.23 .78 .61 .68
WCSTnp -. 15 .91 .70 .87
Stroopw .69 .07 .46 .59
Stroopc .79 -. 18 .59 .73
Stroopcw .75 -.34 .63 .71
% Variance 33.6 26.9

 

 

 

 

 

 

 

Note: Extraction Method — Generalized Least Squares; Rotation Method - Varimax with Kaiser
Normalization; PASAT = Paced Auditory Serial Addition Test; STROOPcw = Stroop Color-Word Test —
color-word naming trial; TMTb = Trailmaking Test Part B; Dsptot = Digit Span total score; VMSptot =
Visual Memory Span total score; WCSTcat = Wisconsin Card Sorting Test total categories achieved;
WCSTper = Wisconsin Card Sorting Test perseverative errors; SDMT = Symbol Digit Modalities Test.

Structural Equation Modeliag

Hypothesis 2 that the relationship between age and verbal memory performance is
mediated by prefrontal cortex ﬁrnction was tested by comparing three structural equation
models. Model 1 tested the hypothesized direct effect of age on verbal memory
performance without taking prefrontal cortex ﬁrnction (as represented by interference
control and working memory) into consideration. Model 2 tested the hypothesized
indirect effects of age on verbal memory performance taking prefrontal cortex function
into consideration. Model 3 tested the hypothesized direct and indirect effects of age on

verbal memory performance taking prefrontal cortex ﬁrnction into consideration. AMOS

58

4 (Arbuckle & Wothe, 1999) was used to obtain generalized least squares estimates of the
model coefficients.
Mm

The purpose of model 1 was to evaluate the signiﬁcant direct effect of age on
verbal memory. The Chi—Square [x2 (df = 62, N = 102) = 80.95, p > .05] was not
signiﬁcant for model 1, but the RMSEA (.06) was large, and the GFI (.88) and the CFI
(.80) were small. However, the data ﬁt the rule of thumb that an acceptable ﬁt exists if
two times the degrees of freedom exceeds the x2 [df = 62 x 2 = 124; 124 > 80.95]. Taken
together, the ﬁt indices provide evidence for an adequate ﬁt between the model and the
data.

The standardized path coefﬁcients for model 1 are presented in Figure 4.
Relationships in the model are all in the expected direction and all are signiﬁcant, with one
exception. Age directly affected verbal memory performance, with advancing age
resulting in poorer verbal memory performance (standardized coefficient = -.30). Age
directly affected interference control and indirectly affected working memory
performance. Advancing age predicted poorer performance on interference control
measures (standardized coefficient = -.70). In turn, poorer performance on interference
control measures predicted poorer performance on working memory measures
(standardized coefﬁcient = .29). However, advancing age did not signiﬁcantly, directly
predict poorer performance on working memory measures. This model, with the uni-
directional arrow from interference control and working memory, was chosen because it

offered the best ﬁt to the data, when compared to all other models evaluated.

59

 

 

Verb-l
Memory

12 = 80.95

 

 

 

RMSE-A = .06
‘Signiﬁcant path

 

Figure 4. Standardized Path Coefﬁcients of Model 1 — Direct Effects Model

M01812

The purpose of model 2 was to evaluate the signiﬁcant, indirect effect of age on
verbal memory performance when prefrontal cortex ﬁrnction was taken into account.
Model 2 resulted in a strong ﬁt between the data and the model. Not only was the Chi-
Square non-signiﬁcant [x2 (df = 60, N = 102) = 68.27, p = .22], the GFI (.90) and the CF]
(.91) were large and the RMSEA (.04) was small.

The standardized path coefﬁcients for model 2 are presented in Figure 5. All
relationships in the model are signiﬁcant and in the hypothesized direction with the same
exception as demonstrated in model 1. Two of the same paths, as found signiﬁcant in
model 1, were found signiﬁcant in model 2. Additionally, the direct paths from working
memory and interference control to verbal memory were signiﬁcant. Age indirectly
aﬂ‘ected verbal memory performance. Advancing age resulted in poorer performance on
interference control measures (standardized coefficient = -.58). Poorer performance on

interference control measures predicted poorer performance on verbal memory measures

60

 

Working

 

 

 

.28‘
.15 Memory
Verb-l
A” Memory
40‘
-.58‘ .

'28 12 = 68.27

Interference 4f: 50

Control 2 = 22
GP] = .90
CFI = .91

RMSEA = .037

‘Significant path

 

Figure 5. Standardized Path Coefﬁcients of Model 2 — Indirect Effects Model

(standardized coefﬁcient = .28). Poorer performance on interference control measures
also predicted poorer performance on working memory measures (standardized coefficient
= .40). In turn, poorer performance on working memory measures resulted in poorer
performance on verbal memory measures (standardized coefﬁcient = .28). The results
indicate that performance on measures of interference control and working memory
signiﬁcantly inﬂuence the relationship between advancing age and declining verbal
memory performance.
M_OAeI_3

The purpose of model 3 was to demonstrate that the direct effect between age and
verbal memory became insigniﬁcant when it was considered in conjunction with the
indirect effect between these two constructs, taking prefrontal cortex ﬁrnction into
consideration. This would conﬁrm that prefrontal cortex function was mediating the

relationship between age and verbal memory. The results of model 3 were nearly identical

61

to those of model 2 and demonstrated a strong ﬁt between the data and the model. The
Chi-Square [x2 (df = 59, N = 102) = 68.00, p = .20] was non-signiﬁcant, the GFI (.90)
and the CFI (.91) were large, and the RMSEA (.04) was small.

The standardized path coefﬁcients for model 3 are presented in Figure 6. The
same four paths, as found signiﬁcant in model 2, were found signiﬁcant in model 3.
However, the direct eﬁea of age on verbal memory was insigniﬁcant. Age indirectly
affected verbal memory performance. Advancing age resulted in poorer performance on
interference control measures (standardized coefficient = -.58). Poorer performance on
interference control measures predicted poorer verbal memory performance (standardized
coefﬁcient = .32). Poorer performance on interference control measures also predicted
poorer performance on working memory measures (standardized coefﬁcient = .40). In
turn, poorer performance on working memory measures predicted poorer verbal memory

performance (standardized coefﬁcient

 

Working
-.1 5 Memory 39'

.07 Verbd

.40‘

 

 

 

 

-.58‘ 32¢
Interference
Control

7’ = 68.00

df= 59

2 = .198

on = .90

on = .91
RMSEA = .039

‘Signiﬁcent path

 

Figure 6. Standardized Path Coefficients of Model 3 — Mediation Model

62

= .29). The direct effect of advancing age on verbal memory performance became non-
signiﬁcant when considered along with the signiﬁcant effect that age had on performance
on prefrontal cortex function measures, which in turn signiﬁcantly negatively impacted
verbal memory performance. The results indicate that performance on measures of
interference control and working memory mediate the effects of advancing age on

declining verbal memory performance, supporting hypothesis 2.

DISCUSSION

In the past decade, there has been considerable effort to attempt to identify the
mechanism(s) by which cognitive aging occurs. Because of the expanding body of
evidence suggesting a selective, age-related decline in performance on neuropsychological
measures identiﬁed as sensitive to ﬁ'ontal lobe dysﬁrnction, a growing number of
researchers, clinicians, and theorists have suggested that the cognitive processes supported
by the frontal lobes, and more speciﬁcally the prefrontal cortex, are among the ﬁrst to
decline with advancing age. West (1996) proposed a prefrontal cortex ﬁrnction theory of
cognitive aging which predicted that ﬁmctions largely dependent on frontal regions are
selectively negatively impacted by aging. Speciﬁcally, age-related declines in cognitive
processes supported by the prefrontal cortex should emerge at an earlier age and should
be of greater magnitude than age-related declines observed in cognitive processes
supported by nonfrontal regions. West identiﬁed four speciﬁc processes underlying
preﬁontal cortex function (prospective memory, retrospective memory, interference
control, and inhibition of prepotent responses) that demonstrated strong ability to account
for age-related declines in a number of cognitive ﬁrnctions including verbal memory.

The purpose of this study was to evaluate the applicability of West’s prefrontal
cortex ﬁinction theory of cognitive aging to a sample of healthy, independently living older
adults. The ﬁrst step involved examining to what extent the four speciﬁc processes (of

which the ﬁrst two were combined in this study and labeled as “working memory”)

63

underlying prefrontal cortex ﬁrnction, as identiﬁed by West, were discemable with
multiple neuropsychological measures that have been consistently identiﬁed as measuring
preﬁ'ontal cortex or executive function. It was hypothesized that the data would ﬁt a
three-factor model of prefrontal cortex ﬁrnctioning, with factors representing interference
control, inhibition of prepotent responses, and working memory. Secondly, the mediating
strength of the identiﬁed underlying prefrontal cortex processes was examined in the
relationship between age and verbal memory performance. Verbal memory performance
was chosen to investigate the hypothesized mediating role of prefrontal cortex ﬁrnction
because of its wide use in investigations of cognitive aging and because of its direct,
pragmatic application to daily ﬁrnctioning. It was hypothesized that the three-factor
measurement model of prefrontal cortex functioning would mediate the relationship

between advancing age and declining performance on verbal memory measures.

The Model of Prefrong Cortex Function

Neither correlational analyses nor initial principatl components analysis (PCA)
supported the proposed 3-factor model of prefrontal cortex ﬁrnction. Tests proposed to
measure interference control, the ability to “clear out” action-irrelevant information from
working memory, were all signiﬁcantly related. In contrast, only two of the three tests
proposed to measure working memory, the ability to maintain relevant task information
“on-line” while performing mental operations, were signiﬁcantly related. Only two of the
four tests pr0posed to measure the ability to inhibit prepotent responses, preventing an
action sequence from being “captured” by a dominant internal response pattern or highly
salient environmental stimuli, were signiﬁcantly related.

The originally proposed measures of interference control were tests that required
sustained attention, concentration, processing speed, and the ability to quickly shift
conceptual sets. Thus, none directly measured what West (1996) described as interference

control; however, stronger performances on all of these tests undoubtedly related to how

64

effectively a person could keep task-irrelevant information from interfering with
subsequent efforts to attend, processes information quickly, and set shift. Numerous
studies have shown that older adults demonstrate greater buildup of interference over
trials, as evidenced by measures such as proactive interference, and declining ability to
sustain attention, both of which can be attributable to a decreased ability to clear task-
irrelevant information from working memory (Nyberg et al., 1997; O’Donnell et al., 1994;
Shimamura & Jurica, 1994; Parasuraman & Giarnbra, 1991; and Dobbs et al., 1989). It
was based on this rationale and research that that aforementioned tests were selected.
However, a more effective, efﬁcient way to have measured interference control would
have been to utilize a version of the Brown-Peterson Consonant Trigram Test or other
tasks on which strong performance depends upon release from proactive interference.
Consistent declines with advancing age on the Brown-Peterson task have been
demonstrated, regardless of the length of the retention interval (West, 1996). In this
study, a measure of proactive interference from the CVLT was included as part of the
subsequent EFA’s to determine a prefrontal cortex measurement model. However, this
measure, CVLTpi, was found to be signiﬁcantly related to only one other measure of
prefrontal cortex function (Vﬂuency) and had one of these lowest factor loading scores of
all of the prefrontal cortex ﬁrnction measures. This suggests that it was either a poor
measure of proactive interference for this sample (due to lack of speciﬁcity and/or
sensitivity) or that the originally proposed measures of interference control represented
constructs that were not sensitive or speciﬁc enough measures of interference control.

Of the three originally proposed measures of working memory, only two were
signiﬁcantly related (DSp and VMSp). Correlations between these measures and the ﬁnal
working memory measure, WC STcat, were extremely low (5 .1). While scores on the
WCST have been consistemly described as measures of working memory (Robbins et al.,
1998; Cabeza & Nyberg, 1997; Ragland et al., 1997; Libon et al., 1994; Parkin & Walter,
1991; Spreen & Strauss, 1991), data on performance on both Digit and Visual Memory

65

Span tests is less consistent (Reynolds, 1997; Ryan, Lopez, & Paolo, 1996; and Lezak,
1995). Both Digit and Visual Memory Span have been purported to measure
auditory/visual attention, short-term memory, and working memory. Backward memory
span, in particular, requires not only sustained attention, but the ability to hold information
“on-line” while actively reversing the sequence of digits presented. Thus, backward
memory span performance relies more on working memory ability. Reynolds (1997)
reviewed previous studies and conducted factor analyses including forward and backward
span recall, concluding that these scores should not be combined for clinical analysis
because the cognitive skills required for both are distinct. Thus, using only digit and visual
memory span backwards scores in this study may have provided a more speciﬁc and
sensitive measure of working memory. Combining both forward and backward
performances may have resulted in a measure that represented both more passive attention
or short-term memory skills with more active attentional and working memory ability.
This may have made it a less speciﬁc measure of working memory (yet still a potentially
good measure of general prefrontal cortex function), making it insigniﬁcantly related to
WC ST performance.

None of the four originally proposed measures of the ability to inhibit prepotent
responses were signiﬁcantly related. Furthermore perseverative errors on the RUFF was
not signiﬁcantly related to any other prefrontal cortex ﬁrnction measure and perseverative
errors on the CVLT was related to only two other prefrontal cortex function measures
(W C STcat and WCSTnp). Perseverative errors represent an inability to prevent an action
sequence from being captured by a dominant internal response pattern that may have been
previously rewarded. Thus, measures of perseverative errors provide a method of
analyzing the ability to suppress a prepotent response. Numerous studies have
demonstrated a relationship between perseverative errors and prefrontal cortex function
(Raz et al., 1998; Burgess and Shallice, 1996; Stuss et al., 1994). It is therefore puzzling

that none of the perseverative error scores were signiﬁcantly related to one another in this

66

study. While each measure of perseverative errors was generated from tests that each
required somewhat different cognitive skills (i.e., the WCST is a working memory test; the
CVLT is a verbal memory test; the RUFF is a test of ﬁgural ﬂuency), the reason for
making perseverative errors on each of these tests is theorized to be the same mechanism.
Processing speed may have played a role in the lack of a relationship between RUFFper
and other measures. Speciﬁcally, this was the only perseverative error measure that was
generated under a time constraint.

The remaining measure of the ability to inhibit prepotent responses, the color-word
naming trial of the Stroop, has been one of the most widely documented and accepted
measures of this construct (Spieler, Balota, & Faust, 1996; West, 1996; Vendrell et al.,
1995; Uchiyama et al., 1994; and Houx et al., 1993). Strong performance on the
Stroopcw requires inhibitory control, as well as the ability to process information quickly.
While it was insigniﬁcantly related to the perseverative measures previously discussed,
Stroopcw was strongly, signiﬁcantly related to all other measures of preﬁ'ontal cortex
ﬁrnction.

Results of the EFA that conﬁned the data to three factors indicated that the
proposed three-factor model of preﬁontal cortex ﬁrnction did not ﬁt the current data set.
This may have been attributable to a number of issues, including: 1) a more appropriate
underlying factor structure of prefrontal cortex ﬁrnction existed; 2) tests were
misidentiﬁed as to the ﬁrnctions measured; 3) tests were unable to speciﬁcally discern
underlying ﬁrnctions comprising prefrontal cortex ﬁrnction; 4) selection of measures was
not well suited to evaluation of West’s model; and 5) sample characteristics (such as level
of education) signiﬁcantly inﬂuenced the measurement and identiﬁcation of the underlying
factor structure of prefrontal cortex ﬁrnction.

Alternative Models oﬂactors Comm Prefrontal Cortex Function

A series of exploratory factor analyses (EFA) conducted suggested alternative

two- and three-factor prefrontal cortex ﬁrnction solutions. Notably, some of the originally

67

identiﬁed measures of prefrontal cortex function demonstrated weak factor loadings
during the series of EF A’s and were subsequently dropped from analyses. These included:
perseveration error scores on the C VLT and RUFF, the CVLT proactive interference
score, Trail Making Test A, and total scores on Digit Span and Visual Memory Span.
Furthermore, some tests not originally proposed as measures of one of the underlying
constructs of prefrontal cortex ﬁrnction were found to signiﬁcantly contribute to the
measurement model of prefrontal cortex function in this study. These included: non-
perseverative errors on the WC ST and scores from the Stroop word- and color-naming
trials.

Factor loading scores for CVLTper, RUFFper, and CVLTpi were extremely low
(5 .13), suggesting that these scores represent ﬁrnctions essentially different from the
remaining prefrontal cortex function measures. Likewise, factor loadings for DSp and
VMSp were also low (3 .35). As previously discussed, both of these scores may represent
a combination of passive and active attentional abilities, short-term memory, and working
memory. Although both were signiﬁcantly correlated with a number of prefrontal cortex
ﬁrnction measures, sensitivity and speciﬁcity of these measures may be less than that of
other preﬁ'ontal cortex function measures, resulting in them being related to, yet not
strongly representative of, the speciﬁc constructs underlying prefrontal cortex function.
The lower factor loading score of TMTa may be a function of its validity as a measure of
prefrontal cortex ﬁrnction. Speciﬁcally, it may predominantly represent processing speed
and visuomotor tracking, versus performance on TMTb, which requires both of these
skills, in addition to mental ﬂexibility and cognitive set-shifting skill. Processing speed
measures have been shown to be signiﬁcantly negatively affected by advancing age
(F ristoe, Salthouse, & Woodard, 1997; Fisk & Warr, 1996; Salthouse, Fristoe, & Rhee,
1996; Salthouse, 1994; and Tun et al., 1992) and may be reliant, in part, on the integrity of
preﬁ'ontal cortex firnction. The extent to which performance on TMTa is related to

processing speed may make it essentially diﬁ’erent from ﬁrnctions measured by the

68

remaining prefrontal cortex ﬁrnction tasks in this sample. In conclusion, measures
eliminated through EF A appear to be a combination of test scores that are less sensitive
and speciﬁc than other measures of prefrontal cortex ﬁrnction.

Rationale for including the ﬁve additional measures of prefrontal cortex function,
as previously addressed, was based predominantly on: previous research that had
demonstrated consistent relationships between selected measures and activation of areas
of the prefrontal cortex in neuroirnaging studies; other research demonstrating
relationships between selected measures and other tests purported to measure prefrontal
cortex function; and the strong correlational relationships between selected measures and
other prefrontal cortex ﬁrnction measures in this study. WCSTnp was added because it
was strongly correlated with the two other scores from the WC ST, as well as with other
measures of prefrontal cortex function. It was added with the assumption that it would
provide another potential measure of working memory. This assumption was supported,
as all WC ST scores consistently loaded independently of other preﬁontal cortex measures
during EFA’s. Numerous recent factor analytic studies examining the WCST and other
neuropsychological measures have demonstrated that subtest scores on the WC ST
typically load independently of other neuropsychological measures or with other tests
designed to be administered, scored, and interpreted in the same manner (such as the
Booklet Category Test — Reitan & Wolfson, 1985) (Boone et al., 1998; Greve, Ingram, &
Bianchini, 1998; and O’Donnell et al., 1994). This is consistent with current EF A results
that demonstrated that the three selected measures of the WCST loaded together and
contributed uniquely as a construct of prefrontal cortex function. Notably, other studies
examining the factor structure and construct validity of the WC ST have demonstrated
both two- and three—factor solutions (Greve, Bianchini, Hartley, & Adams, 1999; Greve
Ingram, & Bianchini, 1998; Greve et al., 1997; and Goldman et al., 1996). However, the
sample in each of these studies was signiﬁcantly different (i.e., healthy, middle-aged adults,
persons who have had strokes) ﬁom the current study. No WCST factor analytic studies

69

with a sample of healthy older adults could be located through extensive literature
reviews. It is possible that the WCST could be measuring slightly different constructs
depending upon the sample of participants used, or that the WCST is more or less
sensitive to the construct(s) measured, depending upon the sample. This is supported by
Burgess et al.’s (1998) study of the ecological validity of tests of executive functioning
that included the WCST. Study ﬁndings indicated that while the WCST was found to be
one of the most highly sensitive measures of overall neuropsychological dysfunction, it
was strongly related to behavioral characteristics deﬁned as reﬂecting inhibition (response
suppression problems, irnpulsivity, disinhibition) and executive memory (confabulation,
temporal sequencing problems, perseveration).

While the word- and color-naming trials from the Stroop have been
predominantly identiﬁed as measures of processing speed, because both were signiﬁcantly
related to the majority of preﬁ'ontal cortex measures and because they also require
sustained attention ability, they were added to subsequent EFA’s. During all subsequent
EF A’s, the added Stroop scores demonstrated strong factor loadings (2 .52). When
prefrontal cortex ﬁrnction variables were constrained to three factors, all three scores ﬁom
the Stroop loaded together on a factor, independent of all other preﬁ'ontal cortex ﬁrnction
variables. This factor was interpreted to represent a combination of processing speed and
inhibitory control. When prefrontal cortex ﬁrnction variables were constrained to two
factors, the WC ST scores loaded independently on one factor and all three Stroop scores
loaded with the remaining prefrontal cortex function variables (PASAT, TMTb, and
SDMT). This latter factor was interpreted to represent a construct comprised of
interference control, sustained attention, processing speed, inhibitory control, and mental
ﬂexibility.

Notably, although both the RUFF and Vﬂuency have been associated with
prefrontal cortex ﬁrnction (Gourovitch et al., 2000; Parks et al., 1988; Jason, 1985) and

were related to the majority of prefrontal cortex function measures in this study, they

70

demonstrated weak factor loadings during EFA’s. Thus, both were eliminated from
ﬁrrther EF A’s. Burgess et al. (1998) found that verbal ﬂuency was highly insensitive to
general neurological pathology and executive ﬁrnctioning in sample of both mixed etiology
neurological patients and normal, age-matched controls. It is possible that either RUFF
and Vﬂuency were not sensitive enough measures of the constructs identiﬁed in this study
or that each measured a separate but related prefrontal cortex ﬁrnction construct from the
remaining tests in this study.

The ﬁnal two- and three-factor models of prefrontal cortex ﬁrnction generated
both included a factor comprised of all three WCST measures that was labeled working
memory. This is consistent with West’s (1996) theory, as well as with previously cited
factor analytic studies of prefrontal cortex ﬁmction measures. However, the remaining
factors generated in both the two- and three-factor models did not directly ﬁt the
remaining secondary processes of prefrontal cortex function, as proposed by West. In the
three-factor model, the factor comprised of all three Stroop scores appears to represent a
combined measure of processing speed and inhibitory control. The remaining factor is
comprised of measures of sustained attention, interference control, processing speed, and
mental ﬂexibility. In the two-factor solution, the ﬁnal factor is a combination of measures
of sustained attention, processing speed, inhibitory and interference control, and mental
ﬂexibility. These results are generally consistent with other factor analytic studies of tests
identiﬁed as prefrontal cortex function measures. Boone et al. (1998) conducted a factor
analytic study including WCST, Vﬂuency, Stroop, DSp, a sustained attention test similar
to the PASAT, and a test nearly identical to SDMT on data collected from a sample
consisting of patients referred for neuropsychological testing and healthy volunteers (mean
age = 56.0; mean education = 14.0 years). Results yielded a three-factor solution. One
factor contained all WCST scores, the second was comprised of Vﬂuency, all Stroop
scores, and SDMT, and the ﬁnal factor contained DSp, the sustained attention test, IQ,

and memory retention scores. A factor analytic study of the PASAT, WCST, TMTb, a

71

visual attention test, and a test similar to the WC ST conducted on a sample of persons
referred for rehabilitation services at a state agency (mean age = 30.2; mean education =
12.4) yielded two factors (O’Donnell et al., 1994). The ﬁrst factor included PASAT,
TMTb, and the visual attention test and the second factor was comprised of WCST and
the test similar to the WC ST. A factor analytic study of numerous attention tests
conducted on a sample of outpatients referred for neuropsychological evaluation (mean
age = 34.9; mean education = 12.9) that included PASAT, TMT, DSp, VMSp, and Stroop
found that all of these tests loaded on the same factor (Schmidt et al., 1994). Notably, in
all of these studies, participants were signiﬁcantly younger and less educated than persons
in this study. Additionally, most participants in these studies were referred for
neuropsychological testing because of some identiﬁed neurologic problem. Nonetheless,
results from these studies are similar to current study ﬁndings. This suggests that
regardless of the age, level of education, and type of impairment a person may have, the
tests of prefrontal cortex ﬁrnction utilized in this study appear to be related in a manner
reﬂecting a working memory factor and an attentional/processing speed factor. This is
consistent with the two-factor model of prefrontal cortex ﬁrnctioning being a better ﬁt for
this data set (as well as possibly other data sets). It is possible that the
attention/processing speed factor may be a general measure of a person’s inhibitory and
interference control and that current prefrontal cortex ﬁrnction measures are not sensitive
enough to discern these underlying constructs.

Measuremenj of Prefrontal: Cortex Function with Neuropsychological Tests

The neuropsychological tests used in this study to measure prefrontal cortex
ﬁrnctioning are some of the most popular and widely researched and used instruments to
assess frontal or executive functioning. These tests are proposed to be selectively
sensitive to frontal lobe dysﬁrnction. However, there are problems with both the
sensitivity and speciﬁcity of these tests as indicators of prefrontal cortex dysfunction
(Phillips & Della Sala, 1999; Phillips, 1997; Reitan & Wolfson, 1994). In addition to

72

measuring prefrontal cortex ﬁrnction, performance on these tests may also be affected by
lesions/impairment elsewhere in the brain, in part, because of the extensive connections
between the frontal lobes and other cortical and subcortical areas. Some studies have
demonstrated that persons with frontal lobe lesions and persons with extensive frontal lobe
damage who have demonstrated problems in managing their daily lives perform well on
numerous tests of frontal functioning (Anderson et al., 1991; Shallice and Burgess, 1991).
Furthermore, some studies have demonstrated insigniﬁcant correlations between tests of
ﬁ'ontal functioning and behavioral problems commonly presented by persons with frontal
lobe lesions (Burgess et al., 1998; Baddeley et al., 1997; Alderman, 1996). However, all
of these studies demonstrating weak or a lack of a signiﬁcant relationship between ﬁontal
lobe lesions/damage, behavior, and performance deﬁcits on tests purported to measure
frontal functioning were conducted on persons with identiﬁed, signiﬁcant ﬁontal cortex
damage through neuroirnaging studies. Even though numerous neuroirnaging studies of
older adults have demonstrated neurological changes in the prefrontal cortex, these
changes are usually described as being less severe than the acquired brain lesions/damage
that result from cerebral vascular disease or other head trauama. Presumably, the effects
of age-associated prefrontal cortex neurological changes would be less than those
evidenced by more severely neurologically impaired persons. This suggests that the
speciﬁc measurement of prefrontal cortex dysfunction in older adults by identiﬁed tests of
frontal lobe functioning is an even more difficult endeavor.

Additionally, as was the case in this study, age differentially relates to and eﬁ’ects
performance on various measures of frontal lobe ﬁrnctioning (Phillips & Della Sala, 1999;
Moscovitch & Winocur, 1995; Shimamura, 1994). While some prefrontal cortex ﬁmction
measures in this study were not signiﬁcantly related to age (Vﬂuency, DSp, perseverative
errch on CVLT and RUFF, Stroop word naming trial, CVLT proactive interference
score) others were very strongly related (p3 .40; TMTa, SDMT, Stroop color- and color-
word naming trials, and WCSTnp).

73

Furthermore, numerous studies have demonstrated that there is a signiﬁcant
amount of shared variance between tests of prefrontal cortex ﬁrnction (Robbins et al.,
1998; Fisk & Warr, 1996; Salthouse et al., 1996; Salthouse 1992). The majority of these
studies have shown that controlling for the effects of processing speed eliminates a
signiﬁcant proportion of the age effects on performance on tests of prefrontal cortex or
executive functioning. This may be an explanation for the EF A results in this study.
Speciﬁcally, the ﬁnal two-factor solution of prefrontal cortex function in this study
contains one factor comprised of WC ST scores that do not involve a time constraint and
thus are most likely not related to processing speed. The remaining factor is comprised of
numerous tests that are all dependent, to some degree, on processing speed. Thus, it
provides a measure of how quickly a person can process information while sustaining their
attention, shifting mental sets, and inhibiting prepotent responses.

Relationsh_ip Between Measured Cpgnitive Functions aad Neuroanatomical
Substrates

The aging brain loses weight, volume, and neurons and experiences an increased
number of pathological structures and changes in neurochemistry. These changes have
been demonstrated to be disproportionally greater in the prefrontal cortex when compared
to other areas of the brain in post-mortum and neuroirnaging studies (F abiani, Friedman,
& Cheng, 1998; Raglund et al., 1997; Haug & Eggers, 1991; Gur et al., 1987; Squire,
1987; Gerard & Weisberg, 1986; Strubel et al., 1985; Shaw et al., 1984; Scheibel et al.,
1975). However, some ﬁndings have been less conclusive and have indicated that other
brain regions demonstrate similar or even greater deterioration that the prefrontal cortex
(Coffey et al., 1992; Giaquinto, 1988; de Leon et al., 1984). In her critical review of the
frontal aging hypothesis, Greenwood (2000) concluded that while there were clear effects
of aging on brain structure and physiology which may underlie the altered patterns seen in
blood ﬂow and metabolism during task performance, the evidence that prefrontal areas are

more vulnerable than other brain regions to these aging effects is weak.

74

Others have emphasized that the changes in cognition with advancing age that
have been largely attributed to a decline in frontal lobe ﬁrnction cannot exclude the role
that the frontal-striatal and frontal-temporal-parietal circuits play in promoting cognitive
ﬁrnction. For example, Rubin (1999) demonstrated, through a neuroimaging study, that
the decrease in the size of the caudate is proportional to that of the prefrontal cortex and
that this volume change is accompanied by declines in inhibitory processes, executive
control, and cognitive speed similar to that demonstrated in normal aging. Schumacher et
al. ’3 (1996) PET study investigating performance on a verbal working memory task
consistently demonstrated overlapping activation of the dorsolateral prefrontal cortex,
Broca’s area, bilateral superior and posterior parietal cortices, anterior cingulate, and right
cerebellum. Thus, in addition to measuring prefrontal cortex ﬁrnction, performance on
tests purported to measure prefrontal cortex ﬁrnction such as those used in this study are
most likely also affected by lesions/impairment elsewhere in the brain, in part, because of
the extensive connections between the frontal lobes and other cortical and subcortical

areas.
The frontal lobes have been described as an executive for the rest of the brain and

have thus been referred to by many as the central executive. Parkin (1998) has argued
that the central executive does not exist. Speciﬁcally, he has stated that there is no
localization evidence, based on his review of neuroradiological and neuropsychological
studies, for a central executive and that tests identiﬁed as measuring executive ﬁrnction
should do so in a “qualitative way” rather than assuming that a range of speciﬁc tests
represent a unitary construct. Greenwood’s (2000) conclusion is similar in her critical
review of West’s (1996) prefrontal cortex theory of cognitive aging. She stated that while
the frontal lobes are subject to age-related changes reﬂected in both behavior and

pathology, evidence for a prefrontal cortex ﬁrnction theory of cognitive aging is both weak

75

and conﬂicting. She argues that a network-based theory of cognitive aging spanning more
that one area of the brain is more appropriate. This is consistent with one of Salthouse et
al. ’5 (1996) conclusions that the source of the shared variance in different cognitive
measures purported to measure frontal functioning is not discrete and localized, but
instead corresponds to anatomical and/or physiological characteristics distributed across
many regions of the brain.

However, both West (2000) and Baddeley (1998) have argued in response to both
Parkin and Greenwood that signiﬁcant and sufﬁcient evidence exists to demonstrate that
the prefrontal cortex/central executive plays a substantial role in cognitive aging. Both
acknowledge that the prefrontal cortex/central executive does not exclusively and unitarily
control the aging process or the effects of cognitive aging as demonstrated by current
neuropsychological measures. Baddeley (1998) argues that the central executive exists as
a “concept” that provides a useﬁrl basis for studying the complexities of executive control
and identifying subprocesses that may then be mapped on to their anatomical substrate(s).

The theoretical debate about the accuracy and applicability of the prefrontal cortex
function theory to cognitive aging contributes to increased difficulties in validity of
measurement of prefrontal cortex functions. While some neuropsychological tests such as
those employed in this study may not be able to be deﬁnitively identiﬁed as measures of
speciﬁc prefrontal cortex functions because of the aforementioned issues, a review of
more recent neuroimaging studies with many of the measures in this study demonstrates a
signiﬁcant, strong trend towards preﬁ'ontal cortex involvement. Additionally, one of the
ways to interpret the large amount of shared variance between tests purported to measure

frontal ﬁrnctioning is that all of these measures may be dependent upon the integrity of a

76

single neuroanatomical structure or system. The prefrontal cortex circuits (i.e., prefrontal-
striatal and prefrontal-parietal circuits) may be the best candidate currently for this single
structure or system.

In conclusion, the neuropsychological tests used in this study to measure preﬁontal
cortex ﬁrnction most likely do not reﬂect ﬁrnctioning in discrete and independent
structures, nor are the age-related effects on them exclusively determined by distinct and
unique sets of inﬂuences. However, this does not necessarily mean that some level of
speciﬁcity and localization could not exist. Many of the prefrontal cortex meausures in
this study that have been shown, through ﬁrnctional MRI studies, to be related to speciﬁc,
increased activity in the preﬁontal cortex were found to be strongly related. Furthermore,
when combined with study ﬁndings demonstrating that preﬁ'ontal cortex function strongly
mediated the relationship between aging and poorer verbal memory performance, results
ﬁom this study are believed to support West’s and Baddeley’s belief that the preﬁ'ontal
cortex/central executive plays a substantial role in cognitive aging. Functional MRI
studies with healthy older adults that allow for the elimination of the contributions of other
areas of the brain to performance on the prefrontal cortex function tests used in this study
would provide ﬁrrther insight as to the localizability of such measures.

Sample Characteristics

The older adults in this study comprised a unique sample of highly educated, non-
depressed, independently living individuals. Education has been proposed to provide
protection against both normal and pathological aging processes (Stern et al., 1994;
Katzrnan, 1993; Mortiner, 1988). Capitani, Barbarotto, and Laicana (1996) proposed that

three different patterns of association could be expected between age-related decline and

77

education: 1) parallelism: age-related decline runs the same course in different
educational groups, so that no interaction is observed; 2) protection: age-related decline
is attenuated in well-educated participants; and 3) conﬂuence: the initial advantage of
well-educated participants in middle age is reduced in later life. Several studies have
demonstrated a strong association between educational level and performance on various
neuropsychological tests (Ardila, Ostrosky-Solis, Rosselli, and Gomez, 2000; Ardila,
Rosselli, & Ostrosky, 1992; and Heaton, Grant, & Mathews, 1986). Given these ﬁndings,
it could be assumed that the high level of education in the current sample, coupled with
the fact that many of the participants were retired university faculty and thus were more
familiar (and perhaps comfortable) with the use of testing/evaluations, would have
signiﬁcantly attenuated ﬁndings. Interestingly, the data do not appear to directly support
this assumption. Only three measures of prefrontal cortex function were signiﬁcantly
positively associated with education and none of these measures were included in the ﬁnal
measurement model of prefrontal cortex function. One of these measures, Digit Span
total score, had been previously shown to be associated with education (Ryan, Lopez, &
Paolo, 1996 and Powell &Whitla, 1994). However, a number of studies support current
ﬁndings that performance on measures of prefrontal cortex function may be minimally
affected by level of education in studies with older adults. In particular, Lowe and
Reynolds (1999) found that education was signiﬁcantly negatively related only to the
number of random responses made on a test of verbal set-shifting and rule-induction
similar to the WCST. Christensen, Korten, Jorm, and Henderson (1997) found that
education has also been shown to be unrelated to processing speed, verbal memory, and

reaction time in older adults. In contrast, Ardilla et al. (2000) found that while education

78

had a protective inﬂuence on a test measuring phonemic ﬂuency, it had a conﬂuent
inﬂuence Digit Span-backwards scores. However, in this study, the mean level of
education for the most highly educated group of adults ages 66 to 85 was 13.5 years,
substantially lower than that of the current sample. However, in a sample of older adults
with a minimum of 16 years of education, Powell and Whitla (1994) found that education
was signiﬁcantly positively related to performance on tests of auditory attention and verbal
and non-verbal reasoning.

Notably in the Powell and Whitla (1994) study, the cognitive beneﬁt of education
was strongest for women, especially on tests of reasoning. Since it was highly unusual for
women to attain advanced degrees in the 1940’s , it is likely that highly educated women,
like those in this study, may have been brighter, more highly motivated, and/or came from
homes that strongly valued intellectual activities. However, very few signiﬁcant
differences were found between the performance of men and women on tests of preﬁontal
cortex function in this study. Despite the fact that the level of education was not
signiﬁcantly different for men and women in this study, males performed signiﬁcantly
better on the PASAT, completed more designs on the RUFF, and made fewer
perseverative errors on the CVLT. However, males made more non-perseverative errors
on the WC ST. Overall, gender and level of education appeared to have had a minimal
impact on the ﬁnal results of this study.

Depression has been demonstrated to have a negative effect on memory skills,
attention, verbal ﬂuency, processing speed, and performance on tasks similar to the WCST
(King, Cox, Lyness, & Caine, 1995; LaRue, 1992; Rubin et al., 1991; and Emery &

Breslau, 1989). However, other studies have failed to replicate these ﬁndings (Boone et

79

al., 1995 and Abas et al., 1990). Although the mean score on the depression measure
(GDS) for this sample was well below the cut-off suggestive of signiﬁcant depressive
symptomatology, scores on the GDS were signiﬁcantly negatively related to a number of
measures of prefrontal cortex functioning, including four measures used in the ﬁnal
measurement model (PASAT, TMTb, WC ST perseverative errors, and the Stroop color-
word naming trial). This ﬁnding has signiﬁcant treatment and research implications. First,
it suggests that participants in this study, who were highly educated and ﬁrnctioned
independently, may be more sensitive to and thus more affected by even minimal levels of
depressive symptomatology. Professionals working with these individuals would therefore
need to be more alert to even subtle observed or reported changes in mood, as this could
have a signiﬁcant, negative effect on cognitive ﬁrnctioning. Second, it suggests that
research investigating the effects of level of depression on cognitive ﬁrnctioning in older
adults that classiﬁes participants as “depressed” or “non-depressed” may be overlooking
subtle but signiﬁcant differences within classiﬁed groups that could have a major impact

upon cognitive and psychosocial ﬁrnctioning.

The Mediating Role of Prefrontal Cortex Function
The second hypothesis, that prefrontal cortex ﬁrnction mediates the relationship
between age and verbal memory performance, was strongly supported. The direct effect
of age on verbal memory performance became insigniﬁcant when the indirect effect of
prefrontal cortex function, as represented by working memory and interference control,
was considered. In fact, the direct effect of age on verbal memory performance became

almost negligible (standardardized coefﬁcient = .07), indicating that prefrontal cortex

80

ﬁmction accounted for nearly all of the decline in verbal memory performance with
advancing age in this sample of healthy, independently living older adults. While the direct
effects model evaluating the relationship between age and verbal memory performance
was the weakest of the three models evaluated, the direct, signiﬁcant relationship between
increasing age and poorer verbal memory performance had already been previously
demonstrated through correlational analyses.

These results are consistent with both neuroimaging and neuropsychological
studies demonstrating relationships between age, prefrontal cortex ﬁrnctioning, and verbal
memory. Shallice et al. (1994) demonstrated that verbal episodic memory involves a
network of speciﬁc prefrontal and posterior structures in healthy middle-aged adults.
Petrides, Alivisatos, and Evans’ (1995) ﬁrnctional MRI study of healthy young adults
found that the mid-ventrolateral frontal cortex was directly involved in active retrieval
mechanisms during a verbal memory task and the mid-dorsolateral frontal region
participated in free recall, which they concluded was a result of working memory
requirements during the verbal memory task. In Stuss et al. ’3 (1994) study examining
organizational strategies of patients with unilateral or bilateral frontal lobe injuries on a
word list learning task, all patients with frontal lesions demonstrated impaired verbal recall
due to poor higher order organization of learning. In particular, those patients with right
frontal lesions demonstrated excess intralist repetitions/perseverations. Parkin and Walter
(1991) examined the relationship between age, short-term memory, and frontal
dysfunction and found that age-related differences on a verbal memory task reﬂected a
deﬁcit in initial acquisition and a retrieval deﬁcit stemming from the fiontal atrophy known

to be associated with normal aging. In a conditional associative learning (CAL) study that

81

compared the performances of patients with frontal lobe lesions to healthy older adults,
both groups demonstrated impaired CAL performance, but the deﬁcit was greater in the
former group where it was speciﬁc to patients with dorsolateral prefrontal cortex lesions
(Levine, Stuss, & Milberg, 1997). The authors concluded that the congruence between
older adults and frontal lobe patients suggested that age-related decline in inhibitory
processes is due to dorsolateral prefrontal cortex dysfunction. Current study results
support this conclusion. Reuter-Lorenz (2000) studied PET activation patterns of
younger and older adults during both letter identity and spatial location memory tasks.
Younger adults showed left-sided frontal activation for the letter identity task and right-
sided frontal activation for the spatial location task, whereas older adults showed bilateral
frontal activation for both tasks. These results were interpreted to suggest that the greater
demands on storage and maintenance in older adults or an increased need to elaborate
stimulus representations that have been less well-encoded lead to bilateral ﬁontal
activation. Another possibility is that prefrontal/executive processes are selectively
impaired in older adults and that the effort to recruit all available prefrontal cortical
resources results in increased activation of preﬁ'ontal areas in older but not younger adults.
The current mediating model of preﬁontal cortex function indicates a strong, direct
relationship between age and interference control, which then indirectly, signiﬁcantly
affects verbal memory performance. Thus, decline in verbal memory performance is a
result of declining sustained attention, processing speed, interference and inhibitory
control skills, which are negatively inﬂuenced by advancing age. This same indirect
relationship was not demonstrated with working memory. While working memory had a

signiﬁcant, indirect effect on verbal memory performance, its role as a mediator between

82

age and verbal memory performance was signiﬁcantly inﬂuenced by interference control.
This indicates that for the older adults in this study, the negative effect of advancing age
did not have a signiﬁcant direct effect on working memory skills. Rather, advancing age
signiﬁcantly negatively inﬂuenced interference control skills that then signiﬁcantly affected
working memory performance, leading to a decline in verbal recall. Theoretically, this is
consistent with West’s (1996) theory. Task irrelevant information needs to be cleared
from working memory stores (interference control) and an action sequence must be
prevented from becoming inﬂuenced by a dominant internal response pattern (inhibition of
prepotent responses) in order for working memory skills to be optimized. Interpretation
with the current measurement model suggests that attention must be sustained,
information must be processed quickly, interfering information must be controlled, and
previously rewarded internal response patterns must be inhibited in order to be able
optimize working memory ability. Once this occurs, working memory performance then
mediates the relationship between age and verbal memory performance.

Consistent with this interpretation of current study ﬁndings, Chao and Knight
(1997) found that the percentage of perseverative errors and overall performance on the
WC ST were positively correlated with measures of distractibility and sustained attention
during a event-related potential study using an auditory delayed match-to-sample task
comparing the performances of younger and older adults. Additional support for the
aforementioned interpretation of current study ﬁndings is found in the expanding body of
research implicating the signiﬁcant role of processing speed in the relationship between
age and performance on neuropsychological measures. Measures of executive functioning

were found to be less important in accounting for age differences between younger and

83

older adults on a verbal associative learning task than measures of the learners’ perceptual
speed (Fisk & Warr, 1998). These same researchers (1996) demonstrated age diﬁ‘erences
in central executive ﬁrnctioning are primarily attributable to a decline in the rate at which
information is activated within the working memory system. In a study examining the
effect of ﬂuid intelligence and processing speed on performance of frontal lobe function
measures, Parkin and Java (1999) concluded that a large proportion of the age-related
variance on measures of frontal lobe ﬁrnction may be attributable to a more general factor
characterized jointly by ﬂuid intelligence and processing speed. Salthouse has
demonstrated, in a number of studies, that slower processing speed mediates the
relationship between age, performance on executive measures, and memory (F ristoe,
Salthouse, & Woodard, 1997; Salthouse et al., 1996; Salthouse, 1994).

While all of these studies demonstrate that speed of information processing plays a
signiﬁcant role in the relationship between age and performance on both prefrontal cortex
ﬁrnction and memory measures, closer examination of the tests used to assess processing
speed and prefrontal cortex/executive ﬁrnctioning raises questions as to the deﬁnitiveness
of some of the conclusions. Many of the measures of processing speed used in the
aforementioned studies have been labeled by others (including this author) as tests of
prefrontal cortex/executive functioning and nearly all measures require sustained attention
ability. For example, the Digit Symbol test from the WAIS-III (W echsler, 1999), which is
nearly identical to the Symbol Digit Modalities Test used in this study, was employed as a
measure of processing speed in many of the aforementioned studies (Parkin & Java, 1999;
Salthouse, 1994; Fristoe et al, 1997). Performance on this test is not only dependent on

processing speed, it also requires sustained attention and working memory skills. In the

Fisk and Warr (1996) study only one measure of working memory span and one measure
of executive functioning were used, neither of which were standardized
neuropsychological tests. Finally, one of the measures of frontal lobe ﬁrnction, category
ﬂuency, that Parkin and J ave (1999) employed has been consistently shown in
neuroimaging studies to be weakly or unrelated to prefrontal cortex involvement.

In conclusion, while these measurement issues do not negate the consistent
ﬁndings that processing speed plays a signiﬁcant role in the relationship between age,
prefrontal cortex ﬁrnction, and memory performance, it is apparent that these studies are
also plagued by the issue of difﬁculty in speciﬁcally and selectively identifying and
measuring cognitive processes. Salthouse et al. (1996) acknowledge that neuroanatomical
and neurophysiological factors responsible for the age-related reduction in processing
speed still need to be identiﬁed and that speed of information processing plays an
important, although not exclusive, role in cognitive aging. Notably, because many of the
measures in this study have been found to share a proportion of their age-related variance
with other measures, including those related to processing speed, it is likely that the age-
related inﬂuences on current prefrontal cortex ﬁrnction measures are not independent.
However, the use of structural equation modeling to investigate the second hypothesis in
this study allowed the investigation of both the common and speciﬁc age-related
inﬂuences that contributed to the relationship between age, performance on prefrontal
cortex function measures, and verbal memory. Nonetheless, a consistent theme apparent
in current study ﬁndings and those summarized above is that processing speed, sustained
attention, and the ability to control interfering stimuli play a critical role in verbal memory

performance and are negatively inﬂuenced by advancing age.

85

Theoretical Considerations and Future Directions

The results of this study partially supported West’s (1996) prefrontal cortex
ﬁrnction theory of cognitive aging by demonstrating that prefrontal cortex ﬁrnction
strongly mediated the relationship between advancing age and declining verbal memory
performance. However, the measurement model of prefrontal cortex function generated
from this data set did not support the underlying secondary processes of prefrontal cortex
ﬁrnction that West outlined. Instead, it suggested that a factor representing measures of
sustained attention, interference control, processing speed, and mental ﬂexibility and a
working memory factor provided the best representation of prefrontal cortex or executive
ﬁrnction for this sample. While this model still appears consistent with, although less
speciﬁc than, the model that West (1996) proposed, a brief review of alternative
theoretical considerations is warranted.

1n Greenwood’s (2000) critical evaluation of the prefrontal cortex function theory
of cognitive aging, she concludes that there is only “weak and conﬂicting evidence that
frontal regions are selectively and differentially affect by aging” (pp. 705). She suggests
that a better approach to conceptualizing the complex effects of aging on cognition is to
consider the effects of aging on the integrity of a processing network that involves more
than the frontal lobes. Mesularn (1998) has proposed a model of brain organization based
on “transmodal areas”. These transmodal areas work in combination with primary sensory
areas and are divided into ﬁve networks, one of which is a “working memory-executive
ﬁrnction” network. Greenwood (2000) suggests that Mesularn’s theory, in combination

with a myelin-based theory of brain aging (which proposes that age-related changes in

cognition are the result of deterioration in axonal myelination), could account for much of
the evidence on which the frontal aging hypothesis is based. Speciﬁcally, processing
networks more highly dependent on corticocortical connections, such as the working
memory-executive function network, would be increasingly vulnerable to the effects of
demyelination with advancing age. Interestingly, this theory does not appear to
signiﬁcantly depart from West’s (1996) theory in its emphasis on neurophysiology and
prefrontal cortex ﬁmction. Greenwood’s suggested theory emphasized networks as
opposed to a speciﬁc area/lobe of the brain.

Baddeley (1998) has also suggested this as a more appropriate way to view the
role of the prefrontal cortex in cognitive aging. He has proposed that the frontal lobes
represent a large and multi-faceted area of the brain which is unlikely to be unitary in
ﬁrnction and that is best described as a “central executive” system. Executive processes
are likely to involve links between different parts of the brain and hence are unlikely to be
exclusively associated with the ﬁ'ontal lobes. Consequently, persons without clear
evidence of frontal damage may conceivably have executive deﬁcits, such as those
demonstrated with advancing age. Thus, the central executive provides a useﬁrl basis for
studying the complexities of executive control and identifying subprocesses, which may
then be mapped onto their neuroanatomical substrates (Baddeley, 1998).

While current study results strongly supported the mediating role of prefrontal
cortex/executive ﬁrnction in the relationship between declining verbal memory
performance with advancing age, the underlying processes comprising prefrontal cortex
ﬁrnction were somewhat inconsistent with West’s (1996) theory. It is possible that with

larger sample size, a broader range of education among participants, and additional, more

87

sensitive measures of prefrontal cortex ﬁrnction (such as those previously mentioned),
results may have been improved. Furthermore, inclusion of ecologically valid rating scales
to be completed by participants and family members (requesting ratings about
demonstrated behaviors consistent with preﬁontal cortex/executive dysfunction), in
combination with the use of ﬁrnctional MRI procedures, would have provided the most

comprehensive and accurate way to evaluate the hypotheses in this study.

88

APPENDIX A

89

Table A-1: Correlations Between Demographic and Prefrontal Cortex Functioning

APPENDIX A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variables

Age Education
PASAT -.38** .15
RFFTdes -.24** .27"
RFFTper -. 13 .01
STROOPw -. 13 .16
STROOPc -.45** -.05
STROOPcw -.48** .19
TMT a .53“ .16
TMTb .35" -.O9
DSptot -. 17 20*
VMSptot -.29** .08
WCSTcat -.29** -.06
WCSTper .34" -.02
WCSTnp .42** .15
SDMT -.51** .07
CVLTpi -.06 .08
CVLTper -.08 .25M
Vﬂuency -.06 .18

 

 

 

 

 

N = 109; * p < .05, ** p < .01. PASAT = Paced Auditory Serial Addition Test raw score; RFFTdes =
Ruff Figural Fluency Test total designs; RFFTper = Ruﬂ'Figural Fluency Test total perseverations;
STROOPw = Stroop Color-Word Test raw score — word naming trial; STROOPc = Stroop Color-Word
Test raw score - color naming trial; STROOPcw = Stroop Color-Word Test raw score — color-word
naming trial; TMTa = Trailmaking Test Part A total seconds; TMTb = Trailmaking Test Part B total
seconds; Dsptot = Digit Span total score; VMSptot = Visual Memory Span total score; WCSTcat =
Wisconsin Card Sorting Test total categories achieved; WCSTper = Wisconsin Card Sorting Test
perseverative errors; WCSTnp = Wisconsin Card Sorting Test non-perseverative errors; SDMT = Symbol
Digit Modalities Test total score; CVLTpi = California Verbal Learning Test proactive interference score;
CVLTper = California Verbal Learning Test total perseverative errors; Vfluency = COWAT total score.

APPENDIX B

91

Table A-2: Relationship Between Gender and Prefrontal Cortex Functioning Variables

APPENDIX B

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sum of Squares F P
PASAT 11745.88 5.70 .02“ (males perform better)
RFFTdes 2085.04 4.60 .03“ (males perform better)
RFFTper 286.39 1.62 .21
STROOPw 321.88 1.32 .25
STROOPc 198.41 1.39 .24
STROOPcw 6.86 6.86 .79
TMTa 338.68 2.07 .15
TMTb 1993.78 0.95 .33
Dsptot 31.02 2.36 .13
VMSptot 6.23 0.80 .37
WCSTcat 7.43 2.63 .11
WCSTper 138.54 1.20 .28
WCSTnp 1085.15 11.0 .00" (males make more NPE’s)
SDMT 32.29 0.37 .54
CVLTpi 4.09 1.34 .25
CVLTper 89.17 5.45 .02* Quales make fewer errors)
Vﬂuency 1 5.69 . l 3 .72

 

 

 

 

 

 

N = 102; df= 100; * p < .05, ** p < .01. PASAT = Paced Auditory Serial Addition Test raw score;
RFFTdes = Ruff Figural Fluency Test total designs; RFFTper = Ruff Figural Fluency Test total
perseverations; STROOPw = Stroop Color-Word Test raw score — word naming trial; STROOPc = Stroop
Color-Word Test raw score - color naming trial; STROOPcw = Stroop Color-Word Test raw score -
color-word naming trial; TMTa = Trailmaking Test Part A total seconds; TMTb = Trailmaking Test Part
B total seconds; Dsptot = Digit Span total score; VMSptot = Visual Memory Span total score; WCSTcat =
Wisconsin Card Sorting Test total categories achieved; WCSTper = Wisconsin Card Sorting Test
perseverative errors; SDMT = Symbol Digit Modalities Test total score; P1 = California Verbal Learning
Test proactive interference score; CVLTper = California Verbal learning Test total perseverative errors;
Vﬂuency = COWAT total score.

92

APPENDIX C

93

APPENDIX C

Table A-3: Signiﬁcant Pearson Product-Moment Coefficients for Depression and

Prefrontal Cortex Functioning Measures

 

 

PASAT

TMTa

TMTb

VMSp

WC STper

Stroopcw

RFFTdes

 

 

GDS

 

-.20*

 

.23*

 

.32"

 

-.21*

 

.23*

 

-.25*

 

-.33**

 

 

N = 102; * p < .05, ** p < .01. GDS total = Geriatric Depression Scale total score; PASAT = Paced
Serial Auditory Addition Test, total score; TMTa = Trailmaking Test Part A; TMTb = Trailmaking Test
Part B; VMSp = Visual Memory Span total from the Wecshler Memory Scale-Revised; WCSTper =
Wisconsin Card Sorting Test perseverative error score; Stroopcw = Stroop Color-Word Test — color-word

naming trial; RFFTdes = Ruff Figural Fluency Test total designs.

94

 

APPENDIX D

95

APPENDIX D

Table A-4: Pearson Product Correlation Coefﬁcients between Prefrontal Cortex Function
Measures

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DSp VMSp PASAT Vflueng SDMT TMTa TMTb CVLTper CVLTpi
DSp 1.00 .04
VMSp .38” 1.00 .08
PASAT .44” .41” 1.00 -.02
SDMT .30” .44" .58" .25‘ 1.00 .08
TMTa -. 13 ~37“ -.4l ” -.22‘ -.60” 1.00 .03
TMTb -.38” -.43” -.41“ -.23‘ -.56“ .53” 1.00 -. l 7
CVLTper .03 -.01 .02 .06 .08 -. 13 -.09 1.00 -.15
Stroopw .18 .12 .40” .40” .44" -.26” -.28” -.05 —.ll
Stroopc .27” .26“ .45” .27" .60“ - 45" -.34” -.01 -.1 1
Stroopcw .35” .37" .53" .39” .60” -.43” ~53” -.03 .05
WCSTcat .07 .02 . 16 .04 .27" -.24‘ -.25‘ .22‘ .04
WCSTper -.13 -.05 -.28" -.21‘ -.36” .25‘ .34“ -.11 .04
WCSTnp -. l O —.04 -.29” -.05 -.33” .33“ .21’ -.27“ -.03
RFFTdes .37” .32” .52“ .26” .51“ -.39“ -.47” -.02 .01
WE -.09 .02 .00 .10 .16 -.12 o.14 -.04 .08
Vﬂuency .25‘ .08 .37“ 1.00 .25‘ -.22‘ -.23‘ .06 -.26“
Stroopw Stroopc Stroopcw WCSTcat WCSTper WCSTnp RFFTdes
Stroopw 1.00
Stroqrc .60” 1.00
Stroopcw .47“ .66” 1.00
WCSTcat .10 .25' .40” 1.00
WCSTper -. l 4 -.27“ -.44“ -.72” 1.00
WCSTnp -.01 -.32“ -.46“ -.76" .71“ 1.00
RFFTdes .27” .36” .46” .19 -.25‘ -.21' 1.00
RFFTper . 12 .09 .01 .04 -.O8 .01 .00 1.00

 

 

 

 

 

 

 

 

 

 

 

N = 102; * p < .05, ** p < .01. WCSTcat = Wisconsin Card Sorting Test total categories achieved; DSp =
Digit Span total from the Wecshler Memory Scale-Revised (WMS-R); VMSp = Visual Memory Span total
from the WMS-R; PASAT = Paced Auditory Serial Addition Test total raw score; SDMT = Symbol Digit
Modalities Test; TMTa = Trailmaking Test Part A; TMTb = Trailmaking Test Part B; CVLTper =
California Verbal Learning Test — total perseverative errors; CVLTpi = California Verbal Learning Test
Proactive Interference Score; Stroopw = Stroop Color-Word Test, word naming trial raw score; Stroopc =
Stroop Color-Word Test, color naming trial raw score; Stroopcw = Stroop Color-Word Test, color-word
trial raw score; WCSTper = Wisconsin Card Sorting Test perseverative error score; WCSTnp =
Wisconsin Card Sorting Test non-perseverative error score; RFFTdes = Ruff Figural Fluency Test total
designs; RFFTper = Ruff Figural Fluency Test total perseverative errors; Vﬂuency = Verbal Fluency total
score. '

96

 

APPENDIX E

97

APPENDIX B

Table A-5: Pearson Product Correlations Between Prefrontal Cortex Function

Variables and Verbal Memory Measures

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Lmtot CVLTtot CVLTsf CVLTlf CVLTrec

PASAT .33" .27" .16 .17 .16
RUFFdes .18 .21* .15 .14 .05
RFFTper .10 .02 .00 .12 -.04
TMTa -.16 -.25* -.14 -.17 -.01
TMTb -.16 -.26** -.18 -.17 -.06
SDMT .35" .32** 24* .21* .02
DSptot .30" .20* .16 .10 .12
VMSptot .09 .15 .07 -.01 -.11
CVLTper -.08 .32" .19 23* -.01
CVLTpi .17 .15 .16 .15 .02
Vﬂuency . 13 . 14 .08 .07 .06
WCSTcat 25* .29" .15 21* .15
WCSTjrer -.28** -.28** -.15 -. 19 -. 18
WCSTnp -.36** -.45** -.36** -.41** -.38**
Stroopw . 16 . 12 .02 .06 .04
Stroopc .18 .31" 22* 25* .12
Stroopcw .34** .38“ 25* .26" 23*

 

 

 

 

 

 

 

 

N = 102, ‘ p < .05, “‘ p < .01. LMtot = Logical Memory immediate memory total score; CVLTtot = California
Verbal Learning Test total recall trials 1-5; CVLTsf = California Verbal Learning Test short delay free recall total;
CVLTlf = California Verbal Learning Test long delay free recall total; CVLTrec = California Verbal Learning Test
recognition trial total; PASAT = Paced Auditory Serial Addition Test raw score; RFFTdes = Ruff Figural Fluency
Test total designs; RFFT per = Ruff Figural Fluency Test total perseverations; STROOPw = Stroop Color-Word Test
raw score - word naming trial; STROOPc = Stroop Color-Word Test raw score - color naming trial; STROOPcw =
Stroop Color-Word Test raw score -— color-word naming trial; TMTa = Trailmaking Test Part A total seconds; TMTb
= Trailmaking Test Part B total seconds; Dsptot = Digit Span total score; VMSptot = Visual Memory Span total
score; WCSTcat = Wisconsin Card Sorting Test total categories achieved; WCSTper = Wisconsin Card Sorting Test
perseverative errors; Wisconsin Card Sorting Test non-perseverative errors; SDMT = Symbol Digit Modalities Test
total score; CVLTpi = California Verbal Learning Test poactive interference score; CVLTper = California Verbal
Learning Test total perseverative errors; Vﬂuency = COWAT total score.

98

REFERENCES

Abas, M.A., Sahakian, B.J., & Levy, R. (1990). Neuropsychological deﬁcits and
CT scan changes in elderly depressives. Psycholgjcal Medicine, 3, 507-520.

Alderman, N. (1996). Central executive deﬁcit and response to operant
conditioning methods. Neuropsychological Rehabilitation, a, 161-186.

Anderson, S.W., Damasio, A., Jones, R.D., & Tranel, D. (1991). “frsconsin Card
Sorting Test performance as a measure of frontal lobe damage. Journal of Clinical and
Experiemental Neuropsychology, 13 909-922.

—’

Arbuckle, J.L. & Wothke, W. (1999). Amos 4.0 user’s guide. USA: Small
Waters Corporation.

Ardila, A., Ostrosky-Solis, F., Rosselli, M., & Gomez, C. (2000). Age-related
cognitive decline during normal aging: The complex effect of education. Archives of
ClinicalﬁNeuropsycholcgy, 1_5, 495-513.

Baddeley, A. (1998). The central executive: A concept and some misconceptions.
Journal of the International Neuropsychological Socim, 4, 523-526.

Baddeley, A., Della Sala, S., Papagno, C., Spinnler, H. (1997). Dual-task
performance in dysexecutive and nondysexecutive patients with a frontal lesion.
Neuropsychology, M, 187-194.

Barch, D.M., Braver, T.S., Nystrorn, L.E., Forrnan, S.D., Noll, D.C., & Cohen,
J .D. (1997). Dissociating working memory ﬁ'om task difficulty in human prefrontal
cortex. Neuropsychologia, 3_5_, 1373-1380.

Benton, AL. (1968). Differential behavioral effects in frontal lobe disease.
Neuropsychologia, Q, 53-60.

Benton, A.L. & Hamsher, K. (1983). Multi-lingpal Aphasia Examination. Iowa
City: AJA Associates.

Boone, K.B., Lesser, I.M., Miller, B.L., Wohl, M., Berrnan, N., Lee, A., Palmer,
8., and Back, C. (1995). Cognitive ﬁrnctioning in older depressed outpatients:
Relationship of presence and severity of depression to neuropsychological test scores.
Neuropsychology, 2, 390-398.

Boone, K.B., Ponton, M.O., Gorsuch, R.L., Gonzalez, J .J ., & Miller, BL. (1998).

Factor analysis of four measures of prefrontal lobe functioning. Archives of Clinicg
Neuropsychology, 13, 585-595.

99

Bowden, S. & Bell, R. (1992). Relative useﬁrlness of the WMS and WMS-R: A
comment on D’Elia et al. (1989). Journal of Clinical and Experimental Neuropsychology,
l_4, 340-346.

Braver, T.S., Cohen, J.D., Nystrorn, L.E., Jonides, J., Smith, E.E., Noll, DC.
(1997). A parametric study of prefrontal cortex involvement in human working memory.
NeuroImage, 5, 49-62.

Brittain, J., LaMarche, J., Reeder, K., Roth, D., & Boll, T. (1991). Eﬂ‘ects of age
and IQ on paced Auditory Serial Addition Test (PASAT) performance. The Clinical
Neuropsychologiat, é, 163-1 75.

Burgess, P.W., Alderman, N., Evans, J., Emslie, H., & Wilson, BA. (1998). The
ecological validity of tests of executive function. Journal of the International
Neurapsychological Socim, 5, 547-558.

Burgess, P.W. & Shallice, T. (1996). Response suppression, initiation and
strategy use following frontal lobe lesions. Neuro cholo 'a, 3_4_, 263-273.

Cabeza, R., Grady, C.L., Nybert, L., McIntosh, A.R., Tulving, E., Kapur, S.,
Jennings, I.M., Houle, S., & Craik, F .I.M. (1997). Age-related difference in neural
activity during memory encoding and retrieval: A positron emission tomography study.
Journal of Neuroscience. l_7, 391-400.

Cabeza, R. & Nyberg, L. (1997). Imaging cognition: An empirical review of PET
studies with normal subjects. Journal of Cognitive Neuroscience. 2, 1-26.

Capitani, E., Barbarotto, R., & Laicana, M. (1996). Does education inﬂuence age-
related cognitive decline? A further inquiry. Developmental Neuropgchology, 12, 231-
240.

Cattell, RB. (1982). The psychomtry of objective motivation measurement: A
response to the critique of Cooper and Kline. Britiah Journal of Educational Paychology,
§_2_, 234-241.

Chao, L.L. & Knight, RT. (1997). Prefrontal deﬁcits in attention and inhibitory
control with aging. Cerebral Cortex. 1, 63-69.

Christensen, H., Korten, A.E., Jorm, A.F., & Henderson, AS. (1997). Education
and decline in cognitive performance: Compensatory but not protective. International
Journal of Geriatric Pavehiatg, _1_2_, 323-330.

Comrey, A.L. & Lee, H.B. (1992). A ﬁrst course in factor analysis. (2" Ed.)
Hillsdale, NJ : Erlbaum.

100

Craik F .I.M., Morris, L.W., Morris, R.G., & Loewen, ER. (1990). Relations
between source amnesia and frontal lobe ﬁrnctioning in older adults. Psychology and
Agiag, a, 148-151.

de Leon, M.J., Ferris, S.H., George, A.E., Christman, D.R., Fowler, J.S., Gentes,
C., Reisberg, B., Gee, B., Emmerich, M., Yonekura, Y., Brodie, J ., Kricheﬂ‘, 1.1., & Wolf,
AP. (1983). Positron emission tomography studies of aging and Alzheimer's disease.
American Journal of Neurcﬂdiolgy, 5, 568-571.

de Leon, M.J., Ferris, S.H., George, A.E., Christman, D.R., Fowler, J.S., Gentes,
C., Brodie, J ., Reisberg, B., & Wolf, AP (1984). Positron emission tomography and
computed tomography: Assessment of the aging human brain. Journal of Computer

Assisted Tomogpaphy, 8, 88-94.

Delis, D., Squire, L.R., Bihrle, A., & Massman, P. (1992). Componential analysis
of problem-solving ability: Performance of patients with frontal lobe damage and amnesic
patients on a new sorting test. Neurgpsycholog'aa, 3_0, 683-697.

Delis, D., Freeland, F., Kramer, J ., & Kaplan, E. (1988). Integrating clinical
assessment with cognitive neuroscience: Construct validation of the California Verbal
Learning Test. Journal of Consultingand Cliniaal Psychology, 56 123-130.

—,

Delis, D., Kramer, J ., Kaplan, E., & Ober, B. (1984). Qﬁfornia Verbal Learning
Test Manual. San Antonio: Harcourt Brace Jovanovich.

Denckla, M.B. & Reiss, AL. (1997). Pre-frontal-subcortical circuits in
developmental disorders. In N.A. Krasnegor, G. R. Lyon, & P.S. Goldman-Rakic (Eds),
Development of the Prefrontal Cortex: Evol_utiona Neurobiology, and Behavior (pp. 283-
294). Baltimore: Paul H. Brooks Publishing Company.

Dobbs, A.R., Aubry, J .B., & Rule, BC. (1989). Age-associated release from
proactive interference: A review. Canadian Psychology 30 588-595.

,—,

Dobbs, AR. & Rule, BC. (1989). Adult age difference in working memory.
Psychology and Aging, 5., 500-503.

Drebing, C.E., Van Gorp, W.G., Stuck, A.E., Mitrushina, M., & Beck, J. (1994).
Early detection of cognitive decline in higher cognitively ﬁrnctioning older adults:
Sensitivity and speciﬁcity of a neuropsychological screening battery. Neuropsychology, 8,
31-37.

Drewe, E. (1974). The effect of type and are of brain lesion on Wisconsin Card
Sorting test performance. Cortex, 19, 159-170.

 

101

Dulaney, C.L. & Rogers, WA. (1994). Mechanisms underlying reduction in
Stroop interference with practice for young and old adults. Journal of Experimental

Paychology: Learning, Memogg, and Cogm'tion, _2_0, 470-484.

Duncan, J. (1995). Attention, intelligence, and the frontal lobes. In M.S.
Gazzaniga (Ed), The Cognitive Neurosciences. Massachusetts: MIT Press.

Egan, V. (1988). PASAT: Observed correlations with IQ. Personalig and
Individual Differences. _9_, 179-180.

Einstein, G.O., Holland, L.J., McDaniel, M.A., & Guynn, M.J. (1992). Age-
related deﬁcits in prospective memory: The inﬂuence of task complexity. Psychology and
Agiag, 1, 471-478.

Emery, O.B. & Breslau, LB. (1989). Language deﬁcits in depression:
Comparisons with SDAT and normal aging. Journal of Gerontology, 44 85-92.

—’

Eslinger, P.J. & Grattan, L.M. (1993). Frontal lobe and frontal-striatal substrates
for different forms of human cognitive ﬂexibility. Neuropsychologia, M, 17-28.

Fabiani, M., Friedman, D., & Cheng, J .C. (1998). Individual differences in P3
scalp distribution in older adults and their relationship to frontal lobe function.

Psychophysiology, M, 698-708.

Fisk, J.E. & Warr, P. (1996). Age and working memory: The role of perceptual
speed, the central executive, and the phonological loop. Psychology and Aging, M, 316-
323.

Folstein, M. ,Folstein, S ,& McHugh, P (1975). "Mini- Mental State":
practical method for grading the cognitive state of patients for the clinician. Journal of
P alchratrrc Resear__ch, 1_2,189-198.

Fristoe, N.M., Salthouse, T.A., & Woodard, J .L. (1997). Examination of age-
related deﬁcits on the Wisconsin Card Sorting Test. Neuropsychology, M, 428-436.

Fuster, J.K. (1980). The Prefrontal Cortex: Anatomy, Physiology, and
Neuropgchology of the FrontMLobes. New York: Raven Press.

Fuster, J .K. (1989). The Prefrontal Cortex. New York: Raven Press.

Gerard G & Weisberg, L. (1986). MRI periventricular lesions in adults.
Neurology, 3Q, 998-1001.

Giaquinto, S. (1988). Agiagand the Nervous System. New York: John Wiley &
Sons, Ltd.

102

Gourovitch, M.L.,Kirkby, B.S., Goldberg, T.E., Weinberger, D.R., Gold, J .M.,
Esposito, G., VanHorn, J ., & Berrnan, KR (2000). A comparison of rCBF patterns
during letter and semantic ﬂuency. Neuropsychology, M, 353-360.

Grady, C.L., Maisog, J .M., Horwitz, B., Ungerleider, L.G., Mentis, M.J., Salerno,
J .A., Pietrini, P., Wagner, E., & Haxby, J .V. (1994). Age-related changes in cortical
blood ﬂow activation during visual processing of faces and location. Journal of
Neuroscience. M, 1450-1462.

Grady, C.L., McIntosh, A.R., Horwitz, B., Maisog, J .M., Ungerleider, L.G.,
Mentis, M.J., Pietrini, P., Schapiro, M.B., & Haxby, J .V. (1995). Age-related reductions
in human recognition memory due to impaired encoding. Science, E, 218-221.

Greenwood, PM. (2000). The frontal aging hypothesis is evaluated. Journal of
the International Neuropsychological Socim, _6_, 705-726.

Greve, K.W., Bianchini, K.J., Hartley, S.M., & Adams, D. (1999). The Wisconsin
Card Sorting Test in stroke rehabilitation: Factor structure and relationship to outcome.
Archives of ClinicaMIeuropsychology, M, 497-509.

Greve, K.W., Brooks, J ., Crouch, J.A., Williams, M.C., & Rice, W.J. (1997).
Factorial structure of the Wisconsin Card Sorting Test. Britis_h Journal of Clini_c_al
Psychology, 3a, 283-285.

Greve, K.W., Ingram, F ., & Bianchini, K]. (1998). Latent structure of the
Wisconsin Card Sorting Test in a clinical sample. Archives of Clinical Neuropsychology,
M, 597-609.

Greve, K.W., Williams, M.C., Hassa, W.G., Littell, R.R., & Reinoso, C. (1996).
The role of attention in \Vrsconsin Card Sorting Test performance. Archives of ClinicM
Neuropsychology, M, 215-222.

Gronwall, D. & Sampson, H. (1974). The Psychological Effects of Concussion.
Auckland: Auckland University Press.

Gronwall, D. & Wrightson, P. (1974). Delayed recovery of intellectual ﬁrnction
after minor head injury. Lancet, 2, 605-609.

Goldberg, E. & Bilder, RM. (1987). The frontal lobes and hierarchical

organization of cognitive control. In E. Perecman (Ed), The Fron_ta1 Lobes Revisited (pp.
159-187). New York: The IRBN Press.

103

Goldman-Rakic (1987 ). Circuitry of primate prefrontal cortex and regulation of
behavior by representational memory. Handboolaof Physiology - The Nervous System. 5,
3 73-41 7.

Goldman-Rakic, P.S. & Brown, RM. (1981). Regional changes of monoamines in
cerebral cortex and subcortical structures of aging rhesus monkeys. Neuroscience, Q, 177-
187.

Gur, R.C., Gur, R.E., Obrist, W.D., Skolnick, B.E., & Reivich, M. (1987). Age
and regional cerebral blood ﬂow at rest and during cognitive activity. Arches of Genﬂ
Psychiatg, M, 617-621.

Guttman, C., Jolesz, F ., Kikinis, R., Killiany, R., Moss, M., Sandor, T., & Albert,
M. (1998). White matter changes with normal aging. Neurology, M, 972-978.

Haaland, K.Y., Vranes, L.F., Goodwin, J.S., & Garry, PJ. (1987). Wisconsin
Card Sort Test performance in a healthy elderly population. Journal of Gerontology, 4_2,
345-346.

Hanninen, T., Hallikainen, M.., Koivisto, K., Partanen, K., Laakso, M.P.,
Riekkinen, P.J., & Soininen, H. (1997). Decline in frontal lobe ﬁrnctions in subjects with
age-associated memory impairment. Neurology, 4_8, 148-153.

Harlow, J .M. (1868). Recovery from the passage of an iron bar through the head.
Publications of the Massachusetts Medical Socie_ty, 2, 327-346.

Hartley, A.A., Speer, N.K., Jonides, J ., Renter-Lorenz, P.A., & Smith, BE.
(2001). Is the dissociability of working memory systems for name identity, visual-object
identity, and spatial location maintained in old age? Neuropsychology, 1_5_, 3-17.

Hasher, L. & Zacks, R. T. (1979). Automatic and effortﬁrl processes in memory.
Journal of Eaperimenaal Pavchology: General, 108. 356-388.

Haug, H., Barnwater, U., Eggers, R., Fischer, D., Kuhl, S., & Sass, N.L. (1983).
Anatomical changes in aging brain: Morphometric analysis of the human prosencephalon.
In J. Cervois-Navarro & H. I. Sarkander (Eds), Brain Agjag: Neuropathology and
Neuropharrnacology Vol. 21: Aging (pp. 193-225). New York: Raven Press.

Heaton, WC. (1981). Wiscon_ain CaMrd Sorting Test. Odessa, TX: Psychological
Assessment Resources.

Hecaen, H. & Albert, M. (1978). Human Neurgrgychology. New York: Wiley.
Hermann, B., Wyler, A., & Richley, E. (1988). Wisconsin Card Sorting test

performance in patients with complex partial seizures of temporal lobe origin. Journal of
Clinical____ and Experimenaal Psycholqu, 10 467-476.

_n-n.’

104

Hultsch, D.F., Hertzog, C., Small, B.J., McDonald-Miszczak, L., & Dixon, RA.
(1992). Short-term longitudinal change in cognitive performance in later life. Psychology
and Aging, 2, 571-584.

Haug, H. & Eggers, R. (1991). Morphometry of the human cortex cerebri and
corpus striatum during aging. Neurobiology of Agiag, M, 336-338.

Heilbroner, P.L. & Kemper, T. (1990). Cytoarchitectonic distribution of senile
plaques in three aged monkeys. Acta Neuropathologigt, 81, 60-65.

Houx, P.J., Jolles, J ., & Vreeling, F .W. (1993). Stroop interference: Aging effects
assessed with the Stroop Color-Word Test. Em’ ental Agm' g Research, M, 209-224.

Janowsky, J .S., Shimamura, A.P., & Squire, LR. (1989). Source memory
impairment in patients with frontal lobe lesions. Neuropsychologia, 21, 1043-1056.

Jason, G.W. (1985). Gesture ﬂuency after focal cortical lesions.
Neuropsycholggra, M, 463-481.

Jonides, J., Smith, E.E., Koeppe, R.A., Awh, E., Minoshima, S., & Mintun, MA.
(1993). Spatial working memory in humans as revealed by PET. Nature, 3_6_3, 623-625.

Katzrnan, R. (1991). Education and the prevalence of dementia and Alzheimer’s
disease. Neurology 43 13-20.

,—,

Kausler, D.H., Salthouse, T.A., & Saults, J .S. (1987). Frequency of occurrence
memory over the adult lifespan. Experimental Aging Research, 13, 159-161.

Kimberg, D.Y. & Farah, M.J. (1993). A uniﬁed account of cognitive impairments
following frontal lobe damage: The role of working memory in complex, organized
behavior. Journal of Experimenaal Psychology: Genera, 4, 411-428.

King, D.A., Cox, 0, Lyness, J .M., and Caine, ED. (1995). Neuropsychological
effects of depression and age in an elderly sample: A conﬁrmatory study.
Neuropsycholagy, 2, 399-408.

Knight, R.T. & Grabowecky, M. (1995). Escape from linear time: Prefrontal
cortex and conscious experience. In N.S. Gazzaniga (Ed.), The Cogm'tive NeurosMciences
(pp. 1357-1371). Cambridge, MA: MIT Press.

Koenig, H., Meador, K., Cohen, H., & Blazer, D. (1988). Self-rated depression

scales and screening for major depression in the older hospitalized patient with medical
illness. Journal of the American Geriatric Society. 36, 699-706.

105

Koss, E., Ober, B., Delis, D., & Friedland, R. (1984). The Stroop Color and
Word Test: Indicator of dementia severity. International 10ml of Neuroscience. M,
53-61.

Kuhl, D.E., Metter, E.J., Riege, W.H., & Hawkins, RA. (1984). The effect of
normal aging on patterns of local cerebral glucose utilization. Annals of Neurolcgy, M,
8133-8137.

LaRue, A. (1992). Cognition in normal aging. In Aging and Neurcpsycholo 'cal
Assessment (pp. 47-77). New York: Plenum Press.

Levine, B., Stuss, D.T., & Milberg, WP. (1997). Effects of aging on conditional
associative learning: Process analyses and comparison with focal frontal lesions.
Neuropsychology, 1 1 367-381.

—,

Lezak, MD. (1995). Neuropsychologiaal Assess_ment (3rd ed.). New York:
Oxford University Press.

Libon, D.J., Glosser, G., Malamut, B.L., Kaplan, E., Goldberg, E., Swenson, R.,
& Prouty Sands, L. (1994). Age, executive ﬁrnctions, and visuospatial functioning in
healthy older adults. Neuropsychology, 8, 38-43.

Lindenberger, U., Mayr, U., & Kliegl, R. (1993). Speed and intelligence in old
age. Psychology and Agipg, 8, 207-220.

Lowe, P.A. & Reynolds, CR. (1999). Age, gender, and education may have little
inﬂuence on error patterns in the assessment of set-shifting and rule induction among
normal elderly. Archives of Cliniaal Neurmpaycholagy, M, 303-315.

Luria, AR. (1973). The Working Brain: An Introduction to Neurmpsychology.
New York: Basic Books.

Luria, AR. (1976). The Neuropsychology of Memogg. New York: John Wiley.
Luria, AR. (1980). we Cortical Function in Ma_n. New York: Basic Books.

Malloy, P., Bihrle, A., Duffy, J ., & Cirnino, C. (1993). The orbitomedial frontal
system. Archives of Clinical Neuropsychology, 8, 185-201.

Mangles, J .A. (1997). Strategic processing and memory for temporal order in
patients with frontal lobe lesions. Neuropgrcholggy, M, 207-221.

Marcusson, J., Morgan, D.G., Winblad, B., & Finch, CE. (1984). Aging and
serotonin receptors in the human brain. Clinical Neuropharmacology, _7_, 38-39.

106

Martin, A., Wiggs, C.L., LaLonde, F., & Mack, C. (1994). Word retrieval to
letter and semantic cues: A double dissociation in normal subjects using interference
tasks. Neuropsychologia, M, 1487-1494.

McDaniel, M.A. & Einstein, GO. (1992). Aging and prospective memory: Basic
ﬁndings and practical implications. Advances in Learning and Behaviﬂ Disabilities 1,
87-105.

McIntyre, J.S. & Craik, F .I.M. (1987). Age differences in memory for item and
source information. Canadian JournM of Psychology, 41, 175-192.

Mesulam, M.M. (1998). From sensation to cognition. Brain, ﬂ, 1013-1052.

Meyer, J .S., Kawamura, J ., & Terayarna, Y. (1994). Cerebral blood ﬂow and
metabolism with normal and abnormal aging. In M.L. Albert & J .E. Knoefel (Eds),
Clinical Neurology of Aging. New York: Oxford University Press.

Miceli, G., Caltagirone, C., Gainotti, G., Masullo, C., & Silveri, MC. (1981).
Neuropsychological correlates of localized cerebral lesions in non-aphasic brain-damaged
patients. Journal of Clinical Neuropsychology, 8, 53-63.

Milner, B. (1963). Effects of diﬂ’erent brain lesions on card sorting. Archives of
Neurology, 2, 90-100.

Mittenberg, W., Seidenberg, M., O'Leary, D.S., & DiGiulio, D.V. (1989).
Changes in cerebral functioning associated with normal aging. Journal of Clinical and
Eaperimental Neuropgchology, M, 918-932.

Mortiner, J .A. & Graves, AB. (1993). Education and other socioeconomic
determinants of dementia and Alzheimer’s disease. Neurology, 13, 39-44.

Moscovitch, M. (1994). Cognitive resources and dual-task interference eﬂ‘ects at

retrieval in normal people: The role of the frontal lobes and medial temporal cortex.
Neuropsycholgy, 8, 524-534.

Moscovitch, M. & Winocur, G. (1995). Frontal lobes, memory, and aging.
Annals of the New York Academy of Sciences. 769. 119-150.

Nyberg, L, Winocur, G., & Moscovitch, M. (1997). Correlation between frontal
lobe ﬁrnctioning and explicit and implicit stern completion in healthy elderly.
Neuropgchology, M, 70-76.

O'Donnell, J .P., MacGregor, L.A., Dabrowski, J.J., Oestreicher, J .M., & Romero,
J .J . (1994). Construct validity of neuropsychological tests of conceptual and attentional
abilities. Journal of Clinical Psychology 50 596-600.

’—’

107

Panek, P.E., Rush, M.C., & Slade, LA. 91983). Locus of the age-Stroop
interference relationship. Journal of Genetic Psycholgy, Ma, 209-216.

Parasuraman, R. & Giarnbra, L. (1991). Skill development in vigilance: Effects of
event rate and age. Pachology and gang, 8, 155-169.

Parkin, AJ. (1998). The central executive does not exist. Joumal of the
International Neuropsychological Sociag, 5, 518-522.

Parkin, A.J. & Walter, BM. (1991). Aging, short-term memory, and frontal
dysfunction. Psychobiology, 1_9, 175-179.

Parkin, A.J. & Walter, BM. (1992). Recollective Experience, Normal Aging, and
Frontal Dysﬁrnction. chhology and Aging, 1, 290-298.

Parkin, Al, Walter, B.M., & Hunkin, NM. (1995). Relationships between
normal aging, frontal lobe ﬁrnction, and memory for temporal and spatial information.

Neuropggchology, _9_, 304-312.

Parks, R.W., Loewenstein, D.A., Dodrill, K.L., Barker, W.W., Yoshii, R., Chang,
J .Y., Emran, A., Apicella, A., Sherarnata, W.A, & Duara, R. (1988). Cerbral metabolic
effects of a verbal ﬂuency test: A PET scan study. Journal of Clinical and Experimental
Neuropsychology, 10 565-575.

—)

Parrnalee, P., Katz, 1., & Lawton, M. (1989). Depression among institutionalized
aged: Assessment and prevalence estimation. Journal of Gerontology, M, 22-29.

Petrides, M. (1985). Deﬁcits on conditional associative-learning tasks after
frontal-and temporal-lobe lesion in man. Neurop' sychologia, 23, 601 -614.

Petrides, M., Alivisatos, B., & Evans, AC. (1995). Functional activation of the
human ventrolateral frontal cortex during mnemonic retrieval of verbal information.
Proceeds from the National Academy of Science USA. 22, 5803-5807.

Phillips, L.H. (1997). Do ‘frontal tests’ measure executive function? Issues of
assessment and evidence ﬁ'om ﬂuency tests. In P.M.A. Ribbitt (Ed), Methodology of
frontal and executive ﬁrnction (pp. 191-214). Hove: Psychology Press.

Posner, M.I. & Raichle, ME. (1996). Images of Mind (pp. 153-179). New York:
Scientiﬁc American Library.

Powell, D.H. & Whitla, D.K. (1994). Proﬁles in Cogm'tive Agiag (pp. 112-133).
Cambridge: Harvard University Press.

108

Pribram, K.H. (1997). The work in working memory. In N.A. Krasnegor, G.
Reid Lyon, & P.S. Goldman-Rakic (Eds), Development of the Prefrontal Cortex (pp.
359-376). Baltimore: Paul H. Brooks Publishing Company.

Prigatano, G. (1978). Wechsler Memory Scale: A seletive review of the
literature. Journal of Clinical Psychology, M, 816-832.

Raglund, J.D., Glahn, D.C., Gur, RC, Censits, D.M., Smith, R.J., Mozley, P.D.,
Alavi, A, & Gur, RE. (1997). PET regional cerebral blood ﬂow change during working

and declarative memory: Relationship with task performance. Neuropsychology, M, 222-
231.

Raykov, T., Tomer, A., & Nesselroade, J .R. (1991). Reporting structural equation
modeling results in Psychology and Aging: Some Proposed Guildelines. Psychology and
agigg, _6, 499-503.

Raz, N., Gunning, F .M., & Acker, J .D. (1998). Neuroanatomical correlates of
cognitive aging: Evidence from structural magnetic resonance imaging.
Neuropsychology, M, 95-114.

Raz, N., Gunning, F .M., Head, D., Dupuis, J.H., McQuain, J., Briggs, S.D.,
Loken, W.J., Thorton, A.E., & Acker, J .D. (1997). Selective aging of the human cerebral
cortex observed in vivo: Differential vulnerability of the prefrontal gray matter. Cerebral
Cortex, 1, 268-282.

Regard, M. (1981). Cognitive rigidity and ﬂexibility: A neuropsychological study.
Unpublished Ph.D. Dissertation, University of Victoria, British Columbia.

Reitan, R.M.& Wolfson, D. (1994). A selective and critical review of
neuropsychological deﬁcits and the frontal lobes. Neuropgchology Review, 5, 161-198.

Renter-Lorenz, P.A., Jonides, J ., Smith, E.E., Hartley, A., Miller, A., Marshuetz,
C., & Koeppe, RA. (2000). Age differences in the frontal lateralization of verbal and
spatial working memory revealed by PET. Journal of Cogm'tive Neuroscience. M, 174-
l 86.

Reynolds, CR (1997). Forward and backward memory span should not be
combined for clinical analysis. Archives of Clim'cal Neuropsycholagy, 12, 29-40.

Robbins, T.W., James, M.,, Owen, A.M., Sahakian, B.J., Lawrence, AD,
Mcinnes, L., & Rabbitt, P. (1998). A study of performance on tests ﬁ'om the CANT AB
battery sensitive to frontal lobe dysﬁrnction in a large sample of normal volunteers:
Implications for theories of executive ﬁrnctioning and cognitive aging. J ouraal of the
International Neurgrsychological Socim, a, 474-490.

109

Roland, PE. (1984). Metabolic measurements of the working frontal cortex in
man. Trends in Neuroscience. 2, 430-435.

Rubin, DC. (1999). Frontal-striatal circuits in cognitive aging: Evidence for
caudate involvement. Aging, Neuropsychology, and Cogra'tion, a, 241-259.

Rubin, E.H., Kinscherf, DA, Grant, E.A., & Storandt, M. (1991). The inﬂuence
of major depression on clinical and psychometric assessment of SDAT. Ameriaan Journal
of Psychiatg, M, 1 164-1171.

Ruff, R.M., Allen, C.C., Farrow, C.E., Niemann, H., & Wylie, T. (1994). Figural
ﬂuency: Differential impairment in patients with left versus right frontal lobe lesions.
Archives of Clinical Neuropsychology, 2, 41-55.

Ryan, J.J., Lopez, S.J., & Paolo, AM. (1996). Digit span performance of persons
75-96 years of age: Base rates and associations with selected demographic variables.
Psychological Assess_ment. 8, 324-327.

Salthouse, TA (1985). Speed of behavior and its implications for cognition. In J.
Bierren & K. Schaie (Eds), ﬁndbooMkof the PsychologjLangiag (2nd edition), New
York: Van Nostrand Reinhold.

Salthouse, TA. (1994). The aging of working memory. Neuropsychology, 8,
535-543.

Salthouse, T.A., Fristoe, N., & Rhee, SH. (1996). How localized are age-related
effects on neuropsychological measures? Neuropsychology, 10 272-285.

—,

Sarno, M.T., Buonaguro, A. & Levita, E. (1985). Gender and recovery from
aphasia after stroke. Journal of Nervous aad Mental Diseaae, 1_7_3, 605-609.

Schacter, BL. (1987). Memory, amnesia, and frontal lobe dysfunction.
Psychobiology, 1_5, 21-36.

Schacter, D.L., Kaszniak, A.W., Kilrlstrom, J .F., & Valdiserri, M. (1991). The
relation between source memory and aging. chhology and Agag,‘ 6, 559-568.

Schacter, D.L., Osowiecki, 1)., Kaszniak, A.W., Kihlstrom, J.F., & Valdiserri, M.
(1994). Source memory: Extending the boundaries of age-related deﬁcits. Psychology
shim; 2, 81-89-

Scheibel, M.B., Lindsay, R. 1)., Torniyasu, U., & Scheibel, AB. (1975).

Progressive dendritic changes in the aging human cortex. Experimen_t_a1 Neurology, 42,
392-403.

110

Schmidt, M., Trueblood, W., Merwin, M., & Durham, KL. (1994). How much
do 'attention' test tell us? Arches of Clinical Neuropsychology, 2, 383-394.

Schumacher, E.H., Lauber, E., Awh, E., Jonides, J., & Smith, BE. (1996). PET
evidence for an arnodal verbal working memory system. Neuroimage, 8, 79-88.

Seidman, L.J., Oscar-Bermarr, M., Kalinowski, A.G., Ajilore, O., Kremen, W.S.,
Faraone, S.V., & Tsuang, MT. (1995). Experimental and clinical neuropsychological
measures of prefrontal dysﬁrnction in schizophrenia. Neuropsychology, 2, 481-490.

Shallice, T. & Burgess, P. (1991). Deﬁcits in stretegy application following frontal
lobe damage in man. Brain, M, 727-741.

 

Shaw, T.G., Mortel, K.F., Meyer, J.S. Rogers, R.L., Hardenberg, J ., & Cutaia,
M.M. (1984). Cerebral blood ﬂow changes in benign aging and cerebrovascular disease.
Neurology, M, 855-862.

Shimamura, AP. (1994). Neuropsychological perspectives on memory and
cognitive decline in normal aging. Seminars in the Neurosciences _6, 387-394.

Shimamura, AP. (1995). Memory and frontal lobe ﬁrnction. In M.S. Gazzaniga
(Ed), The Cogriitive Neurosciences (pp. 803-813). Cambridge, MA: MIT Press.

Shimamura, AP. & Jurica, PI. (1994). Memory interference effects and aging:
Findings from a test of ﬁontal lobe ﬁrnction. Neuropsycholcgy, 8, 408-412.

Smith, E.E., Jonides, J., Marshuetz, C., & Koeppe, RA. (1998). Compnents of
verbal working memory: Evidence ﬁ'om neuroimaging. Procedes from the National
_A_cademy of Sciences. 25, 876-882.

Squire, LR (1987). Memogaad Brai_n. New York: Oxford University Press.

Spencer, W.D. & Raz, N. (1994). Memory for facts, source, and context: Can
frontal lobe dysfunction explain age-related diﬂ‘erences? Psychology and Aging, 2, 149-
159.

Spreen, O. & Strauss, E. (1991 & 1998). A Compendium of Neuropsychological
Tests. New York: Oxford University Press, Inc.

 

Spieler, D.H., Balota, D.A, & Faust, ME. (1996). Stroop performance in healthy
younger and older adults and in individuals with dementia of the Alzheimer’s type. Journal
of Experimen_tal Psychology: Human Perception and Performance, 22, 461-479.

111

Stern, Y., Gurland, B., Tatemichi, T.K., Tang, M.X., Wilder, D., & Mayeux, R
(1994). Inﬂuence of education and occupation on the incidence of Alzheimer’s disease.
Joumﬂf the American MedicaMXssociartion. 271. 1004-1010.

Struble, R.G., Price, D.L., Cork., L.A., & Price, D.L. (1985). Senile plaques in
the cortex of aged normal monkeys. Brain Research, _3_6_1_, 267-275.

Stuss, D.T. (1991). Interference effects on memory functions in post-leucotomy
patients: An attentional perspective. In H.S. Levin, H.M. Eisenburg, & AL. Benton
(Eds), FrorLtal Lobe Function and Dysfunction. New York: IRBN Press.

Stuss, D.T. & Benson, D. F. (1986). The Fron_tal Lobes. New York: Raven
Press.

Stuss, D.T. & Benson, DR (1984). Neuropsychological studies of the ﬁontal
lobes. chhological Bulletin, 28, 3-28.

Stuss, D.T., Stethem, L.L., & Poirier, CA. (1987). Comparison of three tests of
attention and rapid information processing across six age groups. The Clinical

Neuropgchologist, 1, 139-152.

Suhara, T., Fukuda, H., Inoue, O., Itoh, T., Suzuki, K., Yarnasaki, T., & Tateno,
Y. (1991). Age-related changes in human D1 dopamine receptors measured by positron
emission tomography. Psychopharmacology, 1M, 41-45.

Sullivan, E.V., Mathalon, D.H., Zipursky, R.B., Kersteen-Tucker, Z., Knight.
R.T., & Pfefferbaum, A. (1993). Factors of the Wisconsin Card Sorting Test as measures
of ﬁ'ontal-lobe ﬁrnction in schizophrenia and in chronic alcoholism. Psychiat_ry Research,
4_6, 175-199.

Tayor, L. (1979). Psychological assessment of neurological patients. In T.
Rasmussen and R. Marino (Eds), Functional Neurosuagm. New York: Raven Press.

Tulving, E., Kapur, S., Craik, F .I.M., & Moscovitch, M.H.S. (1994). Hemispheric
encoding/retrieval asynrrnetry in episodic memory: Positron emission tomography
ﬁndings. ProceedirMs of the National Academy of Sciences, 21, 2016-2020.

Tun, P.A., Wingﬁeld, A., Stine, E.A.L., & Mecsas, C. (1992). Rapid speech
processing and divided attention: Processing rate versus processing resources as an
explanation of age effects. Psychplogy and Agai' g, 1, 546-550.

Uchiyama, C.L., Mitrushina, M.N., D'Elia, L.F., Satz, P., & Mathews, A. (1994).
Frontal lobe functioning in geriatric and non-geriatric samples: An argument of
multimodal analyses. Archives of Clinical Neuropsychology, 2, 215-227.

 

112

Uttl, B. & Graf, P. (1997). Color-word Stroop test performance across the adult
life span. Journal of Clinical and ExperimentalMNeurgnsychology, M, 405-420.

Vendrell, P., Junque, C., Pujol, J., Jurado, M.A., Molet, J., & Graﬁnarn, J. (1995).
The role of prefrontal regions in the Stroop Task. Neuropsychologia, 8, 341-352.

Warren, L.R., Butler, R.W., Katholi, C.R., & Halsey, J.H. (1985). Age diﬂ‘erences
in cerebral blood ﬂow during rest and during mental activation measurements with and
without monetary incentive. Journal of Gerontology, 30 53-59.

—)

Wechsler, D. (1987). Wechs_ler Memory che-Revised. New York:
Psychological Corporation.

Wecker, N.S., Kramer, J.H., Wisniewski, A., Delis, D.C., & Kaplarn, E. (2000).
Age effects on executive ability. Neuropsychology, M, 409-414.

West, R. (2000). Reply to Greenwood. J ournai of the International
Neuropsychological Socie_ty, 8, 705-726.

West, R. (1996). An application of prefrontal cortex ﬁrnction theory to cognitive
aging. Psychological Bulletin. 120. 272-292.

West, R. & Bell, M.A. (1997). Stroop color-word interference and
electroencephalogram activation: Evidence for age-related decline of the anterior
attentional system. Neuropaychology, M, 421-427.

Wiggs, CL. (1993). Aging and memory for frequency of occurrence of novel,
visual stimuli: Direct and indirect measures. Pachology and Agmg,’ 8, 400-410.

Wilkins, A.J., Shallice, T., & McCarthy, R. (1987). Frontal lesions and sustained
attention. Neuropsycholog'a, 2_5, 359—365.

Winocur, G., Moscovitch, M., & Stuss, D.T. (1996). Explicit and implicit memory
in the elderly: Evidence for double dissociation involving medial temporal- and frontal-
lobe functions. Neuropsychology, 10, 57-65.

Wong, D.F., Wagner, H.N., Dannals, RF, Links, J.M., Frost, J.J., Ravert, H.T.,
Wilson, A.A., Rosenbaunn, AE., Gjedde, A., Douglasse, K.H., Petronis, J.D., & Folsteirn,
MR (1984). Eﬂ’ects of age on dopamine and serotonin receptors measured by positron
tomogaphy in the living human brain. Science, 226, 1393-1396.

Yesavage, J., Brink, T., Rose, T., Lunn, O., Huange, V., Adey, M., & Leirer, V.

(1983). Development and validation of a Geriatric Depression Screening Scale: A
preliminary report. Journal of Psychiatric Researah, M, 37-49.

113