 

H...

”mm"...
WNW

J

3 ..\\l 5.1. inning-$0..
i. 81...... 39?..nanmu5...
. 5.1.3. I...)
2a....Jnuﬁé .1
..... wﬁd.#§1..._ . .a

..
z ..

c. 2. F1...:.«.......3ﬂ.
:41... ﬁ.....u...»...ﬁr

k

zﬁiwrvrzn...‘

: . .
«in. .1:

.3251... , .51.}.
glil.ova-u“~ n.13'l‘l ‘1"; I‘V:
1.9.”.4 3.2!} x . I z. A... a.
Sivan x...
a: xiii.
..

s .331 \r...
arguing
.413...
I42.-

.
5..
.- .iu'V)
l. :3‘.1-‘d
. aﬁJﬁ é .
v1: .‘. . O 43-
if...“

513355.}!
1.231.}... .1

33:51... In...

1.. .14 .0 $3.352}...

r. .1..vﬂt!.:r£t. ..‘2.!

.v: , 11.1.33... .2... 3 .(Il
. r355:

x.
a... ; .13
.5 £551.15... .2. .
Larizrunrxahﬂdc. .73...‘ .. I. .v
. . . . i . ..
multixaL .1“: . 2”... tilt”...
5.. .5. .o!!.
. rzsﬂrnld} a

a.» 5...... , .33....

v I .c 2.531.... .
.3! I £ 1.. . . , , . .

 

THESIS

2500 llllllllllHlllliHllﬂlllllllllitllllHIIUIHIIHHIHIIIHH

LIBRARY 31293 02048 8916
MWhigan State
University

 

 

 

 

This is to certify that the -
thesis entitled

The Role of Inhibition in Working Memory Performance
Associated with Age
presented by

Carol Catherine Persad

has been accepted towards fulﬁllment
of the requirements for

_._h.D_.___dP egree in Jazmhology

M KM

Major professor

 

Date Aéfl '23. 3000

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

 

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11m chIRCIDaIaDmpBS—p.“

THE ROLE OF INHIBITION IN WORKING MEMORY PERFORMANCE
ASSOCIATED WITH AGE
By

Carol Catherine Persad

A DISSERTATION
Submitted to
Michigan State University
In partial fulﬁllment of the requirements
For the degree of
DOCTOR OF PHILOSOPHY
DEPARTMENT OF PSYCHOLOGY

2000

ABSTRACT
THE ROLE OF INHIBITION IN WORKING MEMORY PERFORMANCE
ASSOCIATED WITH AGE
By

Carol Catherine Persad

The current investigation addressed the view that inefﬁcient inhibitory processes account
for some of the age-related declines in performance on measures of working memory
(WM). Two studies were designed that incorporated WM measures chosen from the
neuroimaging literature. Study I examined age differences between young (18-29) a
young-old (60—74) and an old-old group (75+) on versions of the n-back and item
recognition tasks (J onides et al. 1998; Smith & J onides, 1997). The tasks included critical
trials for which it was presumed that inhibitory processes were particularly important for
successful performance. According to the inhibitory deﬁcit view, older participants
should show particular difﬁculty on these trials. Results generally supported this
hypothesis. Not only did the older adults have more difﬁculty than the young adults on
these tasks overall, they also showed relatively more difﬁculty on those trials that were
theorized to rely more on the integrity of inhibitory processes for successful performance.
No differences were found between the two Older groups. Study 1 also addressed the
suggestion of some researchers that inhibition is partly subserved by the prefrontal cortex
(PFC). In particular, performance on the WCST, often described as a putative measure of
frontal lobe functioning was correlated with performance in the 2-back version of the u-

back task, lending some support for the relationship between the PFC and inhibitory

mechanisms. Study 2 investigated performance on the item recognition task in relation to
age-related information processing speed changes as assessed by two perceptual speed
measures (Salthouse, 1996). Additionally, in an attempt to vary the amount of
interference across trials, Study 2 also included a comparison between categorized and
unrelated word lists. Because of prior strong associative bonds between category
members, it was expected that the categorized list condition would particularly impair
performance in the older adults due to inefﬁcient inhibitory mechanisms. It was also
hypothesized that although age-related changes in processing speed would account for
some of the differences in this working memory task, (a) it would not account for all of
age-related variance, and (b) it would account for less Of the variance in the categorized
list condition than the unrelated condition, presumably because inhibition was also
important to performance, particularly in the categorized condition. The category
manipulation did not achieve the desired results; contrary to expectations, performance
was better in the categorized list condition compared to the unrelated condition.
Hierarchical regression analysis indicated that perceptual speed accounted for much Of
the age-related variance in the item recognition task. Nonetheless, the interpretation of
such ﬁndings as necessarily supporting a processing speed explanation of age-related
changes in cognition is called into question by recent ﬁndings indicating that age
differences on perceptual speed measures are partly a reﬂection of differences in
susceptibility to interference effects from visual distraction. This outcome suggests that
the fundamental factor in regression analysis may be inhibitory efﬁciency rather than
processing Speed. Further studies are needed to evaluate the relative role of processing

speed and inhibition to age changes in performance on WM tasks.

Copyright by
CAROL CATHERINE PERSAD

2000

ACKNOWLEDGMENTS

First and foremost I would like to thank my family for their considerable love, support and
patience. I would like to thank Rose Zacks, my mentor, for her knowledge, guidance and
support, as well as the other members of my dissertation committee for their helpful
comments and suggestions. I would like to thank Ruth Dryden for her invaluable help in
subject recruitment and data collection and Tim Salthouse for providing me with the
perceptual speed measures. Thanks also go to Karin Butler and Carrick Williams for their

assistance in designing the computer programs.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES

INTRODUCTION
Measures of Working Memory
N-Back task
Verbal Item Recognition Task
The Prefrontal Cortex and Inhibition

STUDY ONE
Method
Results
Discussion

STUDY TWO
Method
Results
Discussion

GENERAL DISCUSSION
Role of the Prefrontal Cortex and Inhibition
Inhibition and Processing Speed
Relationship between Inhibitory Processes and age within the older
Age spectrum

FUTURE DIRECTIONS
Working Memory Subsystems and Inhibition

FOOTNOTES

REFERENCES

vi

67

86

LIST OF TABLES

Table 1. Age, Education and Vocabulary scores for the three age groups.

Table 2. Accuracy and Response time data in the l-back condition
for the three age groups.

Table 3. Accuracy and Response time data in the 2-back condition
for the three age groups.

Table 4. Accuracy and Response time data ﬁ'om the Item Recognition
test across the three age groups.

Table 5. Means and Standard deviations for WCST variables for the
three age groups.

Table 6. Age, Education and Vocabulary scores for the ﬁve age groups
in Study 2.

Table 7. Study 2 Accuracy data ﬁ'om the Unrelated and Categorized Item
Recognition test across the ﬁve age groups.

Table 8. Study 2 Response time data ﬁ'om the Unrelated and Categorized
Item Recognition test across the ﬁve age groups.

Table 9. Raw scores from the Perceptual Speed Measures for the
different age groups in Study 2.

Table 10. Results from the hierarchical regression analyses
for the recent negative condition in Study 2.

vii

69

7O

71

72

73

74

75

76

77

78

LIST OF FIGURES

Figure 1. Figural Description of the N—back task.

Figure 2. A Comparison of Accuracy Data in the Negative and Recent Negative
Conditions for the 2-back task.

Figure 3. Study 2 Comparison of Recent Negative and Negative Accuracy
data for the Item Recognition task across the ﬁve age groups.

Figure 4. Response time for the Negative and Recent Negative Conditions
for the Categorized and Unrelated Lists.

Figure 5. Item Recognition Accuracy Performance comparing Young and
Old adults collapsed across studies.

Figure 6. Item Recognition Response Time data comparing Young and
Old adults collapsed across studies.

viii

79

80

81

82

83

84

INTRODUCTION

Age-related deﬁcits have been found on a variety of tasks designed to measure
working memory (Salthouse, 1994; Verhaegen & Salthouse, 1997). Different theories
have been proposed to account for these age differences, including a reduced ability to use
self-initiated processing due to reduced working memory resources (Craik & Anderson,
1999; Craik & Byrd, 1982), a generalized slowing of information processing speed
(Salthouse, 1996), and a reduction in the efﬁciency of inhibitory functioning (Hasher &
Zacks, 1988; Hasher, Zacks, & May, 1999; Zacks & Hasher, 1994; see Light, 1991 for a
review). Although each theory has garnered support, the current investigation focused on
the inhibition theory of aging.

It has been proposed that inhibition, together with excitation, is a necessary
component of attentional processes for successﬁil and eﬁicient execution of a variety of
tasks and behaviors including those that rely on working memory (e.g. Bjork, 1989;
Neumann & DeSchepper, 1992; Tipper, 1985; Tipper, Weaver, & Houghton, 1994). To
this end, inhibition is thought to serve three main purposes in support of eﬂicient working
memory (Hasher et al., 1999). First, by preventing extraneous, distracting information
from entering working memory, inhibition can allow for more efﬁcient processing of target
or goal-relevant information. The distracting information can be ﬁ'om the external
environment as well as from internal thoughts and associations. Along with extraneous
information, previously relevant information can also cause interference if it remains
activated in working memory beyond its applicability to the task at hand. In a constantly

changing environment, task demands ﬁequently change or new goals are chosen, making

stimuli that were originally the focus of processing no longer relevant to the end goal.
Thus, inhibitory processes, by suppressing or deleting old information ﬁ'om working
memory, allow for a re-focusing on new information and aid in reducing possible
interference eﬂ‘ects that could otherwise occur given simultaneous activation of irrelevant
stimuli. Lastly, inhibitory processes can prevent the occurrence of prepotent responses to
stimuli/situations that may not be appropriate to the current context. When inhibitory
processes become inefﬁcient, irrelevant and potentially distracting information enters or
remains in working memory, making it more difﬁcult to carry out other cognitive
processes effectively, presumably due to interference eﬂ‘ects and response competition.

Hasher and lacks (Hasher & Zacks, 1988; Hasher et al., 1999; Zacks & Hasher,
1994) have proposed that inefﬁcient inhibitory processes are one mechanism that accounts
in part for a variety of age-related cognitive deﬁcits. If inhibitory functioning is less
effective in old adults, in comparison to young adults, irrelevant and potentially interfering
information can enter or remain in working memory, making it more difﬁcult to efﬁciently
execute the necessary processing for a given task. Because working memory has been
implicated in a wide variety of cognitive tasks, deﬁcient inhibitory ﬁinctioning can
potentially affect older adults’ performance in many cognitive domains.

There is considerable support for the inhibition theory of aging, although there are
notable exceptions. Research ﬁndings have shown that inefﬁcient inhibitory processes
affect performance on a variety of tasks in old adults. These include language and reading
comprehension skills (Connelly, Hasher & Zacks, 1991; Hamm & Hasher, 1992; Hartman
& Hasher, 1991; See & Ryan, 1995), selective attention processes as measured by

negative priming procedures (Hasher, Stoltzfus, Zacks & Rypma, 1991; May, Kane, &

Hasher, 1995), the ability to inhibit a prepotent response as measured by performance on
the Stroop task (Dempster, 1992) and the antisaccade task (Butler, Zacks, & Henderson,
1999), and the ability to learn and retrieve multiple associations to various facts (Gerard,
lacks, Hasher, & Radvansky, 1991). Older adults do not always exhibit impaired
inhibition, however. For example, although there is considerable evidence demonstrating
reduced negative priming in older adults, there are also an equal number of studies that
have not found these age diﬁ‘erences, particularly under varying task conditions (Langley,
Overmier, Knopman, & Prod’Homme, 1998; Schooler, Neumann, Caplan, & Roberts,
1997; Sullivan & Faust, 1993). In addition, older adults have been found to do as well as
younger adults on a variant of the Stroop task (Li & Bosman, 1996), and on tasks
requiring the individual to ignore distracting information based on spatial cues (Carlson,
Hasher, Connelly & Zacks, 1995; Connelly & Hasher, 1993). These exceptions have led
some researchers to suggest that there may be multiple inhibitory mechanisms, only some
of which are age sensitive (Carlson, Hasher, Connelly & Zacks, 1995; Kramer, Humphrey,
Larish, Logan, & Strayer, 1994; Li & Bosman, 1996).

Older adults do appear to have particular difﬁculty suppressing or deleting
information from working memory that is no longer appropriate for task completion.
Certain types of errors made by older adults provide indirect evidence of this difﬁculty.
For instance, older adults have greater difﬁculty on tasks of cognitive ﬂexibility that
require switching from one strategy or behavior to another. On these tasks, once a
response set has been established, older adults experience difﬁculty altering their behavior
when task conditions change. The result is an increase in perseverative responses (Cronin-

Golomb, 1990; Lezak, 1995; Veroﬁ‘, 1980). The combination of diﬂiculty in shifting set

and perseveration is consistent with the theory of a reduced suppression mechanism. Older
adults also exhibit more intrusions of material from earlier trials during tests of free recall,
a ﬁnding that is also suggestive of a deﬁcient suppression mechanism (Koriat, Ben-Zur, &
Sheﬁ‘er, 1988; Larrabee, Trahan, Curtiss, & Levin, 1988). In addition a recent study has
demonstrated that performance on traditional working memory span measures, generally
used as an indication of an individual’s capacity of working memory, is signiﬁcantly
inﬂuenced by a build up of proactive interference (May, Hasher & Kane, 1999), especially
for older adults.

Other studies have tried to investigate the suppression mechanism more directly.
For example, reading tasks that make use of garden-path sentences (Hartman & Hasher,
1991) and passages (Hamm & Hasher, 1992) provide results consistent with the view that
efﬁciency of inhibitory processes declines with age. The materials are called "garden path"
texts because they are designed to initially mislead the reader to arrive at one
interpretation, which later turns out to be incorrect. Although both young and old adults
do arrive at and maintain the correct interpretation, unlike young adults, the old adults
also maintain the incorrect interpretation. These ﬁndings have been demonstrated both on
indirect tests (sentence completion tests) and direct tests of memory (speeded judgment
tasks).

Zacks, Radvansky and Hasher (1996) used a directed forgetting paradigm to
investigate suppression ability in older adults. In the typical directed forgetting procedure
a list of words is presented to the participant. After each word or aﬁer a block of words
has been presented, a cue is given that tells the subject whether the word(s) are to be

remembered or forgotten. Because the cues are presented aﬁer the word has been

presented and because the subject does not know until then whether the particular item is
to be remembered or forgotten, it is assumed that the person must attend to each word
and process it. Once the entire list has been presented, participants are asked to recall all
of the to-be remembered words (TBR). The assumption is that once the forget cue is
presented, an individual must delete or suppress those words in working memory so that
these items (the "to-be—forgotten words," or TBF) will not intrude or interfere with recall
of the TBR words. On such tasks, not only do young adults have little diﬂiculty recalling
TBR words and give few intrusions of TBF words (Bjork, 1989; Geiselman & Bagheri,
1985), their recall of the TBR words is equivalent to trials in which the TBR words were
presented alone. Performance levels are as if the TBF words were never presented
supporting the view that the young adults successfully suppress the TBF items to the
extent that these items do not interfere with recall of the TBR words. In contrast, old
adults exhibited higher intrusion rates of TBF words at recall in comparison to trials in
which the TBR words are presented alone (Zacks, et al., 1996). The older adults appeared
to have problems suppressing the irrelevant information leading to reduced performance.
These results suggest that the inability to suppress TBF words can affect recall
performance in two ways; by increasing the difficulty in correctly recalling the target
material and by increasing the number of intrusions of the irrelevant material.

Taken together, these results support the view that older individuals have difﬁculty
suppressing material that is no longer relevant in working memory. As a result,
interference can occur with the presence of this distracting material, leading to reduced
performance at both the information processing/encoding stages (as seen in language

comprehension) as well as at the time of retrieval (for example, lowered performance on

directed forgetting tasks). The result of impaired inhibitory processing can be manifested
in a variety of ways including increased errors of omission, perseverations, and/or
intrusions.

The present research was designed to further study the role of inhibition in
cognitive changes associated with advancing age. Because working memory has been
hypothesized to be essential for many cognitive tasks, such as language comprehension,
memory, and reasoning (Baddeley & Hitch, 1994; Just & Carpenter, 1992), the choice in
the current investigation was to focus on age-related changes in working memory and the
role of inhibition. To this end, two working memory measures were chosen. It will be
argued that efﬁcient inhibitory processes are a necessary component to successﬁrl
performance on these tasks. Therefore, because the position taken for the current research
is that older individuals have deﬁcient inhibitory functions, it was predicted that they
would show poorer performance on these working memory tasks, as compared to younger
adults. In addition to an overall age-related decrement in performance, reduced inhibitory
functioning would predict a speciﬁc pattern of responses to certain types of foils on these
tasks. The analyses of these responses can provide stronger evidence that age-deﬁcits in
inhibition partially cause reduced performance on these tasks. The speciﬁc type of errors
that were expected will be outlined later during the discussion of the working memory
tasks.

Two studies were conducted in order to investigate more fully the role of
inhibition in working memory performance. The ﬁrst study had a two-fold goal. The ﬁrst
was to compare performance differences between young and old individuals on two

working memory measures, with the expectation of age-differences in performance.

Although an overall age decline might be expected, it was hypothesized that inhibitory
processes are more instrumental to performance in certain conditions of the tasks; thus,
greater age-related changes were expected in these particular conditions. The second aim
goal was to look for performance differences within an older sample to further understand
the role of inhibition within this age group. This part of the study was a follow-up to an
earlier study that demonstrated continuous changes in inhibitory functioning within an
older age group (Persad, Abeles, Zacks, & Denburg, in press). A growing body of
research suggests that performance on a number of cognitive and neuropsychological
measures continues to decline with advancing age (Christensen, MacKinnon, J orm &
Henderson, 1994; Osterweil, Mulford, Syndulko, & Martin, 1994). These studies have
generally divided the older population into a young-old group (typically between the ages
of 60 and 74) and an old-old group (individuals over the age of 75). The usual ﬁndings are
lowered performance rates on a variety of tasks in the old-old group as compared to the
young-old group (Christensen et al., 1994; Osterweil et al., 1994; but see Speiler, Balota,
& Faust, 1996). For the current study, older adults were divided into two groups,
classiﬁed as young-old (between the ages of 60 and 74) and old-old (75 years and older).
By assessing young-old and old-old groups, a clearer picture concerning age changes in
inhibitory functioning can be drawn.

Although there is much support for the inhibition theory, it is not the only
explanation of age-related cognitive changes. An alternative theory with considerable
support is the processing speed theory (see Salthouse, 1996, for a review). The second
study aimed to investigate the relative roles of processing speed and inhibition in

accounting for age-related differences on the working memory data, by designing test

conditions that result in different predictions based on the two theories. A review of this

theory and its implications for the inhibition viewpoint will be presented later in the paper.

Measures of Working Memory

The two working memory tasks that were chosen for each study were the n—back
task and the verbal item recognition task. The speciﬁc variations of the task used here
have been used recently in the neuroimaging literature to identify the neuroanatomical
regions that may be involved in working memory in young adults (Smith & Jonides,
1997). In the sections that follow, a brief description of the measures and previous results
from related studies are presented. Then, the hypothesized contribution of inhibition to

performance on these tasks is outlined, and predicted age differences are described.

N-back Task

The general procedure of the n-back task used here was patterned aﬁer one used
by Smith and Jonides in their studies of the neuroanatomical regions involved in working
memory (1997; Smith, Jonides, & Koeppe, 1996). In this task, letters were visually
displayed one at a time in a continuous string. As each letter appeared, the subject
responded (pressed the appropriate key) to indicate whether the current letter was the
same as (i.e. same in identity) or different from one presented n positions earlier in the
sequence. See Figure l for a visual representation of the task. For this study, only two
conditions were administered, requiring a decision based on the match to the item 2-back
or l-back in the series. This task is usually regarded as a complex working memory task

requiring the continual processing of items (Baddeley & Hitch, 1994; Just & Carpenter,

1992). With the presentation of each new letter, there is a need to update the contents of
working memory while remembering the relative location of the appropriate items. In
studies that have used variations of this task, one of the assumptions is that the longer the
lag (i.e. the number of intervening items between the current item and the item presented
n-back in the series), the greater the memory load in working memory (Dobbs & Rule,
1989; Smith & Jonides, 1997).

It is argued here that inhibition is one of the key components to successful
performance on this task. With each new letter, as one updates the contents of working
memory, one must suppress, or in a sense ignore, earlier presented items that are irrelevant
to the current trial. The ability to suppress this information will aid in the efﬁcient search
of the most current letters and selection of the appropriate response. Otherwise, if
suppression does not occur, the additional material in working memory can be distracting,
making it more diﬂicult to make a correct response presumably due to increased
interference effects from this extraneous information that remains activated in working
memory.

To analyze the task in more detail, let us follow the events that occur on a given
trial of a 2-back condition. With the presentation of a new letter, the letter that was 2-back
in the sequence on the previous trial is now in the 3-back condition. This letter is no
longer relevant for the current trial and, it is argued, needs to be suppressed to prevent it
from interfering with the other items currently in working memory. The greater the
amount of distracting and potentially interfering information in working memory, the

greater the potential difficulty in deciding whether a match has occurred. Thus, the

presence of additional material can lead to slower response times and decreased accuracy
on this task.

Given the hypothesis that the ability to suppress previous information declines with
age, the prediction is that older adults will have relatively more difﬁculty deleting the
irrelevant letters. With the resulting cluttering of working memory, older adults will ﬁnd
this task more difﬁcult to perform. Therefore, the n-back task ought to be sensitive to age
effects due to interference from the additional letters that remain activated in working
memory for older adults. This increase in clutter also would be expected to hurt
performance more in the more diﬂicult 2-back condition, due to the larger amounts of
material that need to be maintained and processed in working memory. Thus, larger age
differences are predicted in the 2-back condition than in the l-back condition.

To provide additional evidence for the role of inhibition in this task, an
examination of the responses to different types of foils on the 2-back task can ﬁrrther
illuminate the processes involved. If older .adults have more difﬁculty suppressing/deleting
earlier letters, these letters would remain in working memory causing interference. To test
this hypothesis more directly, on a proportion of trials the probe letter would actually
match the letter 3-back in the sequence. It is expected that when the current letter matches
the one presented 3-back, older adults would show more errors of commission (increase in
"yes" responses) and an increase in response time in comparison to trials when there is no
match. In contrast, because young adults have more effectively inhibited the earlier stimuli,
comparisons between the 3-back matching trials and trials on which there is no match

should yield smaller or no performance differences.

10

Although age diﬁ‘erences have not been investigated on the proposed version of
the n-back task, age differences on tasks that used the basic procedure as the n-back task
have been reported. Dobbs and Rule (1989) administered a variant of the n-back task to
subjects ranging in age from 30 to 97, divided into ﬁve age groups by decade (the oldest
age group was 70 and above). These authors wanted to design a working memory task
that required the constant updating and processing of information. At the same time, they
wanted a task that did not rely on verbal/reading ability like the typical working memory
reading span measures or the computation span measures that rely on basic arithmetic
skills. The researchers therefore presented digits one at a time, and, depending on the
condition, instructed the participants to repeat the digit just heard (0-back), the one just
prior (l-back) or the digit presented 2 trials back (2-back). Participants were administered
all 3 lags, with a possible maximum score of 10 for each condition. They also administered
measures of short-term memory, digit span forwards and backwards, and the Brown-
Peterson task. Results revealed age differences only on the n-back task. Although all age
groups showed a decline in performance with each increasing lag condition, the pattern of
decline varied. The 30-, 40-, and 50-yr-olds showed similar declines across all lag
conditions, whereas the 60-yr-olds performed signiﬁcantly more poorly at lag 1 and 2 than
the younger age groups. By' contrast, the 70+-yr-olds showed even greater declines in
performance than the 60—year-olds. Based on these ﬁndings, the authors concluded that
age declines on the n-back task might be due to a diﬂiculty in the ability to update or
switch processes quickly.

See and Ryan (1995) used the n-back procedure with an older sample to

investigate determinants of language difﬁculties that occur with age. These investigators

11

were interested in studying the relative roles of generalized slowing, inhibition, and
working memory as mediators of age differences on a variety of language tests. In this
study the n-back task was one of the predictor variables, used as a measure of working
memory, whereas the Stroop interference score was used as a measure of inhibitory
functioning. Although the n-back task was not under direct study, the reported raw data
and the data from the hierarchical regression analyses are informative. The n-back
procedure used in that study was patterned after the one used by Dobbs and Rule (1989)
but with two variations. Consonants were used instead of digits, and two additional lags
were tested (3 and 4-back). Both the young and old group showed an overall decline in
performance with increasing lag, with the older individuals exhibiting increasingly more
dimculty across the lag conditions. Furthermore, when the proportion of variance
accounted for by the n-back task was covaried out in a regression analysis, it was found
that this working memory measure mediated some of the age-related decline in language
performance (just as did the measures of speed and inhibition). What is of interest is that
when either the speed or inhibition measures were ﬁrst factored out before the n-back
task, the proportion of age-related variance in working memory performance was no
longer signiﬁcant. This suggests that performance on the n-back task itself may be
mediated in part by both speed of processing and inhibitory functioning.

In summary, there is evidence that old adults, in comparison to young, do
experience more difﬁculty on tasks that presumably require the suppression of earlier
material, as in the n-back task. These ﬁndings are consistent with the hypothesis that
inhibitory functions become deﬁcient with age. It is hypothesized that older adults will

exhibit greater difﬁculty on the n-back task. Unlike previous studies that have only

12

examined overall age diﬁerences in working memory, the use of familiar (i.e. recently
presented) foils in the n-back task in the current study can provide additional evidence for
the role of inhibitory processes (as opposed to processing speed or decreased working
memory capacity; this point will be expanded later in the paper) and allow for a stronger
claim to this effect. In addition, the use of the two older age groups will allow for a more

reﬁned examination of inhibitory functioning within an older sample.

Verbal Item-Recognition task

The verbal item-recognition working memory task (VRWM) modeled aﬁer
Stemberg’s short-term memory task (Stemberg, 1966) is also taken from the
neuroimaging literature (J onides, Smith, Marshuetz, Koeppe, & Renter-Lorenz, 1998;
Smith & Jonides, 1997; Smith et al., 1996). In a typical experiment, on each trial an array
of letters is presented, followed by a probe letter after a speciﬁed delay. Upon presentation
of the probe, one needs to decide whether or not the probe was a member of the target
array. The task requires the participant to maintain the target array in working memory for
the duration of the delay, then compare the letters in memory with the probe, and ﬁnally
make the appropriate response. Stemberg (1966) used this procedure to study short-term
memory search processes and found that as the number of items presented in the target
array increased, the longer it took to respond to the probe. He interpreted this ﬁnding as
support for the idea that memory search occurs serially and that one exhaustively searches
all items in memory before making a response. Many of the earlier studies that have used
this basic procedure to study age differences were interested in presumed serial scanning

abilities. It has generally been found that as target set size increases, older adults take

13

increasingly longer to respond to the probe than younger adults (Anders & Fozard, 1973;
Anders, Fozard, & Lillyquist, I972; Erikson, Hamlin, & Daye, 1973), leading to
conclusions that search processes are slower in older adults. However, these ﬁndings leave
open the question of what underlies the slowed search processes in older adults.

Results from more recent studies suggest that inhibition may play a role in
successful performance on the VRWM. Jonides and his colleagues have used a modiﬁed
version of the Stemberg task to investigate this role (Jonides et al., 1998). Using one set
size, they presented the four target items simultaneously on each trial, with the probe
matching the target array (positive probe) on half of the trials. To study inhibition, two
conditions were designed that manipulated the type of probes in very speciﬁc ways. In the
High-Recency condition, half of the negative probes (those not matching a target item)
were actually members of the target array on the previous trial (designated recent negative
probes). Although a "no" response was currently required, the authors hypothesized that
subjects would have to inhibit a prepotent response tendency to respond "yes" because of
the previous association to the earlier target array. The need to inhibit the 'yes' response
should be demonstrated in longer response times and increased errors on these trials. In
addition, half of the positive probes were also members of the previous target array
(recent positive probes). In this case, because the response was already primed, facilitation
would be expected, manifested in shorter response times. In the Low-Recency condition
probes were not allowed to overlap with targets ﬁ'om the two previous trials, thereby
reducing the need for inhibition. Results generally conﬁrmed the expectations. Response
times were slower and responses less accurate for the recent negative probes in the High-

Recency condition compared to the negative probes in the Low-Recency condition. These

14

results supported the view that inhibitory processes were engaged in this task. However,
facilitation effects (represented by faster response times in the recent positive condition as
compared to the positive condition) were not found.

In a follow-up study, Jonides and colleagues (Jonides, Marshuetz, Smith, Reuter-
Lorenz, Koeppe, & Hartley, 2000) tested a group of older individuals on the same task.
Results demonstrated that the older adults exhibited increased interference effects in the
recent negative condition as compared to the younger adults. The authors interpreted
these ﬁndings to support the notion that inhibitory processes become less efﬁcient with
advancing age.

A similar result was obtained in the directed forgetting study with older adults
reported earlier in this paper (Zacks et al., 1996). Of relevance to this discussion is the use
of an immediate recognition test in the third experiment of this study. Aﬁer each
presentation of the TBF and TBR words, a recognition test was administered. Participants
were instructed to respond yes if a presented probe was in the TBR set of words. On some
trials, the probe was actually a member of the TBF list, whereas on the remaining trials the
probe was a new word. Results showed that younger adults had signiﬁcantly slower
reaction times to the TBF probe as compared to the probe that was neither a TBR nor a
TBF item. These results are similar to Jonides et al.'s ﬁndings (2000) and to other ﬁndings
that have used this recognition procedure to study inhibition in young adults (Bjork,
1989). With respect to the older adults' performance, it was expected that the inefﬁcient
inhibitory processes would result in a greater increase in reaction time when the probe was

a TBF item as compared to a new item. Results conﬁrmed these expectations, supporting

15

the claim that older adults have greater diﬁiculty suppressing information in working
memory that becomes irrelevant to the task.

Although the VRWM task proposed in the current study does not explicitly
instruct the participant to forget items ﬁ'om earlier trials, as in the directed forgetting task,
there is an implicit awareness that one needs to ignore earlier information and instead
attend to the current target array for successﬁrl performance (Bjork & Bjorlg 1996). An
inability to suppress the items ﬁ'om the earlier trials on the VRWM, presumably leaving
these items activated in working memory, can lead to interference effects from the
presence of this additional material. It follows then, that with the presumed deﬁcit in
inhibitory ﬁrnctioning, older adults will have more difficulty on the inhibition trials (i.e.
recent negative), as evidenced by longer response times and lower accuracy. This was the
very result found by Jonides and colleagues (2000).

Although additional studies would have to be conducted, it is possible that
inhibitory deﬁcits could partially account for the reported age deﬁcits from the earlier
studies of search processes (i.e. search rate in the Stemberg task) in older adults. In these
studies, typically, multiple target lists are presented to participants with little delays
between trials. The slowing in response time in older adults may actually partially be a
result of inefficient inhibitory processes rather than a slowed serial search process per se.
Items from the earlier trials may have remained activated in working memory, causing
interference effects and thus making it difﬁcult to respond to the probe (May, Hasher, &
Kane, 1999). It would be important to design studies to control for the effects of
interference in the Stemberg task, before drawing conclusions about search processes in

older adults.

16

In summary, as a result of hypothesized inhibitory deﬁciencies that accompany old
age, it is predicted that older adults will exhibit more diﬂiculty than young adults on two
working memory tasks. Speciﬁcally, performance should be differentially worse in
conditions where inhibitory processes are assumed to be more important for successful
outcomes. Finding a greater decrease in performance on trials that require greater
inhibitory control in both working memory tasks would offer especially strong support for
the inhibition theory of aging deﬁcits in working memory. In contrast, the processing
speed theory (Salthouse, 1996) does not necessarily predict differences between trials that
involve familiar foils (the 3-back match condition in the n-back task, and the inhibition
trials in the VRWM task) as opposed to those trials that don't employ this manipulation. In
addition, supporting previous ﬁndings of a continuous decline in inhibitory ﬁrnctioning
within an older population, old-old individuals are expected to do more poorly than

young-old individuals.

The Prefrontal Cortex and Inhibition

In addition to providing a test of the inhibition deﬁcit hypothesis, the rationale and
design of the current studies were influenced by ﬁndings ﬁom the neurocognitive
literature. Neuroimaging studies suggest that the prefrontal cortex, especially the
dorsolateral aspect, is involved in both working memory and inhibitory ﬁrnctioning. (Chao
& Knight, 1997; Fuster, 1997;Gevins & Cutillo, 1993; Jonides et al., 2000; Jonides et al.,
1998; Schumacher, Lauber, Awh, Jonides, Smith & Koeppe, 1996; Smith & Jonides,
1997; Smith et al., 1996). Both PET and ﬂ\dRI studies have demonstrated increased

activation of the dorsolateral prefrontal cortex (DLPF C) during the n-back task (see Smith

17

& Jonides, 1997 for a review) as part of a larger neural network that has been proposed to
subserve a working memory system. The role of the PFC in inhibitory processes has been
examined more fully with the VRWM task. In the Jonides et a1. study (1998) described
above, young participants showed an increase in activation in the left lateral prefrontal
cortex but only during the inhibition trials (recent negative trials), suggesting a role for the
PFC in inhibition. In contrast, not only did the older adults demonstrate behaviorally more
problems inhibiting responses in the follow-up study (Jonides et al., 2000), no reliable
activation in same prefrontal region was found.

Clinical research studies also lend indirect support for the possible role of the
prefrontal cortex in working memory and inhibitory processes. Findings from lesion
studies have shown that patients with frontal lobe lesions have particular diffrculty
inhibiting responses and suppressing previous response sets (Cronin-Golomb, 1990; Stuss
& Benson, 1984). lmpainrrents in executive functions, including working memory tasks
have also been reliably demonstrated in these patients (Lezak, 1995).

In addition, neuropsychological and neuroanatorrrical evidence suggest that there is
a differential decline in the functioning of the frontal lobes associated with the aging
process. Atrophy is more marked in the frontal lobes compared to other cortical areas
(Coffey, 2000; Haug & Egger, 1991), accompanied by a reduction in cerebral blood ﬂow
(Gur, Gur, Obrist, Skolnick, & Reivitch, 1987; Pietrini & Rapoport, 2000), and older
adults have difﬁculty on neuropsychological tasks that are assumed to reﬂect frontal lobe
functioning (Cronin-Golomb, 1990; Shimamura & Jurica, 1994). Chao and Knight (1997)
have also demonstrated more directly a relationship between inhibitory functioning, age,

and the prefrontal cortex. Using electrophysiological measures, changes in amplitude were

18

demonstrated in the preﬁ'ontal cortex in older adults in conjunction with behavioral
measures of decreased inhibitory functioning.

Although the main focus of this paper is the relationship between inhibitory
processes and working memory performance in older adults, a neuropsychological
measure that has been associated with ﬁ'ontal lobe functioning was administered in Study
1 to provide indirect evidence of the link between the frontal lobes and performance
deﬁcits seen on the working memory measures due to aging. In particular, the Wisconsin
Card Sorting Test (WCST; Heatorr, Chelune, Tally, Kay, & Curtis, 1993; Milner, 1964), a
test that has been used as an indicator of frontal lobe functioning (see Anderson, Damasio,
Jones, & Tranel, 1991, for an opposing view) was used. Recently, PET data have shown
that the DLPFC is activated during the WCST (Berman, Ostrem, Randolph, & Gold,
1995). In a previous study, following the argument of Arbuckle and Gold (1993; Gold &
Arbuckle, 1995), we have argued that inhibition is one of the processes that is measured
by performance on the WCST (Persad et al., in press). As a result, it was expected that
performance on the working memory measures (that are thought to be mediated, in part,
by inhibition) and the WCST would be signiﬁcantly correlated. If this relationship were

borne out, it would provide indirect evidence for the role of the DLPFC in inhibition.

STUDY 1

METHOD

Participants
Twenty-one younger adults under the age of 30 and 41 adults over the age of 65

were tested. The older adults were further divided into two groups based on age. Twenty

19

adults between the ages of 60 and 74 were classiﬁed as young-old and twenty-one adults
over the age of 75 were classiﬁed as old-old. The young adults were undergraduate
students who received course credit for their participation. The older adults were recruited
from the local community and were paid ten dollars for their participation. Subjects were
excluded on the basis of a history of stroke, head injury, signiﬁcant periods of loss of
consciousness, attention deﬁcit disorder or a learning disability, or other health problems
that have been documented to aﬂ‘ect general cognitive functioning. One young adult with
an ADD diagnosis and one old-old adult with carpal tunnel syndrome who was unable to
make the required motor response were excluded, resulting in equal sample sizes across
the three age ranges. In addition, the Mini Mental State Examination (MMSE; Folstein,
Folstein, & McHugh, 1975) was administered to the older adults as a screen for possible
cognitive impairment. On this test a score of 23 or less is an indication of the presence of
cognitive difﬁculties. No older adult was excluded based on this measure; MMSE scores
for the young-old group were 27.2 (SD. = 1.61) and 27.6 (SD. = 1.64) for the old-old
group.

Data on mean age, education and vocabulary ability are provided in Table 1.
Results from a one-way AN OVA demonstrated signiﬁcant differences in education (F
(2,57) = 4.47, MSe = 5.70, p<.00) between the three age groups. Bonferroni post hoc
analyses revealed signiﬁcant differences in education between the young and the old-old
group, with the older adults having achieved a higher level of education. Verbal ability as
measured by the Vocabulary test from the Shipley Institute of Living Scale (Shipley, 1946)
demonstrated signiﬁcant differences between the three groups (F (2, 54) = 9.76, MSe =

17.54, p<.00). Post Hoc analyses showed that vocabulary scores for the old-old group

20

were signiﬁcantly higher than the two other age groups, but the young-old and the young
were not signiﬁcantly different from each other.

In addition, the presence of depressive symptomatology was assessed through self-
report questionnaires. The Beck Depression Inventory (BDI; Beck, 1978) was used with
the young adults and the Geriatric Depression Scale (GDS; Brink et al., 1982) was used
with the older adults. On both instruments a score of 10 or above is used an indication of
depression. Scores on these measures for the three groups were Mean = 3.3 (SD. = 2.28)
for the young group, Mean = 4.9 (SD. = 3.65) for the young-old and Mean = 3.7 (SD. =
2.89) for the old-old group. None of the young received a score above the cut-off, 2 of the
young-old (scores of 12 and 14 respectively, classiﬁcation as mild depressive symptoms)
and 1 of the old-old (score of 11, very mild depressive symptoms) were above the cut-off.
Inspection of the data from these 3 subjects revealed no apparent differences from their
relevant age groups. All analyses were performed with and without including these
subjects with similar results. Thus, to maintain equal n’s across the groups, these 3

subjects were retained in all of the analyses described below.

Measures
Working Memory Measures

Administration of the working memory measures was counterbalanced. Half of the
participants were given the item recognition task ﬁrst, followed by the n-back task; the
other half were given the reverse order. In addition, counterbalancing was done for the n-
back task; the l-back condition was administered ﬁrst for half the subjects, and the other

half were given the 2-back ﬁrst.

21

N-back task
Materials

All 21 consonants were used as stimuli. Five lists of letters were constructed for
the l-back condition, each of which had 36 letters. Within a list, letters were not repeated
unless in the positive or recent negative condition. For each list there were: 8 positive
trials in which the letter matched the one presented just before, or l-back, in the sequence;
22 negative trials, in which the letter did not match the item l-back; 6 recent negative
trials in which the letter actually matched the item presented 2-back in the sequence. The
ﬁrst item in the list was considered ﬁller. This led to a total of 40 positive trials, 110
negative trials, and 30 recent negative trials.

The 2-back stimuli were constructed in a similar fashion. Another ﬁller was added
at the beginning of each list for a total of 37 items. Each string included 8 positive trials; in
this case a positive trial was one in which the letter matched the one presented 2-back in
the sequence. There were 6 recent negative trials deﬁned as trials in which the letter
actually matched one presented 3-back in the sequence, still requiring a “no” response.
Twenty-one negative trials as well as two l-back (foil) trials were included. Just as before,
a l-back trial was deﬁned as one in which the letter matched the l-back. This condition
was added to check that the participant was performing the task, and not just responding
indiscriminately based on simple repetition of a stimulus. Across the ﬁve lists, there were a

total of 40 positive, 105 negative, 30 recent negative and 10 foil trials.

22

Procedure

For both the l and 2-back conditions, a series of practice trials were administered
followed by the test. The practice consisted of three different components. First the
individual was shown a series of numbers printed on a card. The presentation allowed the
participant to see all of the numbers at the same time, while making the appropriate yes/no
response. Then numbers were presented on cards one at a time, and again the subject was
asked to make the appropriate response. Numbers were used initially in an attempt to
reduce possible proactive interference effects. Finally, a practice series of letters was
presented on computer. Feedback regarding performance was given and practice trials
were repeated as necessary until the individual appeared to understand the task and
achieved a perfect score.

Stimuli were presented in the center of a computer screen. Each trial was 3000
msecs in duration: a stimulus presentation of 1000 msecs followed by a 2000 msec delay
during which time a cross was presented on the screen. Participants were instructed to
press the appropriate key on a button box as quickly as possible. For half of the subjects,
the yes response was made with the right index ﬁnger and the no response with the leﬁ
index ﬁnger, with the remaining subjects receiving opposite instructions. Subjects were
instructed to respond “no” to the ﬁrst one (l-back) or two (2-back) letters in a list. A one-
minute rest period was provided between each list presentation. Response times measured

in milliseconds and accuracy were recorded.

23

Verbﬂtem Recrgnition test
Materials

320 four-letter nouns were chosen ﬁ'om the Kuchera and Francis (1967) frequency
norms ranging in ﬁ'equency from 5 to 393. These were used to develop eighty arrays, each
containing four words. Each array was randomly assigned to one of the four experimental
conditions. Each word was presented only once, unless in the interference or facilitation
condition. Four different lists were developed that counterbalanced the stimulus arrays

across the four experimental conditions.

Procedure

Each array was presented in a 2x2 matrix in the center of the computer screen for
2000 msecs. This was followed by the presentation of a cross for 2500 msec. Then the
target word was presented and participants were instructed to decide if the probe word
was in the array. Just as in the n-back task a button press was used for responding; for half
of the individuals the leﬁ button was used for “yes” and the right button was “no” with the
reversed order for the remaining participants. A practice sequence of ﬁve trials was
administered to ensure comprehension.

The 80 arrays were presented in pseudorandom order. For 40 trials the probe
matched one from the stimulus array. On half of these trials, the probe was a member of
only the current array (positive condition). For the other half of these trials the probe was
also a member of the array from the previous trial (recent positive condition). On the

negative trials half of the probes were new words, never presented on any previous trial

24

(negative condition). The remaining probes were members of the previous array, but not

the current one thereby requiring a “no” response (recent negative condition).

Wisconsin Card Sorting Test (WCST; Heaton, Chelune, Tally, Kay, & Curtis, 1993;
Milner, 1964),

The participant was given a set of 128 cards on which were printed one to four
symbols of either stars, crosses, triangles or circles in one of four different colors. The
task was to place each of the cards under one of four stimulus key cards according to a
principle that the person must deduce from the pattern of the examiner's responses to the
person's placement of the cards (i.e. color, form, or number had to be matched). After a
run of ten correct placements in a row, the to-be sorted principle was changed, and this
was indicated to the subject only in the changed pattern of "right" and "wrong" feedback.
This was continued until six category shifts were achieved or all cards were used.
Although a variety of indices can be obtained from this task, ﬁndings from the
neuropsychological literature have demonstrated that the number of perseverative
responses (deﬁned as a response based on a sorting strategy that was previously deemed
incorrect) exhibited on the WCST is most highly correlated with ﬁ'ontal lobe functioning
(Lezak, 1995). The perseverative response score was used in subsequent analyses with

higher scores representing greater impairment.

25

RESULTS

Unless otherwise indicated, signiﬁcance levels were set at p s .05.
N-back task

Reliability estimates were ﬁrst computed for the n-back reaction time results for
each age group separately. Only the response times for correct responses were included in
the reliability analysis. An inter-item analysis was performed by comparing results ﬁ'om the
5 presentation lists and computing an alpha score. Reliability as measured was consistently
high for all three age groups in both the l-back and 2-back conditions. In particular, for
the l-back condition, an alpha of 0.94 was obtained for the young adults, 0.96 for the
young-old, and 0.97 for the old-old. Similar results were obtained for the 2-back task
(alpha=0.96 for the young group, 0.94 for the young-old and 0.94 for the old-old).

Reliability estimates for the accuracy data were computed the same way. For the
1-back condition an alpha of 0.57 was obtained for the young group, 0.84 for the young-
old, and 0.81 for the old-old. The 2-back results were 0.77 for the young, 0.77 for the
young-old, and 0.82 for the old-old. The low reliability estimate for the young adults in the
1-back condition is due to a problem of range restriction in the results. The young adults
were at ceiling, making very few errors and the resulting restriction in variation led to a
underestimate of the reliability of the measure.

Tables 2 and 3 depict the raw response time and accuracy data for the l-back and
2-back tasks across the three age groups. The reduced inhibition theory of aging would
predict poorer performance on this task in the older age groups, particularly in the 2-back

condition. In order to examine this hypothesis, ﬁrst mixed 3 (age) X 2 (n-back task: 1-

26

back vs. 2—back) X 2 (condition: positive vs. negative) repeated measures ANOVA’s were
conducted separately for the response time and accuracy data.

An ANOVA of the accuracy results found all comparisons to be signiﬁcant,
including the three way interaction between age, n-back task, and condition, [F (2,57) =
3.290, MSe=26.91, eta2 = 0.103]. In order to more fully understand the three-way
interaction, two-way ANOVA’s were conducted for the l-back and 2-back task
separately. In the l-back task, only the condition effect was signiﬁcant, [F (1,57) = 75.41,
MSe = 24.32, eta2 = 0.57], indicating that overall participants were less accurate in the
positive condition than the negative condition. No age or age X condition effects were
signiﬁcant. Results were diﬁ‘erent for the 2-back task. A signiﬁcant condition effect, [F
(1,57) = 221.54, MSe =43.73, etaz = 0.80], age effect, [F (2,57) =8.32, MSe =53.34, etaz
= 0.23], and an age X condition effect [F (2,57) = 5.94, MSe =43.73, eta2 = 0.17] were
found. There were no accuracy diﬂ‘erences between the three age groups in the negative
condition; however, both of the older-age groups made signiﬁcantly more errors in the
positive condition compared to the young.

For the response times, only data from correct responses were included in all
subsequent analyses. Results of the response time data found an overall task effect;
participants were slower to respond in the 2-back in comparison to the l-back condition,
[F (1,57) = 95.37, MSe= 14897.72, eta2 = 0.63]. The age effect was also signiﬁcant, [F
(2,57) = 8.712, MSe = 43327.99, eta2 = 0.23]. Tukey’s honestly signiﬁcant difference
(Tukeys HSD) post hoc analysis demonstrated that the young group responded
signiﬁcantly more rapidly than either of the two older groups, but the two older groups

were not signiﬁcantly diﬁ‘erent from each other.

27

Taken together these results demonstrate that the older adults have more difficulty
than the younger adults do on the n-back task, particularly in the 2-back condition. The
reduction in accuracy and slower response times in the 2-back task is consistent with the
view of reduced inhibitory processing in older adults. Analysis of the recent negative
condition can further lend support for this view. As the next step, comparisons between
the negative condition and the recent negative condition were performed.

Inhibition Comparison: Accuracy

An age (3 age groups) X n-back (1 vs. 2) X condition (recent negative vs.
negative) analysis was performed with the accuracy data. All main effects and interactions
were signiﬁcant except for the three-way interaction (power = 0.46). Results for the
accuracy data showed a signiﬁcant main effect for age [Young M = 92.94, Young-old M
= 87.85, Old-Old M =88.01; F (2,57) = 4.79, MSe =139.47, eta2= 0.14], n-back effect [1-
back M = 97.21, 2-back M = 82.00; F (1,57) = 166.37, MSe =83.47, etaz = 0.75] and n-
back X age interaction [F (2,57) = 2.96, MSe =83.47, eta2 = 0.09] with the two older
groups overall making more errors in the 2-back task compared to the young. The
condition [negative M = 98.99, recent negative M = 80.21; F (1,57) = 176.34, MSe
=119.95, eta2 = 0.76] and condition X age interaction were also signiﬁcant [F (2,57) =
3.81, MSe =119.95, eta2 = 0.12] indicating that the two older groups made more errors in
the recent negative condition compared to the young adults (Figure 2). The n-back X
condition interaction was also signiﬁcant [F (1,57) = 159.63, MSe =74.39, eta2 = 0.74].
Tukey’s post hoc analysis showed that the young adults were more accurate than the two
older groups, but no difference between the two older groups was evident. To explore

further the possible relationship between age and performance within the older age range

28

correlations were computed between age and accuracy scores for the recent negative and
negative conditions just for the two old groups combined. None of the correlations
between age and the recent negative and negative conditions for either n-back condition
approached signiﬁcance.

Given the lack of age differences between the two older groups, these data were
collapsed into a single old group and the analyses were repeated to increase the power of
the tests by increasing sample size. This time, all of the comparisons were signiﬁcant
including the three way interaction (i.e. age X n-back X condition; F (1,58) = 4.45, MSe
=73.52, eta2 = 0.07). In order to more firlly understand this interaction separate two-way
ANOVA’s were conducted for the l-back and 2-back condition separately. In the l-back
task, only the main effect of condition was signiﬁcant [F (1,58) = 15.62, MSe =29.48, eta2
= 0.21] demonstrating that overall accuracy was lower in the recent negative condition in
comparison to the negative condition. Although not signiﬁcant there was a trend for an
overall age effect (ps 0.07, eta2 = .06) with the older adults experiencing more difﬁculty
(i.e. less accuracy) than the young. In comparison, the 2-back found a signiﬁcant age
effect [Young M = 87.3, Old M = 79.34; F (1,58) -—- 9.04, MSe =187.10, etaz = 0.14],
condition effect [negative M = 98.64, recent negative M = 68.00; F (1,58) = 154.58, MSe
=161.96, eta2= 0.73] and condition X age interaction [F (1,58) = 7.19, MSe =161.96, eta2
= 0.11]. In particular, in the 2-back task the older adults experienced relatively more
dimculty in the recent negative condition than the young adults did with no age differences

in accuracy in the negative condition.

29

Inhibition Comparison: Response time

A similar analysis (as computed for the accuracy data) was performed for the
response time data. First an age (3 age groups) X n-back (1 vs. 2) X condition (recent
negative vs. negative) analysis was done. The main eﬂ‘ects of age [Young M = 638.82,
Young-Old M = 791.87, Old-Old M =766.26; F (2,57) = 8.89, MSe =60468.09, eta2 =
0.24], n-back [l-back M = 646.2, 2-back M = 818.43; F (1,57) = 93.82, MSe =18969.1,
eta2 = 0.62] and condition were signiﬁcant [negative M = 665.99, recent negative M =
798.64; F (1,57) = 187.19, MSe =5640.44, eta2 = 0.77]. In addition the n-back X
condition was signiﬁcant [F (1,57) = 6.71, MSe =3179.97, eta2 = 0.11]. No other
interactions were signiﬁcant. Post hoc analyses once again found no differences between
the two older age groups; however these two groups were signiﬁcantly slower than the
young adults. Correlations were computed for age and the recent negative and negative
response times within the older age range to further examine the relationship between age
and performance. None of the correlations with age were signiﬁcant for either n-back task.

The two older age groups once again were collapsed into one group. Results ﬁ'om
the ANOVA revealed similar results. Main effects of age, n-back, and condition were all
signiﬁcant. The n-back X condition interaction was also signiﬁcant [F (1,57) = 4.13, MSe
=3150.29, eta2 = 0.07]. In addition, the age X condition was signiﬁcant with the older
adults experiencing relatively more difﬁculty (i.e. slower response times) in the recent
negative condition than the young [F (1,57) = 3.65, MSe =5572.13, eta2 = 0.06] with

smaller age differences shown in the negative condition.

30

Item Recognition tag

Reliability estimates for the response time data were ﬁrst done by conducting a
split half analysis on the response time data for correct trials for each age group. Alpha
coefficients indicated satisfactory reliability for all three age groups (young, alpha = 0.97;
young-old alpha = .91; old-old, alpha =. 90). Alpha coeﬂicients for the accuracy data were
—0.01 for the young, 0.87 for the young-old and 0.83 for the old-old. The lack of a
relationship for the young adults is due to the very restricted range of accuracy scores in
this measure. At most only 4 errors were made by the young adults (only 2 subjects)
across each half, making it difﬁcult to compute the alpha coefﬁcient. With the young
adults essentially at ceiling on this task, it is felt that the obtained accuracy data is reliable
for this group.

Raw data for the verbal item recognition test is presented in Table 4. The response
time data were ﬁrst trimmed to eliminate possible outliers due to anticipatory responding
or other factors such as distractionsl. Those responses that were more than 3 standard
deviations ﬁom the mean were trimmed. Few responses were eliminated based on this
procedure. Speciﬁcally, 1.5% (SD. = 1.34) of the responses were removed from analyses
for the young group, 0.88% (SD. = 1.10) for the young-old group and 1.96 (SD. = 1.47)
for the old-old group.

First an overall analysis was conducted looking at all four conditions across the
three ages for accuracy and response time data separately. An age X condition ANOVA
was ﬁrst conducted with the accuracy data. Only the condition effect was signiﬁcant [F
(3,171) = 16.49, MSe =46.38, eta2 = 0.22] with lower performance in the recent negative

condition across all age groups. A similar analysis for the response time data found a

31

signiﬁcant condition effect [F (3,171) = 45.52, MSe = 7767.41, eta2 = .44] and age effect
[F (2,57) = 16.21, MSe = 80237.27, eta2 = 0.36]. Post hoc analysis found no diﬂ‘erences
between the two older groups, whereas they were both slower than the younger adults.

Because the stimuli used in the item recognition task were words, level of
vocabulary skill could be a factor in performance on this taskz. Thus, all analyses were
repeated using the Shipley vocabulary score as a covariate. Overall, covarying verbal skill
did not affect the results except where noted below. Repeating the age X condition
analyses, using vocabulary skill as a covariate did not alter the results for either the
accuracy or response time data.
Inhibition Comparison

It was hypothesized that the older adults should experience relatively more
difficulty in the recent negative condition as compared to performance in the negative
condition in the item recognition task, presumably due to age-related inhibitory deﬁcits
resulting in interference effects. Due to the lack of differences between the two older age
groups, these data were ﬁrst collapsed and analyses were conducted looking just at the
comparison between the recent negative and negative conditions. An age X condition
(recent negative vs. negative) comparison was ﬁrst done for the accuracy data. A main
eﬂ‘ect of condition was found with poorer performance seen in the recent negative
condition overall [negative M = 98.13, recent negative M = 90.38; F (1,58) = 52.04, MSe
=30.77, eta2 = 0.47] whereas the main effect for age showed a trend towards signiﬁcance
(p s 0.08, eta2 = .05).

A similar analysis was done for the response time data. In this case a signiﬁcant

main effect for condition [negative M = 782.08, recent negative M = 932.97; F (1,58) =

32

86.35, MSe =7031.29, eta2= 0.60], and age [Young M = 743.74, Old M = 971.3; F (1,58)
= 27.17, MSe =50441.62, eta2 = 0.32] were found as well as a trend for the age X
condition interaction [p s 0.06, F (1,58) = 3.58, MSe =7031.29, eta2 = 0.06, power =
0.46]. Although only approaching signiﬁcance these data suggested that the older adults
experienced more difﬁculty in the recent negative condition as evidenced in slower
response times.

A similar pattern of results was obtained for both the accuracy and response time

data when vocabulary skill was used as a covariate in the analyses.

Eacilitzgion Comparison

Although not the focus of the study, following Jonides et al.’s (1998) design of the
item recognition procedure, possible facilitation effects were examined by comparing
performance in the positive and recent positive condition. In the recent positive condition,
items are presented twice, once in the preceding array and then again immediately aﬁer in
the target array. This repetition of a stimulus may maintain activation of the item, possibly
leading to better performance on these trials. Again the two oldest age groups were ﬁrst
collapsed for the analyses. None of the comparisons in an age X condition ANOVA of the
accuracy data were signiﬁcant. The response time data showed a signiﬁcant effect of
condition (F (1,58) = 4.89, MSe=2220.24, eta2 = 0.08] demonstrating that response times
were faster in the recent positive condition compared to the positive condition [this is
contrary to results reported by Jonides et al. who did not ﬁnd facilitation effects for their
young adults (1998)]. A main effect of age was also found [F (1,58) = 27.89, MSe

=43081.87, eta2 = 0.33] demonstrating that overall the older adults were slower than their

33

younger counterparts. Although in this study facilitation effects were demonstrated by
faster rates of responding, there were no age diﬁ‘erences in this effect
As in the inhibition comparison, no change in results were found when vocabulary

skill was covaried.

Relationship to WCST Measures

The possible role for the prefrontal cortex in inhibitory control of performance on
the working memory measures was explored by examining the relationship between the
WCST and the two working memory measures. Means and standard deviations for the
WCST are shown in Table 5. A one-way ANOVA of the Perseverative response score
revealed a signiﬁcant age difference, [F (2,55) = 3.54, MSe = 332.05]. Tukey’s HSD post
hoc analysis showed that the old-old group had more perseverative responses than the
young group, while there was a trend for the old-old group to be signiﬁcantly different
from the young-old group (p < 0.09). Bivariate correlations were computed between the
number of perseverative responses obtained on the WCST and the accuracy and response
time scores in the recent negative condition for the n-back and the item recognition task.
Perseverative responses were not signiﬁcantly correlated with accuracy or response time
in the l-back task or the item recognition task. In the 2-back condition, although there
was not a signiﬁcant correlation with response time, perseverative responses were
signiﬁcantly correlated with accuracy (r = -0.26) indicating the more difficulty experienced
on the WC ST, as represented by a higher perseverative response score, the greater the
interference effect in the 2-back condition as indicated by lower accuracy. Perseverative

responses were not correlated with accuracy or response time in any other condition (e.g.

34

positive, negative, recent positive) with the exception of a negative correlation with

accuracy in the positive condition in the 2-back task (r = -0.33).

DISCUSSION

Results from the n-back task support the hypothesis that age-related deﬁcits in
inhibition contribute to age differences in working memory performance. Less eﬂicient
inhibitory ﬁrnctioning would lead to greater difficulty supprssing information from
working memory, thus allowing irrelevant material (in this case letters from earlier trials)
to remain in working memory leading to performance decrements due to increased
interference eﬁ‘ects. This interference effect would be expected to be greater in the more
diﬂicult 2-back task. The results showed that overall, older adults had more difﬁculty on
the n-back task than younger adults, as reﬂected both in slowed response time and
decreased accuracy. As predicted by the inhibition theory, these age differences were more
pronounced in the 2-back condition than in the l-back condition.

Although these data are consistent with the inhibition view, the pattern of results
also could be partially accounted for by other theories, such as an age-related decrease in
working memory capacity. However, examination of responses in the recent negative
condition lends additional support for the inhibition hypothesis. If older adults have more
difﬁculty suppressing or deleting the previous n-back item from working memory because
of inefﬁcient inhibitory processes, then older adults would be expected to show greater
interference effects, as reﬂected both in response time and accuracy results, when the
target item actually matched the n +1 back item. This is what was found. Compared to

younger adults, the older adults experienced increased interference in the recent negative

35

condition than in the negative condition, particularly in the 2-back task. In effect, these
results imply that older adults held more information in working memory compared to
younger adults (predicted by the inhibition theory) rather than less information as would
be predicted if working memory capacity decreased with age.

However, results ﬁom the verbal item recognition task did not provide strong
support for the inhibition hypothesis of aging. Overall, performance was lower in the
recent negative condition as compared to the negative condition, but results were only
suggestive of an age difference in this eﬂ‘ect. In particular, results ﬁ’om the inhibition
comparison showed a trend for greater age related diﬂiculties in the recent negative
condition as compared to the negative condition; however, the difference was not
statistically signiﬁcant. In contrast, Jonides et al. (2000) found evidence for age diﬁ‘erences
in the recent negative condition. There are a number of notable differences in procedure
between the current study and the one by Jonides et al. that may account, in part, for this
difference. The current study used words as stimuli, whereas Jonides et al. used single
letters. The use of words may have provided enough context for the older adults to
facilitate recognition, overcoming any possible decrements that might have occurred due
to decreased inhibition. Unlike letters which, being only 26 in number, required ﬁ'equent
repetition (and most likely increasing interference effects), no word was repeated in the
study arrays (except for items in the recent positive and recent negative conditions),
making each stimulus relatively unique from each other and potentially easier to
discriminate. Earlier studies have shown that age differences in inhibition can be

attenuated if not eliminated by slight changes in the task (such as location of distracting

36

stimuli) that allow for easier discrimination between stimuli, reducing the role of inhibitory
processes for successful performance (Carlson et al., 1995; Connelly & Hasher, 1993).

In addition, contrary to earlier ﬁndings in young individuals (Jonides et al. 1998),
there was a facilitation eﬁ‘ect in both the young and old age groups, with no age
diﬁ‘erences in the size of the eﬂ'ect.

An alternative explanation for the lack of age differences in the current study
pertains to the power of the current study to detect an age effect. In the Jonides et al.
study (2000), because the focus was on the correlates of performance and neuroimaging
results, participants in their study were extensively trained to ensure a high level of
accuracy across all conditions and to reduce error variance prior to conducting the study,
making any age diﬁ‘erences more detectable by statistical analysis. The subjects ﬁ'om the
current study did not have the beneﬁt of prior training. Inspection of the data reveals
patterns consistent with Jonides et al.’s study, with the older age groups experiencing
greater amounts of interference as reﬂected in longer response times and lower accuracy
(in fact the present study demonstrated numerically larger interference effects for the older
age groups), yet the effects here were not statistically signiﬁcant.

Limited indirect evidence was also provided for the possible role of the prefrontal
cortex to inhibitory functioning. Perseverative responses from the WCST, a measure that
is often argued to be associated with dorsal lateral prefrontal functioning, was correlated
with performance on the 2-back task for the inhibition condition (i.e. recent negative).
This relationship was not evidenced with the l-back or the VRWM. The l-back is an
easier task, and it can be argued that inhibition may not be as crucial a process to this task

as the 2-back, when more material is dealt with in working memory. In contrast, although

37

a signiﬁcant relationship was expected between the WCST and the VRWM interference
scores, since the expected interference effects associated with age were not clearly
demonstrated on the VRWM task, it is not surprising that there is a lack of relationship
between these two measures making interpretation of this null effect difﬁcult.

One ﬁnal point of the ﬁrst study is the lack of expected age differences between
the two older groups. This result is contrary to ﬁndings reported by a number of
researchers, including those by the author (Persad et al., in press). But, other studies have
failed to ﬁnd age differences within the older age groups (Keys & White, 2000; Speiler et
al., 1996). At this time, it is unclear why there is this discrepancy in age differences. One
factor may be differences in task diﬂiculty, although in the current study the n-back task
could be argued to be a much more difﬁcult task than the item recognition task. Yet
neither task showed the expected age differences between the older age groups. If
continued age declines within the older population are found consistently with only a
subset of cognitive tasks, then this pattern of results would suggest more focal changes
that occur within certain systems as opposed to a global process that is believed to

underlie all age-related cognitive changes.

STUDY 2
Although one can argue for the role of inhibition in the VRWM and n-back tasks
by demonstrating that predicted results based on the theory do occur, it is also important
to demonstrate that other theories cannot adequately account for the results. One such
theory is the processing speed theory, championed by Salthouse and his colleagues (see

Salthouse 1996 for a review). According to the processing speed theory, as one ages,

38

there is an accompanying generalized slowing of cognitive processes, leading to
performance decrements on a variety of tasks. The eﬁ‘ects of slowing result from two
proposed mechanisms. The ﬁrst is the limited time mechanism. Important cognitive
processes are assumed to operate more slowly with increased age, which can reduce
performance on tasks requiring rapid on—line processing, as in reaction time measures or
language comprehension. Slowed initial processes also will cause difﬁculties in later stages
of processing, not just because of time constraints but because of the limited or incomplete
information that arises from the earlier stages. The second mechanism, related to the ﬁrst,
is called the simultaneity mechanism. Higher level, more complex cognitive processes are
assumed to integrate and evaluate products of earlier processes. For this stage to be
effective, the information must be simultaneously available. If, however, information is not
readily available at the same time due to slowed processing effects, the ﬁnal products of
the higher level processes may be inadequate.

Like the inhibition theory, the processing speed theory has garnered considerable
support. Many of the studies that have evaluated the role of processing speed have used
different types of cognitive tasks along with measures of perceptual speed. Performance
on the perceptual speed measures has been taken as an indicator of processing speed. The
general analytical procedure used in these studies is hierarchical regression analysis, which
shows that the perceptual speed measures mediate much of the relationship between age
and the cognitive task in question. Measures of processing speed have been found to be
signiﬁcant mediators of a variety of cognitive tasks, including working memory (Salthouse
& Meinz, 1995), arithmetic (Salthouse & Coon, 1994), paired associate learning

(Salthouse, 1993) and ﬁee recall (Salthouse, 1993; 1994; Salthouse & Coon, 1993). One

39

study in particular has found that measures of perceptual speed account for as much, and
in some cases more, of the age-related variance on two working memory span measures
than a measure of inhibition (Salthouse & Meinz, 1995).

It is important to note, however, that two recent studies have called into question
the validity of the perceptual speed measures as relatively pure measures of processing
speed (Lustig, Tonev & Hasher, 2000; Tonev, Lustig & Hasher, 2000). It has been
demonstrated that age differences in performance on the variety of tasks generally used as
measures of processing speed (letter and number comparison tasks, symbol digit
substitution, arrow task) are in part a result of interference eﬂ‘ects that occur as a result of
the way the tests are administered. Typically items on perceptual speed measures are
presented simultaneously, and age differences are consistently found. However, when
items are presented one at a time, age diﬁ‘erences in performance are much smaller,
suggesting that older adults are more vulnerable to visual distraction, as reﬂected by
relatively poorer performance. This susceptibility to visual distraction in older adults is
consistent with inhibitory declines associated with age.

More recently the results from a number of studies have suggested that both
generalized slowing processes and declines in inhibitory functioning must be taken into
consideration to provide a fuller account of age-related deﬁcits in cognition. As described
above, See and Ryan (1995) found that both speed and inhibition were signiﬁcant
independent sources of the age-related variance on their battery of language tasks. Persad
et al. (in press) showed that both measures of reading speed and inhibition accounted for a
signiﬁcant proportion of the age-related variance on a complex attention task and a verbal

memory task. Nettelbeck and Rabbitt (1992) reported that measures of processing speed

40

predicted only some of the age-related variance on a variety of cognitive tasks (including
tests of recognition memory, list learning and free recall of 30 nouns), suggesting that
slowed processing rate in older adults does not fully account for age-decrements in
performance on these tasks. After controlling for level of complexity across tasks of
processing speed and executive functioning, Keys and White (2000) also found that
although processing speed accounted for a signiﬁcant amount of the variance in the
executive tasks, age was still a signiﬁcant contributor.

West and Baylis (1998) have also investigated the relative contributions of slowing
and inhibition to performance on the Stroop task. For the standard administration of the
Stroop task, measures of perceptual speed accounted for a signiﬁcant proportion of the
age-related variance in the interference score. This, however, still left a signiﬁcant
proportion of unexplained variance. They then administered a variation of the Stroop
color-word task. Some words were presented in their respective color (congruent trial)
whereas other words were presented in the usual way (i.e. a different color; incongruent
trial). Incongruent and congruent trials were mixed during presentations. However, two
conditions were administered that varied the proportion of congruent/incongruent trials.
Half of the participants received a test series with more congruent than incongruent trials,
whereas the remaining participants received more incongruent than congruent trials. The
hypothesis was that in the mainly incongruent condition, the participants had to use a color
naming strategy while inhibiting automatic word reading for successﬁrl performance. In
the mainly congruent condition, by contrast, participants could choose either a color
naming or word reading strategy and still do relatively well. If, according to the inhibition

theory, inhibitory processes in older adults are deﬁcient, it was expected that this age

41

group would show more interference in the mainly incongruent condition. In contrast, no
age differences were expected in the congruent condition, because the task allowed for the
use of a word reading strategy, which the older adults could use to achieve the correct
response, Results showed that older adults exhibited more interference in the incongruent
as opposed to the congruent condition when compared to younger adults. These age
differences were maintained even after performance responding rate was covaried out.
From the results, the authors concluded that the contributions of processing speed and
inhibition to successful performance varied according to the requirements of the task.

These results highlight a need to investigate the relationship of both processing
speed and inhibitory ﬁrnctioning to age-related cognitive deﬁcits in working memory. In a
review of the literature on processing speed, Salthouse (1996) himself has stated that the
postulation of a generalized slowing mechanism does not preclude the possibility of
additional speciﬁc processes that are also affected with age.

The aim of study two was to investigate the potential contribution of processing
speed to performance in the inhibition condition of the VRWM task. The inhibition theory
states that the inability to suppress irrelevant material in working memory can lead to
interference from this distracting information. The goal of the study was to change the
task demands of the VRWM by changing the type of stimuli used,'with the intention of
increasing the amount of interference assumed to occur within the task and thus the need
for inhibitory processes. Age differences in performance were expected to increase as the
amount of interference increased. In contrast, the processing speed theory would predict
equal age differences regardless of the type of material to be processed, so long as the task

parameters themselves did not change.

42

In an attempt to increase interference eﬁ‘ects, stimuli for the item recognition task
were chosen based on category membership. Words from the same category are thought
to have stronger associative bonds than unrelated words, which in theory may be more
diﬁcult to suppress, therefore causing greater interference effects. It has been
demonstrated in a variety of tasks that words ﬁom the same category can cause greater
proactive interference eﬁ‘ects than words ﬁ'om different categories. For instance, the
classic proactive interference in short term memory paradigm (Wickens, 1972; Wixted &
Rohrer, 1993) has demonstrated a decrease in free recall performance across trials when
the stimuli used are words ﬁ'om the same category, presumably due to a build-up in
proactive interference. This is followed by a return to the original level of performance
when the category is changed.

As discussed earlier, ineﬁicient inhibitory processes will result in increased
interference effects as reﬂected in lowered accuracy and slower response times. Given that
the inhibitory mechanisms are deﬁcient with age, irrelevant material will remain activated
in working memory. The use of a categorized word list is expected to increase the amount
of interference that is experienced because of this sustained activation across a pair of
trials, involving the same category in comparison to an unrelated list. Thus, age differences
in the recent negative condition should be greater in the categorized list presentation.

Although the inhibition theory does predict a different pattern of results with the
two word list conditions, the processing speed theory would not necessarily predict a
difference in the proportion of age-related variance accounted for by perceptual speed.
Because the task is identical in both list conditions, with the exception of the stimuli used,

the processing speed theory would assume that the underlying processes are the same in

43

both conditions. The use of different stimuli should not affect these processes; therefore
the processing speed theory would not predict age-related diﬁ‘erences in performance
across the two word lists.

To examine more speciﬁcally the relationship between processing speed and
inhibition with age, a hierarchical regression analysis was used (Hertzog, 1996; Salthouse
1991; 1994). If speed mediates the declines in working memory that accompany age, then
controlling for speed, the proportion of age-related variance in working memory
performance should be substantially reduced. Processing speed was assessed using two
perceptual speed measures, often used in studies investigating cognitive slowing and age
(Salthouse, 1996). These were the pattern comparison and the letter comparison tasks.
Following the evidence reviewed earlier, it was expected that both processing speed and
inhibitory functioning would affect performance in the older age group. It was expected
that the perceptual speed measures may mediate some of the age-related variance on the
item recognition task, but not all because of the hypothesized role of inhibition in this task.
The processing speed theory would predict that the proportion of age-related variance
accounted for by the perceptual speed measures would be roughly equivalent for the two
list types. In contrast, the inhibition view would predict that inhibitory processes
differentially come into play for each word list condition, with inhibition being a more
important factor as the amount of interference increases. In particular, although speed is
expected to account for a signiﬁcant proportion of the age-related variance in item-
recognition performance, there would still be a signiﬁcant amount of variance still
unaccounted for, which it is argued is a reﬂection of inhibitory processes. However, there

will be a weaker, non-signiﬁcant relationship between the speed measures and

44

performance in the increased interference condition, presumably because it is the added
deﬁcit of inhibition that accounts for the additional age-related decline in performance.

In addition, although secondary to the aim of the study, not only was the category
manipulation expected to increase interference effects, it was also thought that the use of
categorized lists would increase facilitation effects across a pair of trials. The stronger
associative connections of words from the same category would more likely maintain
activation of the item in the recent positive condition, thus leading to greater facilitatory
effects with the categorized list. However, it is uncertain whether there should be age
differences in the amount of facilitation evidenced in the recent positive condition as
compared to the positive condition. Therefore, it was predicted that both the recent
negative and recent positive conditions would show increased effects overall, while
performance in the positive and negative conditions should be better in the categorized

list.

METHOD

Participants

One hundred adults divided into ﬁve equal age groups (18-29; 30-44; 45-59; 60-
74; 75 years and older) were tested. The young adults were undergraduate students who
received course credit for their participation. The remaining adults were recruited from the
local community and were paid ten dollars for their participation. Exclusion criteria were
the same as in Study 1. Two participants were excluded based on a diagnosis of bipolar
disorder and were taking Lithium (one adult from the 18-29 range and one from the 46-59

range). Two additional adults were tested to maintain equal sample sizes.

45

Descriptive data are presented in Table 6. One-way analyses revealed signiﬁcant
diﬁ‘erences in level of education (F (4,95) = 2.73, MSe = 4.19) and Shipley scores (F
(4,95) = 16.05, MSe = 13.38) between the ﬁve age groups. Post hoc comparisons for
education showed that the 60-74 year olds had obtained a signiﬁcantly higher level of
education than the other four age groups. No other differences were signiﬁcant. Post hoc
analysis for the Shipley scores revealed that the two youngest groups had signiﬁcantly
lower vocabulary scores than the three older groups, and were signiﬁcantly different from
each other.

Presence of depressive symptoms was assessed as in Study 1 using the Beck
Depression Inventory for the three youngest groups and the Geriatric Depression Scale for
the two older groups. Mean scores on the self-report questionnaires were 6.1 (SD. =
6.49) for the 18-29 year olds, 4.6 (SD. = 3.3) for the 30-44 year olds, 5.55 (SD. = 5.38)
for the 45-59 year olds, 4.35 (SD. = 3.3) for the 60-74 year olds and 4.05 (SD. = 1.7) for
the 75 and older group. Using a cut-off of 10, 4 individuals from the 18-29 group, 2 from
the 30-44 and 2 from the 45-59 group scored above the cut-off value. Analyses were
conducted both with and without these individuals with similar results. As a result these
data were retained in the following analyses to maintain equal sample sizes and greater

power.

46

VerbaLl Item Recognition test
Materials

Two lists were developed for this study. The structure of the lists was the same as
in Study 1, with 20 trials in each of the four conditions (i.e. positive, facilitation, negative
and recent negative). The ﬁrst list consisted of nouns of varying length again chosen from
the Kuchera and Francis (1967) ﬁequency norms with a range of 4 to 450, referred to as
the unrelated list. Although the basic list of items was essentially the same as that of Study
1, some of the words were replaced with longer words in an attempt to equate for word
length between list conditions. A second list of items was developed that used nouns
chosen from the Battig and Montague (1969) category norms. Words were chosen from
40 separate categories with the constraint that there were enough category instances that
were less or equal to 8 letters in length to fulﬁll the necessary conditions. The most
representative members of each category were chosen. Each array of words consisted of
four items from the same category; on each trial the array had words from a different
category unless it was in a facilitation or recent negative condition, for which the category
remained the same. Four separate lists for the unrelated and categorized lists were

designed to counterbalance the position of each array in each of the four conditions.

Procedure

The procedure was similar to Study 1. Each participant completed both the
unrelated and the categorized conditions. Administration of the two list types was
counterbalanced. The perceptual speed measures were completed between the

administration of the two list types. Shipley scores were also obtained for each individual.

47

Perceptual Speed Measures (Salthouse. 1994;

Two paper and pencil measures of processing speed were administered: the letter
comparison and the pattern comparison task. The letter comparison task consisted of pairs
of three, six, or nine letters presented on a page. The participant was asked to respond as
quickly as possible with an S if the pairs were the same or D if they were different. Thirty
seconds was allotted for this task, and the score was computed as the number correct
minus the incorrect responses. The procedure of the pattern comparison task was
equivalent to the letter comparison task except that the stimuli consisted of pairs of
patterns with three, six, or nine line segments. Two trials of each the letter and pattern
comparison tasks were administered and an average score was computed separately for

both speed measures.

RESULTS

The raw data for the response time and accuracy score across the ﬁve age groups
are shown in Tables 7 and 8 for the unrelated and categorized lists respectively. The
response time data were ﬁrst trimmed to remove outlying responses that were more than 3
standard deviations from the mean. Just as in Study 1, the amount of data lost to this
procedure was minimal. For the unrelated condition, the percent of data lost across the
ﬁve age groups (beginning with the youngest age range) were 0.96, 1.05. 1.50, 1.3, and
1.16. A one-way ANOVA failed to ﬁnd any differences in lost data between the age

groups. A similar pattern of results was obtained for the categorized condition, with again

48

no age differences noted. Percent data removed for the ﬁve age ranges, starting with the
youngest in order, were 0.84, 1.80, 1.76, 1.62, and 1.88.

An age (5) X list type (2) X condition (4) was conducted ﬁrst for the accuracy
data. A main effect for condition (F(3,285) = 23.42, MSe = 49.20, eta2 = 0.20) and age
(F(4,95) = 3.91, MSe = 130.19, eta2 = 0.14) were found. In addition, the condition X age
interaction was signiﬁcant (F(12,285) = 3.2, MSe = 49.20, eta2 = 0.12) as was list X
condition (F(4,285) = 1.36, MSe = 29.08, eta2 = 0.04).

The parallel analysis for the response time data showed a main effect for condition
(F(3,285) = 64.60, MSe = 16087.31, eta2 = 0.41), and age (F(4,95) = 10.38, MSe =
188521.85, eta2 = 0.30) as well as a list by condition interaction (F(43285) = 3.85, M86 =
4626.47, eta2 = 0.04). Since the comparisons of interest were between the inhibition and
facilitation conditions, separate analyses were done for each, with a goal to understanding
the interactions.

Inhibition Comparison

Age X list (categorized vs. unrelated) X condition (recent negative vs. negative)
analyses were conducted to test for the expected increased interference in the categorized
condition. For the accuracy data neither the main effect of list type nor any of the
interactions with list type were signiﬁcant. Contrary to expectations, the recent negative
condition did not show larger age effects for the categorized vs. the unrelated lists. A main
effect of condition (F(1,95) = 57.47, MSe = 47.23, eta2 = 0.38) and age were found
(F(4,95) = 4.82, MSe = 112.39, eta2 = 0.17) as well as a condition x age (F(4,95) = 6.97,
MSe = 47.23, eta2 = 0.23) eﬂ‘ect shown in Figure 3. In particular for the ﬁrst four age

groups little interference effects were noted as evidenced by a lack of difference between

49

accuracy in the recent negative and negative conditions. In contrast, the oldest age group
showed a decrease in accuracy in the recent negative condition compared to the negative
condition, suggesting increased interference eﬁ‘ects restricted to the oldest age group.

The response time data showed a main effect of age (F(4,95) = 9.06, MSe =
135501.07, eta2 = 0.27) and condition (F(1,95) = 170.33, MSe = 13431.04, eta2 = 0.64).
In addition the list X condition (F(1,95) = 5.08, MSe = 4522.63, eta2 = 0.05) and the 3
way interaction were signiﬁcant (F(4,95) = 2.54, MSe = 4522.63, eta2 = 0.10). The 3-way
interaction is plotted in Figure 4. The signiﬁcant interaction seems to reﬂect some slight
reversals across age in the relative response times in the categorized vs. unrelated recent
negative conditions. Although statistically signiﬁcant, this interaction was not predicted by
the study and the pattern does not appear to be theoretically signiﬁcant. Of importance to
this study, there was a trend for the age X condition interaction; however it did not reach
signiﬁcance (p s .13, eta2 = 0.07). When vocabulary skill was used as a covariate in the
analyses, results were similar except that there was only a trend for the 3-way interaction
to be signiﬁcant.

Taken together, results from the accuracy and response time data are consistent
with the idea that compared to younger adults the older adults have more difﬁculty
suppressing information ﬁ'om earlier trials, leading to increased interference in the recent
negative condition as demonstrated by reduced accuracy and slower response times.
However, this was only a trend in the response time data. To increase the power of ﬁnding
possible age differences, the data from this study was joined with the data from Study 1.
The data from the youngest age group and the two oldest age groups were collapsed for ~

the unrelated list condition. Participants were chosen from the same age range in both

50

studies allowing for a combination of the two data sets. This procedure allowed for an
increase in the sample size to 40 young and 80 old adults. An age (young vs. old) X
condition (recent negative vs. negative) X study (study one vs. study 2) analysis was
performed. The accuracy data showed signiﬁcant main effects of age and condition but no
differences between Study 1 and Study 2. In addition, the age X condition interaction
reached signiﬁcance (F (1,116) = 8.86, MSe = 50.84, eta2 = 0.07), with the older adults’
performance in the recent negative condition declining relative to the young adults (Figure
5). Response time analyses demonstrated the same pattern of results. Both the main effects
of age and condition were signiﬁcant as was the age X condition interaction (Figure 6; F
(1,116) = 10.83, MSe = 7957.96, eta2 = 0.09). The older adults had relatively more
difﬁculty in the recent negative condition than the young as evidenced in their slower
response times than in the negative condition. A similar pattern of results was found using
vocabulary as a covariate in the analyses.
ﬂcilitation Comparison

An age X list type by condition (recent positive vs. positive) analysis was
conducted ﬁrst for the accuracy data. Results showed an effect of list type (F(1,95) =
5.64, MSe = 29.91, eta2 = 0.06) and condition (F(1,95) = 15.11, MSe = 25.94, eta2 =
0.14). In addition the list X condition comparison was signiﬁcant (F (1,95) = 4.19, MSe =
28.90, eta2 = 0.04) showing that overall facilitation effects were demonstrated in the
categorized list but not the unrelated list. The age X list interaction approached
signiﬁcance (pg .06, eta2 = 0.08, power = 0.63) showing a trend for the two oldest groups
to show lowered performance in the categorized list condition, unlike their younger

counterparts who did not show a change in accuracy.

51

A similar analysis was conducted with the response time data. A main effect of list
(F(1,95) = 4.41, MSe = 12908.51, eta2 = 0.04) was found with overall slower response
times demonstrated in the categorized list. There was also a main effect of condition
(F(1,95) = 16.76, MSe = 3907.07, eta2 = 0.15) with faster response times evidenced in the
recent positive condition. In addition response times increased with age (18-29 yr. olds M
= 685.14; 30-44 yr. olds M = 778.2; 45-59 yr. olds M =744.92; 60-74 yr. olds M =
889.12; 75+ yr. olds M = 924.77; (F(4,95) = 9.56, MSe = 83944.64, eta2 = 0.29).
Processing Speed

Raw data for the processing speed measures are shown in Table 9. One way
AN OVAs were performed for the letter and pattern comparison speed measure separately.
In both cases, there was a signiﬁcant age effect, with the older adults obtaining lower
scores on these measures (Letter Comparison F(4,95) = 12.56, MSe = 5.23; Pattern
Comparison F(4,95) = 24.20, MSe = 8.13) To evaluate the contribution of processing
speed to item recognition performance, the aim was to run hierarchical regression
analyses, by ﬁrst looking at the amount of age-related variance and then by evaluating the
proportion of this variance that remained once the effects of processing speed were
partialled out. However, the relative role of processing speed in the interference effects of
the item recognition test cannot be ﬁrlly tested as hypothesized because the manipulation
designed to increase interference effects failed. Nonetheless, there were signiﬁcant
correlations between age and the accuracy and response time data in the recent negative
condition, so the relative contribution of processing speed to this relationship was
examined. First, because no diﬂ‘erences were found between the categorized and unrelated

lists, data were collapsed across the two list conditions. As a measure of interference,

52

difference scores were calculated by subtracting the negative condition from the recent
negative condition for both the accuracy and response time data. This difference was
thought to reﬂect the increased interference effects in the recent negative condition. A
signiﬁcant relationship was found between age and the accuracy difference score (r =
0.36) as well as age and the response time difference score (r = 0.23).

To collapse data from the letter and pattern comparison measures, z-scores were
ﬁrst computed for each measure and then averaged together to arrive at one composite
speed measure which was then used in the regression analyses.

A hierarchical regression analysis was performed on accuracy and response time
separately. Vocabulary skill was used as a covariate in the analyses. Because of the strong
correlation between age and vocabulary, age was ﬁrst regressed onto vocabulary scores
and the unstandardized residuals were saved. These unstandardized residuals were then
covaried out by entering them into the equation ﬁrst. Table 10 demonstrates the order of
entry into the regression model. Age accounted for a signiﬁcant proportion of the variance
in accuracy performance in the recent negative condition. However, when processing
speed was entered into the equation, age no longer accounted for a signiﬁcant proportion
of the variance for accuracy performance. Similar results were found for the response time

data.

53

DISCUSSION

The results ﬁ'om this study did not support the expectation of overall increased
interference effects in the categorized list condition as compared to the unrelated list. Nor
did they show a differential increase in interference effects associated with age, as
predicted by the inhibition theory of aging. In an attempt to understand these contrary
ﬁndings, we need to look more closely at the manipulation used in this study so that we
can say whether it increased interference. The categorized words were presented in
blocks, i.e. each array of fours words was from one category with items in the next array
from a different category unless it was the facilitation or interference condition, where it
was necessary to continue with the same category. On ﬁrrther consideration, this use of
the blocking procedure may actually parallel more closely a release from proactive
interference experiment, one in which less, rather than more, interference would be
expected. It has been shown that although one sees declining recall performance over trials
when words from the same category are used, recall rates rebound when the next
presentation uses stimuli from another category; i.e. release ﬁ'om proactive interference
(Wickens, 1972; Wixted & Rohrer, 1993). Thus, in this study although interference may
have built up across a few trials due to inefﬁcient inhibitory mechanisms, each time the
category shifted it can be argued that this manipulation most likely “released” any buildup
of proactive interference so that an increase in interference effects in the categorized list
condition would not be expected. Therefore, unfortunately, the very manipulation intended
to increase interference effects may have worked contrary to the original plan and
produced the reverse. Given this line of reasoning, these results imply that there may have

been an overall increase in proactive interference in the item recognition task across all of

54

the trials that may have been contributing to age diﬂ‘erences, rather than just an increase
within a given pair of trials in the recent negative condition.

Nonetheless, consistent with results of Study 1, the oldest age group demonstrated
increased interference eﬂ’ects on the item recognition task (regardless of list type)
compared to the other age groups, presumably due to decreased inhibitory functioning.
Even stronger support for the view of age related declines in inhibition were found when
the unrelated list condition data from participants chosen from the same age brackets were
combined with data from the ﬁrst study. In particular, the hypothesized age differences in
performance in the interference condition were evident. Speciﬁcally, the older individuals
showed a differential decline in performance in the recent negative as opposed to the
negative condition. This pattern of results suggests that material ﬁom earlier trials remains
activated in working memory longer in older adults, due to ineffrcient inhibitory
mechanisms that would otherwise suppress or delete no longer relevant information so
that it does not interfere with subsequent processing. As a result of the maintenance of
information in working memory from earlier trials, when the probe item actually matches
one from an earlier list, it becomes relatively more difﬁcult for older adults to respond
appropriately because a match is found in working memory, albeit not the correct match,
leading older adults to make more errors as evidenced in decreased accuracy and slower
response times.

The category manipulation, on the other hand, did work in increasing overall
facilitation effects. Seeing a word presented on two successive trials produced better
performance when presented within a categorized structure. The strong associative links

between members of a common category most likely aided in maintaining activation of the

55

word across trials. In addition, when trials were presented in succession with stimuli from
the same category, this method of blocking may have alerted participants more to
consistencies between trials, and highlighting the fact that when a word was repeated
across trials, it would also be the probe target.

One further point, it is unclear why only the very oldest age group would
demonstrate increased interference effects whereas the 60-74 year olds failed to show this
effect. Results of Study 1 failed to ﬁnd differences between a group of young-old and old-
old individuals in either the n-back or the item recognition task. It is unclear why only the
oldest group, similar in age to the old-old in Study 1, would show a interference effects,
whereas the next youngest group, chosen from the same age range as the young-old of
Study 1, would show a more similar pattern of performance in the interference comparison
to their younger counterparts. It may be that there were diﬂ‘erences in sample
characteristics or individual differences between the groups of participants, even though
the average age and education level and vocabulary skill of each group were very similar.
Unfortunately, the Mini Mental State Examination was not administered to the older
subjects in study 2, so it is possible that some of the oldest subjects were mildly
cognitively impaired. If so, that would skew the results in the direction of age diﬂ‘erences
when in reality no differences are present. However, it is still important to stress that when
the data were combined across studies, thus increasing the power of the analyses, a

reliable age difference was evident in the inhibition condition.

56

W

The aim of the current studies was to evaluate the role of inhibitory processes in
working memory performance diﬁ‘erences associated with advancing age. Given the view
that inhibitory mechanisms become less eﬂicient with age, decreased performance on two
measures of working memory, the n-back and the item recognition tasks, was expected.
Overall, the results supported the premise that inefficient inhibitory mechanisms account,
at least in part, for age-related decrements on working memory tasks. This was most
clearly seen in results with the n-back task. As task complexity increased in the 2-back
condition, requiring the processing of relatively more information than in the l-back
condition, older adults showed greater dimculty as evidenced in decreased accuracy and
slower response times.

In addition, the examination of performance on speciﬁc foil trials in the n—back task
provided stronger support for the inhibition viewpoint. The inclusion of the critical 3-back
condition provided additional evidence for decreased inhibitory functioning associated
with age. As originally hypothesized, if older adults have more difﬁculty deleting
information from working memory, this could lead to increased interference effects due to
an increase in “clutter”. In the n-back task, with the continual need to update the contents
of working memory as each new letter is presented, it was hypothesized that for the older
adults, letters presented earlier in the sequence would remain activated in working
memory, thus making it more difﬁcult for the older adults to successﬁrlly complete the
task because of a resulting increase in interference effects. To test this hypothesis on some
trials the probe item actually matched an item that was in the 3-back position rather than

the 2-back item. Although the appropriate response would be “no” (i.e. it does not match

57

the letter in the 2-back position), it was hypothesized that older adults would experience
more diﬂiculty on these trials presumably because the 3-back item was still activated in
working memory. Results from Study 1 supported this hypothesis. Although overall, all
participants had more difﬁculty when the target letter actually matched the 3-back item,
the difﬁculty for older adults was greater, based on response time and accuracy measures,
than their younger counterparts.

Although the age effects were not signiﬁcant in the item recognition task in Study
1, there was a trend toward increased interference eﬂ‘ects with advancing age.
Furthermore, when comparable data from both Study 1 and 2 were combined, increasing
the overall sample size and thus the power of the analyses, the age X condition
interactions were signiﬁcant, indicating that compared to the younger adults older adults
showed larger interference effects in the recent negative condition than the negative
condition. The larger age effects in interference are consistent with the view that inhibitory
functioning decreases with age.

Taken together, these results suggest that in older adults inhibitory ﬁrnctioning
becomes less efﬁcient leading to decrements in performance in working memory tasks.
This inefﬁciency allows irrelevant information to remain activated in working memory,
leading to decreased performance due to interference that is caused by this extra
information.

Role of the PrefroLal Cortex and Inhjbition

 

Results from Study 1 provided indirect evidence for the role of the prefrontal
cortex in inhibition. A signiﬁcant relationship was found between performance on the

WCST and the n-back task. In particular, this relationship was only signiﬁcant for the 2-

58

back version, the one in which clear age differences were found. This result supports other
studies that have reported links between ﬁontal lobe functioning and inhibitory deﬁcits
(Chao & Knight, 1997; Cronin-Golomb, 1990; Stuss & Benson, 1984). However, it is
important to note that WCST is a complex neuropsychological measure, one that cannot
be easily classiﬁed as a speciﬁc measure of any one cognitive process, such as inhibition.
In fact the primary role of the prefrontal cortex in performance on the WCST has been
questioned (Anderson, Damasio, Jones, & Tranel, 1991) with additional cortical and/or
subcortical areas implicated. Additional studies are necessary to more fully examine the
possible link between speciﬁc neuroanatomical areas and inhibitory ﬁrnctioning. The use of
neuroimaging procedures could provide useful information in this regard.

There is strong evidence in young adults that the prefrontal cortex is important for
working memory and, speciﬁcally for the n-back task. Using positron emission
tomography (PET) (Schumacher, Lauber, Awh, Jonides, Smith & Koeppe, 1996), in
comparison to a control condition (in this case a 0-back condition), the verbal n-back task
showed increased activation in Broca’s area, the left posterior parietal lobule and to a
lesser degree the right posterior parietal and supramarginal area (Brodmann’s area 40), as
well as the left dorsal lateral prefrontal cortex (DLPFC, Brodmann’s area 9, 10, 44, 45,
46). The same pattern was evidenced for a spatial version of the n-back task in the
homologous regions in the right hemisphere. These results replicate earlier ﬁndings that
have implicated these brain areas in working memory. This verbal working memory system
appeared to be amodal, in that these same areas were activated whether the stimuli were
presented visually or auditorily. To understand more fully the role of the different brain

regions in working memory, Awh and colleagues (Awh, Jonides, Smith, Schumacher,

59

Koeppe, & Katz, 1996) used a 2-back task to study the relative contributions of these
areas to storage and rehearsal processes. Results suggested that Broca’s area and
surrounding speech areas were important in rehearsal, whereas the posterior parietal and
DLPFC appeared to be involved in storage of information in working memory. Using a
time series design, Cohen and colleagues (Cohen, Perlsteirr, Braver, Noll, Jonides, &
Smith, 1997) also demonstrated putatively different roles for the various cortical areas that
are activated in the n-back task. Early in the task, Broca’s area showed increased
activation as the memory demands increased presumably due to rehearsal processes, but
this activation actually decreased over the time delay. The posterior parietal areas showed
increasing activation with increased memory load that was sustained across the delay,
supporting its role in the storage of information. Interestingly the DLPFC did not show
increased activation for the 0 and l-back conditions but did show a marked increase
during the 2 and 3-back trials that was sustained over the delay. These results suggest
that the DLPFC may play a role in more complex working memory tasks.

Parametric studies also have been used to investigate the relationship between
neuroanatomical regions and working memory (as opposed to the subtraction method that
was used in the studies described above). The following studies dealt only with the verbal
n-back procedure. Jonides et al. (1998), using four lag conditions (0,1,2 and 3), found
activations in the same areas as described earlier that increased linearly with increasing
memory load. Braver et al. (Braver, Cohen, Jonides, Smith, & Noll, 1997) replicated these
ﬁndings using ﬂvIRI instead of PET.

The use of neuroimaging data to examine the age deﬁcits observed on the n-back

task could prove fruitful. Data from neuroimaging research could provide additional

60

evidence in support of the role of the preﬁ'ontal cortex to inhibition. In particular it would
be interesting to examine the amount of activation that occurs in the prefrontal cortex
when directly comparing the recent negative and negative conditions in this task. If the
PFC subserves inhibitory processes, then it would be expected that one would see a
diﬁerence in activation rates in this area in the recent negative condition as compared to
the negative condition. Furthermore, given that the older adults in this study showed
ineﬂicient inhibitory functioning compared to young adults on the n-back task, one
possible outcome would be that the results of the neuroimaging data would parallel the
behavioral outcome, demonstrating a relative decrease (or increase) in activation in the
PFC for the inhibition (i.e. recent negative) condition in comparison to the young adults.

Jonides et al (2000; Jonides et al., 1998) did ﬁnd results that support this
hypothesis using the item recognition task. In particular not only did they ﬁnd that the
older adults had more difﬁculty behaviorally in the recent negative condition compared to
young individuals, they also found that there was no reliable activation of the PFC in the
older adults during these trials, unlike the increase in PFC activation seen with the young
individuals.
InhibitioL and Processing Speed

Unfortunately, because Study 2 was not successful in increasing interference, the
relative roles of processing speed and inhibition could not be examined across different
tasks that presumably relied differentially on inhibitory mechanisms for successful
performance. Nonetheless, processing speed measures did account for almost all of the
age-related variance demonstrated in performance in the item recognition task. These

results suggest that at least in the item recognition task, most of the age effects seen in

61

study 2 were mediated by processing speed. However, it is important to note two points.
First, the age related declines in performance in the inhibition condition in the item
recognition task were relatively weak. Clear age declines in performance were not evident
until the sample size was greatly increased by collapsing data across the two studies. As a
result, it is unclear at this stage if the processing speed results reﬁrte the inhibition
hypothesis of aging, or if the declines in performance due to inhibition deﬁcits are
relatively small, thus making it difﬁcult to partition the contribution of different processes
to this task.

Second, recent research has called into question the validity of the perceptual
speed measures as good measures of information processing speed (Lustig et al., 2000;
Tonev et al., 2000). As suggested by these researchers, if interference effects account, in
part, for the age differences in the perceptual speed measures, then it is not surprising that
covarying out these measures ﬁ'om the regression analyses would account for much of the
age-related variance in performance. Given this argument, the present ﬁndings then are not
inconsistent with the inhibition view of aging; however, further research is needed to
examine this issue. In particular, using a perceptual speed measure that presents items one

at a time, thereby reducing visual distraction, may provide very different results.

Riationship between inhibitory processes We within the older age spectrum

Contrary to expectations, generally no age differences were observed between the
two oldest age groups in the n-back and item recognition task (except for increased
interference effects in adults over the age of 75 in the categorized list condition of study

2). These results are not consistent with the bulk of literature that has demonstrated

62

continued cognitive declines with advancing age (Christensen et al., 1994; Osterweil et al.,
1994) including continued declines in inhibitory functioning (Persad et al., in press but see
Speiler, Balota, & Faust, 1996). Given that age diﬁ‘erences on other standardized
measures were obtained (i.e. WCST, perceptual speed measures) consistent with the aging
literature, it is diﬁcult to account for this ﬁnding. As already noted, characteristics of the
older adults in this study may have contributed to the lack of differences; in particular the
older individuals used in this study were generally highly educated and may not be a
representative sample of the older population. In comparison, the older adults tested in the
Persad et al. study (in press) were recruited actively from the local community and on
average had lower education levels and thus may have been more representative of the
general older population (Mean education for adults over the age of 75 in Study 1 =15.3,
Study 2 = 14.3, Persad et al. study = 12.6). Further research is needed to determine

whether inhibition does continue to decline with advancing age.

FUTURE DIRECTIONS

A number of relatively simple studies can be designed using minor alterations to
the tasks employed in these studies to further examine inhibition and its role in age-related
decrements in working memory. Since the original aim of Study 2 was to increase
interference eﬁ‘ects and hence the need for inhibition to successful performance, a ﬁrst
goal would be to re-run Study 2 using a word list that was properly designed to increase
interference effects. Simply making some minor alterations to the categorized list
condition could do this. Basically, the same stimuli and procedure could be used; however,

the format of the categorized list would need to be revised. In particular the use of words

63

from a limited number of categories that repeat throughout the list, as opposed to the use
of a blocking procedure, could achieve the desired affect. This format should eliminate the
“release from proactive interference” design allowing for a presumed build-up in
interference effects due to the strong pre-existing associative bonds that have been
hypothesized to exist for categorized words.

A more fruitful line of research would probably make use of the n-back task
because reliable age differences were found with this measure. In addition to needing to
replicate the age ﬁndings, the stimuli could also be designed to increase interference
without changing the underlying task procedure. In this study, it was stipulated that no
letter was repeated within a series (unless part of a positive or recent negative trial).
Increasing the number of times a letter was presented in a list should lead to increased
interference effects; however it would be important to try to distinguish age effects from
other age-related changes such as diminished capacity for recalling the temporal order of
items (Kausler, Salthouse, & Saults, 1988). This task could also prove especially useful
for examining the relative roles of processing speed and inhibition because of the clear
evidence for age differences in performance. In addition, as stated above, further study
with neuroimaging techniques to look at underlying neuroanatorrrical systems that are

involved in age-related declines in inhibition could prove to be very interesting.

Working Memory Subsystems and Inhibition
The present study focused on age differences in two working memory measures,
the n-back and the item recognition task, both presumed to rely on verbal working

memory processes. Yet, ﬁndings from both the neurocognitive and neuroimaging literature

64

suggest the presence of three separate subsystems in working memory, each of which
deals with a different type of stimuli, and each of which is subserved by a diﬁ‘erent region
of the brain. In particular, these three subsystems have been referred to as verbal, spatial,
and object working memory. The verbal working memory system, as described earlier,
comprises Baddeley’s original conception of a phonological loop (Baddeley, 1992;
Baddeley, & Hitch, 1994). However, the original conception of the visual-spatial
Sketchpad has been ﬁrrther divided into a system that is involved in the spatial aspects of
stimuli, such as location or orientation in space, versus the identiﬁcation of an actual visual
object.

As discussed earlier, some researchers have suggested that there are multiple
inhibitory mechanisms, only some of which are affected with age. Different suggestions
have been made, however, about the particular mechanisms associated with age changes.
For example, Carlson et al. have hypothesized that there is an inhibitory mechanism that
deals with spatial location, and that it is this mechanism that is age invariant (Carlson,
Hasher, Connelly & Zacks, 1995). This inhibitory mechanism also has been linked to the
occipitoparietal or dorsal neural pathway, which has been linked to the processing of
location. By contrast, age declines in inhibition associated with indentiﬁcation of stimuli
are thought to be related to the ventral pathway. Kramer and colleagues (1994) also
hypothesized multiple inhibitory mechanisms, but have proposed that only those that rely
on frontal lobe functioning are affected by age. On the premise that there are multiple
inhibitory pathways and that only some are affected by age, it would be interesting to
design studies to look at age-related inhibitory ﬁrnctioning in the three separate working

memory subsystems. A ﬁrst general question is to examine whether there is a global

65

decline associated with inhibition that cuts across all working memory systems or whether
there are more speciﬁc age changes that are dependent both on the task processes and the
type of stimuli used.

In summary, the results ﬁom both studies presented in this paper provided
evidence of increased interference effects with age that are consistent with the viewpoint
that inhibitory mechanisms become less efﬁcient with age, leading to reduced performance

on working memory tasks.

66

FOOTNOTES

1. Due to the design of the n-back task a trimming procedure was not necessary. In the n-
back task stimuli were continually presented at a rate of one every 3 seconds in total and a
response had to be made within this time ﬂame or else it was considered an error. This is
in comparison to the item recognition task where the next trial was presented only after a

response was made by the participant.

2. Although vocabulary skill was not expected to inﬂuence performance as much in the n-
back task, a possible relationship vocabulary skill and n-back performance was examined;
however, none of the correlations between the Shipley vocabulary score and any of the n-

back scores were signiﬁcant.

67

APPENDIX

68

APPENDIX

Table 1. Age, Education and Vocabulary scores for the three age groups.

 

 

 

 

 

 

 

 

 

N Age Education Shipley Vocabulary
Mean (SD) Mean (SD) Mean (SD)
Young 20 (3M, 17F) 19.35 Q23) 13.05 (1.19) 30.16 (3.23)
Young-Old 20 (9M, 11F) 69.40 (4.03) 14.00 (2.45) 32.40 (5.59)
Old-Old 20 (9M, 11F) 78.10 (2.51) 15.30 (3.11) 36.15 (3.13)

 

69

 

3:038 05 E xoﬁtm 82m 05 3:088 88m 395 uo>uewoz 258% 6
x03; 80: 05 :39: 8: 806 So: 389 ”2:692 a
08258 05 E 3.23; So: on“ 8:89: 88m Boa ”338m .

 

 

 

 

 

 

 

 

 

 

A863 933 66.23 5.8 A38 933
3.5 Ems 33$ .62 3.8 3.3 20-20
A833 3.8: $55 ANS: $6.8 ans 20
3.3. Sam. «38 2.8 3.3 an: 98>
swab €6.33 £23 833 $8 3.3
3.5 «.2: 8.38 2.2 2.3 3.8 was.»
o>uwwoz eoZHMNoZ
Booed ozﬁwoz 93mg 808% $3332 eo>Emom ow<
89 $62 685 85 6283 SE 582 9562 688.2

 

 

 

 

._ gnaw E 3:on own 8:: 05 8m aoEvcoo 633.; 05 E 83 08: uncommom Ea 3838.4 .m 033.

NHQZm—mn?‘

70

00000000 0:: E :00:-m 80: 0:0 8:808 :8: 0:03 “030082 00003: e

0.00:-N 80: 0:0 :89: :0: 800 80: 0:0.& “030082 2.

00000000 0:: E 0.00:-N 80: 0:: 8:308 80: 0:08 ”02:00am .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

80.38 20. 88 Godm D 945.2. 3 @008 80.x 3 Rod 9%: C 20

3.33 3...? «~63. 3."? 3.: 2.6: 3.: 3:2. -20

828 :83: 6.50 @200 :08 9?: 399 3.0: 20

3.3: and: was: 3.03. 3.2. 3.": 3.3 3.2. -300?
@343 $6.68 Game: $20: $3: $35 so: as

mmﬁmn ESE. Same maﬁa: 3.3 aﬁnu Nada was” 950%

0>wmm02 6030032

:02: “800% 03882 03:08 0:009: 6003: o0>$0m02 .0250“: 0w<

8m: :82 @095 08C. 08008”: am: :00: 80.5008. “000008:

 

 

.8000» owe. 000:: 0:0 00.: 0030000 030:-N 0:0 5 800 08: 08098:: 000 >000=00< .m 2:3.

XHDmeﬁ

71

>080 80:88 0:: .:0 8:808 0 003 8: >080 :88: 0:: .:0 8:808 0 :0: m: 80:: 0:05 ”03:08:: 80005 0

>080 880: 0:: .:0 8:808 0 :08 m: 80:: 0:05 83:08:: 6

50.20 80:88 0:: :80 >080 880: 0:: :0 8:808 0 m: 80: 0:05 80:08:00.: 2.
50.80 :88: 0:: .:0 8:808 0 0.: 80:: 0:05 ”03:85 .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.0:: $5.02: $0.08: $32: :33 60.3 8.0: $8: 20
8.35: 00:5: 8.8:. 8.80 8.3 8.8 8.3 00.: -20
900mm: $0.80 253 2.0.0:: :80: 30.3 :80: :3: 20-
8.88 «8.8 8.3: 3.80 05.0 8.8 :0: 8.8 08$
:88: 5.8:: 5.8: 3.8:: $3.: :80: :3: 5.3
8.80 8.80 $.08 80% 8.2 3.8 8.? an; 88.»
0>::0w0Z 0 003882 a .
3000M >mwmmoz :Omwwﬁmomm 03:60am 3000M 00>m§woz Gomuﬁmzodm o>£mom ow<
8mm 0002 A008: 085. 08080.5 am: 0002 80800:: 80085

 

 

0500.8 0w0 08:: 0:: 00080 :8: 00580005 80:: 0:: 80: 0:00 0::: 080585 :80 50080014 .0 030,—.

X592m55<

72

APPENDIX

Table 5. Means and Standard deviations for WCST variables for the three age groups.

 

 

 

 

 

 

 

 

 

Age Number Correct Perseverative # of Categories
Responses

Young 70.40 (6.98) 16.06 (10.89) 5.35 (1.27)

Young-Old 73.84 (10.74) 17.40 ( 9.68) 5.00 (1.45)

Old-Old 66.84 (16.80) 30.23 (28.21)‘ 3.32 (2.36) "b

 

' old-old group is signiﬁcantly different than the young group.

b old-old group is signiﬁcantly different than the young-old group.

73

APPENDD(

Table 6. Age, Education and Vocabulary scores for the ﬁve age groups in Study 2.

 

 

 

 

 

 

 

Age N Age Education Shipley Vocabulary
Mean (SD) Mean (SD) Mean (SD)
18-29 20 (9M, 11F) 20.30 (2.58) 13.1 (1.37) 28.06 (4.01)
30-44 20 (9M, 11F) 36.45 (3.76) 14.3 (1.78) 31.75 (4.58)
45-59 20 (6M, 14F) 51.60 (4.33) 14.5 (2.48) 35.00 (3.49)
60-74 20 (12M, 8F) 69.55 (4.07) 15.2 (2.46) 35.12 (3.15)
75+ 20 (6M, 14F) 79.05 (3.56) 14.3 (1.92) 35.94 (2.78)

 

 

 

 

 

74

 

88 00030.8 0:: :0 8:808 0 003 :0: >88 :88: 0:: :0 8:808 0 :00 0: 80: 0:05 ”02882 80005 0
8:8 :0w8: 0:: :0 8:808 0 :00 0: 80:: 0:05 ”030802 0

808 000308 0:: 000 8:8 :88: 0::-:0 8:808 0 0: 80:: 0:05 80:80:05 0

>88 :88: 0::.:0 8:808 0 0: 80: 0:05 ”03:05 .

 

 

 

 

 

 

 

 

 

 

 

 

 

75

90.3: :80: :80: :000: :80: :80: :80: :000:
0000 5.5 00.8 8.: 00.00 00.3 8.00 00.00 +00
:000: :80: :00: :20: :80: :3: :20: :05:
0000 8.00 0...: 0000 00.0: 8.00 00.00 800 £00
:80: :80: :80: :00: :30: :00: :03: :03:
00.00 8.00 00.8 00.00 0000 2:00 0000 00.00 00.00
:02: :80: :03: :80: :03: :00: :000: :80:
00.00 00.00 00.0: 00.00 00.00 02.... 00.00 00.2: 3.00
:80: :05: :03: :000: :20: :80: :03: 80.0:
00.3 00.00 0000 :00 0000 0:... 800 00.00 00.0:
03882 0 00.5302 0
8000:: 2:802 00:08:05 03:05 8000.5 0030802 00:80:00.5 .0285 0?:
80: :82 :00: :82
Ham-H Eamon—00 how-38¢: 300.5% :mm-H toga—0.53 how-50040: :Goouom

 

 

 

 

 

05008 08 0:6 0:: 00080 :00: 00830005 80:: 00800380 000 008—083 0:: 800-: 08:: 58800:». N 500% .0 050-:-

X:Q2m55<

88 00030.8 0:: .:0 8:808 0 003 :0: >88 :88: 0:: .:0 8:808 00 :00 0: 80:: 0:05 ”0.0::802 :00005 0

00:8 :88: 0:: .:0 8:808 0 :00 0: 80:: 0:05 "03:80:: 0

>88 00030.8 0:: 08 00:8 :88: 0::-:0 8:808 0 0: 80:: 0:05 ”00:80:00.5 0
.088 :88: 0:: .:0 8:808 0 0: 80:: 0:05 ”030005 .

 

 

 

 

 

 

76

 

 

 

 

 

 

 

 

00.000: 000.000: 00.00:: 00.30: 20.000: 000.000: :000000 5.00::
00.00:: 00.:00 00.000 00.000 000:: 00.000 00.000 0.0.20 +00
:00000: :0000: :0000: 00.00:: :0.00:: :00000: :00.00:: :00.00::
00.000 00.000 00.000 00.000 3.000: 00.0.0 00.000 00.000 00.00
:00000: :00._0:: :00.00:: 00.00:: :00000: 00.0.2: :00000 :0000:
00.000 00.000 00.0.00 00.000 3.000 00.000 00.000 00.000 00.00
:00000: :00000: :000: :00000: 20.00: :00000: 00.00:: :0000:
00.000 00.000 00.000 00.000 00.000 00.000 00.000 00.000 3.00
:00000: :0003: :0000 00.00:: :00000 :00-.000: :00000: :00.:0::
00.000 00.000 00.000 00.:0 00.000 00.000 00.000 00.000 00-0:
03:80:: 0005802 0
«coco-m o>m~wwuz demand—moan o>Emom anon“ oo>ﬂmwoz coﬂaﬁmoﬁm going-om ow<
000: 20:0 :00 5.0:
:05 08:80:00 00008: 08:9 00000005 :05 008—083 00008: 08:. 00000005

 

 

 

00008 80 02.: 0:: 00080 :00: 005080005 80:: 0000:8080 :00-0 00:20.05 0:: 80.0.: 08:: 0::: 00005005 N 000% .w 2:05.

X52554

 

APPENDIX

Table 9. Raw scores from the Perceptual Speed Measures for the different age groups in

 

 

 

 

 

 

 

Study 2.

Age Group Letter Comparison Pattern Comparison
Mean (SD) Mean (SD)

18-29 10.53 (2.76) 19.03 (3.00)

30-44 9.00 (2.61) 17.08 (3.82)

45-59 9.38 (2.30) 15.90 (2.41)

60-74 6.48 (2.00) 13.03 (2.81)

75+ 6.50 (1.56) 11.20 (1.83)

 

 

 

 

77

APPENDIX

Table 10. Results from the hierarchical regression analyses for the recent negative

condition in Study 2.

 

Controlled Variable

Accuracy Data

R2

Increase in R2

Response Time Data

R2 Increase in R2

 

 

Age 013* 005*

Speed Composite 019* 009*

Age 020* 0.01 009* 0.00
p < .05

78

APPENDIX

 

 

 

Iw
+
l—l

 

 

 

 

 

 

1 000m

RESPONSE

NO NO
(2-back)

Figure l. Figural Description of the N-back task.

 

 

 

 

 

 

(3-back match)

79

+ I L . I
‘——'D
2000111
S
NO YES

 

 

lw

 

 

NO

 

 

ID

 

NO

 

APPENDIX

 

 

 

  

110 7
z
100 j
g 2
g 90 j
2 80 4 INegative 1
‘5' 70 j I Recent Negative
0 .
a 601
CL 9
50 ‘
40

Young Young- Old-Old
Old

Figure 2. A Comparison of Accuracy Data in the Negative and Recent Negative
conditions for the 2-back task.

80

APPENDIX

 

 

 

I Negative
I Recent Negative

 

    

           

 

 

  

 

  

  

ii. a l A ‘l l1 ii1.i
5 0 5 0 5
9 9 8 8 7

38:85 .523.

  

    

              

18-29 30-44 45-59 60-74 75+

Figure 3. Study 2 Comparison of Recent Negative and Negative Accuracy

data for the Item Recognition task across the ﬁve age groups.

81

APPENDIX

 

 

 

 

 

 

 

 

 

 

1200 a
1 100 /}‘
+ Unrelated-
g 1000 Negative
'; / -l- Unrelated Recent
g 900 / negative
5 +Categorized
8- Negative
:2“ 300 ‘ + Categorized
M Recent Negative
700
/
600 r i T i i

 

18-29 30-44 45-59 60-74 75+
Age

Figure 4. Response time for the Negative and Recent Negative Conditions for the
Categorized and Unrelated Lists.

82

APPENDIX

 

 

 

 

>

U

2

3

2 I Young
'5 I Old

6

2

d)

(L

 

Negative Recent Negative

Figure 5. Item Recognition Accuracy Performance comparing Young and Old
Adults collapsed across studies.

83

APPENDIX

1200 T

1000

 
  
   

800

600

I ._ ._..14 -. 4 . 4.. __...._

IYoung
I Old

400

Response Time (msecs)

N
O
O
.-. "-I_441-.

Negative Recent Negative

Figure 6. Item Recognition Response Time data comparing Young and Old adults
collapsed across studies.

84

REFERENCES

85

 

REFERENCES

Anders, T. R. & Fozard, J. L. (1973). Eﬂ‘ects of age upon retrieval from primary
and secondary memory. Developmental Psyphology, 9, 411—415.

Anders, T. R., Fozard, J. L, & Lillyquist, T. D. (1972). Eﬂ‘ects of age upon
retrieval from short-term memory. Developmental Psychology, 6, 214-217.

Anderson, S. W., Damasio, H., Jones, R. D., & Tranel, D. (1991). Wisconsin card
sorting test performance as a measure of frontal lobe damage. Journ_al of Clinical and

Experimental Neuropsychology, 13, 909-922.

Arbuckle, T. Y. & Gold, D. P. (1993). Aging, inhibition, and verbosity. Journal
of Gerontology: Psychological SciencesAB. P225-P232.

Awh, E., Jonides, J., Smith, E. E., Schumacher, E. H., Koeppe, R. A., & Katz, S.
(1996). Dissociation of storage and rehearsal in verbal working memory: Evidence from

PET. Psychological Science, 7, 25-31.

Baddeley, A. D. (1992). Working Memory. Science 225, 556-559.

 

Baddeley, A. D. & Hitch, G. J. (1994). Developments in the concept of working
memory. Neurops'Lchology, 8, 485-493.

Battig, W. F. & Montague, W. E. (1969). Category norms for verbal items in 56
categories: A replication and extension of the Connecticut category norms. Journal of
Experimental Psychology, 80, 1-46.

Beck, A. T. (1978). Beck Depression Inventory. San Antonio, TX: Psychological
Corporation.

Berman, K., F., Ostrem, J. L., Randolph, C., & Gold, J. (1995). Physiological
activation of a cortical network during performance of the Wisconsin Card Sorting Test:
A positron emission tomography study. Neuropsychologja, 33, 1027-1046.

Bjork, R. A. (1989). Retrieval inhibition as an adaptive mechanism in human
memory. In H.L. Roediger and F.I.M. Craik (Eds) V_arieties of Memory a_nd
Consciousness. Essaysin Honour of Endel Tulving. New Jersey: Lawrence Erlbaum
Associates.

Bjork, E. L. & Bjork, R. A (1996). Continuing inﬂuences of to-be-forgotten
information. Consciousness and Cogmtiop, 5, 176-196.

86

 

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

Brink, T. L., Yesavage, J. A, Lum, 0., Heersema, P. H., Adey, M., & Rose, T. L.
(1982). Screening tests for geriatric depression. Clinical Pachologist, 1, 37-43.

Butler, K. M., Zacks, R.T., & Henderson, J. M. (1999). Suppression of reﬂexive
saccades in younger and older adults: Age comparison on an antisaccade task. Memory

and Cogpition, 27, 584-591.

Carlson, M.C., Hasher, L., Connelly, & Zacks, R. T. (1995). Aging, distraction,
and the beneﬁts of predictable location. Pychology and Aging, 10, 427-436.

Chao, L. L. & Knight, R. T. (1997). Prefrontal deﬁcits in attention and inhibitory
control with aging. Cerebral Cortex. 7. 63-69.

Christensen, H., MacKinnon, A., Jorrn, A.F., & Henderson, AS. (1994). Age
differences and interindividual variation in cognition in community dwelling elderly.

Psychology and Aging, 9, 381-390.

Coﬁ'ey, C. E. (2000). Anatomic imaging of the aging human brain. In C. E. Coﬂ‘ey
& J. L. Cummings (Eds) Textboole Geriatric Neuropaychiatg. Washington: American
Psychiatric Press, Inc.

Cohen, J. D., Perlstein, W. M., Braver, T. S., Nystrom, L. E., Noll, D. C., Jonides,
J., & Smith, E. E. (1997). Temporal dynamics of brain activation during a working
memory task. Nature 386, 604-608.

 

Connelly, S. L., & Hasher, L. (1993). Aging and inhibition of spatial location.
Jourpal of Experimental Psychology: Human Perception and Performance. 19. 1238-1250.

Connelly, S. L., Hasher, L. & Zacks, R. T. (1991). Age and reading: The impact of
distraction. Psychology and Aging, 6, 533-541.

Craik, F. I. M. & Anderson, N. D. (1999). Applying cognitive research to
problems of aging. In D. Gopher & A. Koriat (Eds) Attention apd Performance XVII.

Craik, F. I. M. & Byrd, M. (1982). Aging and cognitive deﬁcits: The role of
attentional resources. In F. I. M. Craik & S. Trehub (Eds) Aging and cogm’tive processes.
New York: Plenum Press.

Cronin-Golomb, A. (1990). Abstract thought in aging and age-related neurological

disease. In F. Boller and J. Grafman (Eds) H_andboolﬂf Neuropsychology. Vol 4.
Elsevier Science Publishers.

87

 

Dempster, F. N. (1992). The rise and fall of the inhibitory mechanism: Toward a
uniﬁed theory of cognitive development and aging. Develgpmental Review 12, 45-75.

 

Dobbs, A. R. & Rule, B. G. (1989). Adult age diﬂ'erences in working memory.
Psychology and Aging, 4, 500-503.

Erikson, C. W., Hamlin, R. M., & Daye, C. (1973). Aging adults and rate of
memory scan. Bulletin of the Psychonomic Sociegi, 1, 259-260.

Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-Mental
State: A practical method for grading the cognitive state of patients for the

clinician. Journal of Psychiatric Research, 12, 189-198.

Fuster, J. M. (1997). The Prefrontal Cortex. Anatomy, Physiology, and
Neuropsychology of the FroLtal Lobe. Third Edition. New York: Lippincott-Raven.

Geiselman, R.E. & Bagheri, B. (1985). Repetition effects in directed forgetting:
Evidence for retrieval inhibition. Memory apd Cogm'tiop, 13, 57-62.

Gerard, L., Zacks, R. T., Hasher, L. & Radvansky, G. A. (1991). Age deﬁcits in
retrieval: The fan effect. Journal of Gerontology: Pychological Sciences, 46, P13 1-136.

Gevins, A. & Cutillo, B. (1993). Spatiotemporal dynamics of component processes

in human working memory. Electroencephalography and clinical Neurophysiology, 87,
128-143.

Gold, T. Y., & Arbuckle, D. P. (1995). A longitudinal study of off-target
verbosity. Jourpal of Gerontology: Psychological Sciences, 50B, P307-315.

Gur, R. C., Gur, R. E., Obrist, W. D., Skolnick, B. E., & Reivitch, M. (1987). Age
and regional cerebral blood ﬂow at rest and during cognitive activity. Archives of Geneﬂ
Psychiatry. 44. 617-621.

Hamm, V. P. & Hasher, L. (1992). Age and the availability of inferences.
Psychology and Aging, 6, 587-594.

Hartman, M., & Hasher, L. (1991). Aging and suppression: Memory for
previously relevant information. Paychology and Aging, 6, 587-594.

Hasher, L., Stoltzfus, E. R., Zacks, R. T., & Rypma, B. (1991). Age and

inhibition. Journal of Experimental Psychology: Learning. Memory and Cogpition. 17.
163-169.

88

 

Hasher, L. & Zacks, R. T. (1988). Working memory, comprehension, and aging: A
review and a new view. The Paychology of Learning and Motivatiop, 22, 193-225.

Hasher, L. Zacks, R. T. & May, C. (1999). Inhibitory control, circadian arousal,
and age. In D. Gopher & A. Koriat (Eds) Attention d Performance XVII o 'tive

Regplation of Performance: Interaction of Theory and Application. Cambridge, MA: MIT
Press.

Haug, H. & Eggers, R. (1991). Morphometry of the human cortex cerebri and
corpus striatum during aging. Neurobiology of Aging, 12, 336-338.

Heaton, R. K., Chelune, G. J., Talley, J. L., Kay, G. G. & Curtiss, G. (1993).
Wisconsin Card Sorting Test Manual: Revised and Expanded. Odessa: Psychological
Assessment Resources, Inc.

Hertzog, C. (1996). Research design in studies of aging and cognition.
In J. E. Birren and K. Warner Schaie (Eds) Handbook of the Psychology of
Aging. San Diego: Academic Press.

Jonides, J., Marshuetz, C., Smith, E. E., Renter-Lorenz, P.A., Koeppe, R. A., &
Hartley, A. (2000). Age differences in behavior and PET activation reveal differences in

interference resolution in verbal working memory. Journal of Cognitive Neuros_cience. 12.
188-196.

Jonides, J., Smith, E. E., Marshuetz, C., Koeppe, R. A., & Renter-Lorenz, P. A.
(1998). Inhibition in verbal working memory revealed by brain activation. Proceedings of
the National Acaderry of Science. 95. 8410-8413.

Just, M. A. & Carpenter, P. A. (1992). A capacity view of comprehension:
Individual differences in working memory. Psychological Review. 99. 122-149.

Kausler, D.H., Salthouse, T.A. & Saults, J. S. (1988). Temporal memory over the
adult lifespan. American Joumﬂ of Psychology, 101, 207-215.

Keys, B. A, & Whiter, D., A. (2000). Exploring the relationship between age,
executive abilities, and psychomotor speed. Journal of the International
Neuropsychological Sociegi, 6, 76-82.

Koriat, A., Ben-Zur, H., & Sheﬁ‘er, D. (1988). Telling the same story twice:
Output monitoring and age. JouLnal of Memory and Langpage, 27, 23-39.

Kuchera, H. & Francis, W. (1967). Computational analysis of present-day
American English. Providence, RI: Brown University Press.

89

Kramer, A. F., Humphrey, D. G., Larish, J. F., Logan, G., D., & Strayer, D. L.
(1994). Aging and inhibition: Beyond a unitary view of inhibitory processing in attention.
Psychology and Aging, 9, 491-512.

Langley, L.K., Overmier, B., Knopman, D.S., & Prod’Homme, M.M. (1998).
Inhibition and habituation: Preserved mechanisms of attentional selection in aging and
Alzheimer’s disease. Neuropaychology, 12, 353-366.

Larrabee, G. J ., Trahan, D. E., Curtiss, G., & Levin, H. S. (1988). Normative data
for the verbal selective reminding test. Neuropgichology, 2, 173-182.

Lezak M. D. (1995). Neuropsychological Assessment. Third Edition. New York:
Oxford University Press.

Li, K. Z., H., & Bosman, E. A. (1996). Age differences in Stroop-like interference

as a function of semantic relatedness. Aging, Neuropgzchology, and Cogm'tiop, 3, 272-
284.

Light, L. L. (1991). Memory and aging: Four hypotheses in search of data. Annual
Review of Pavchology, 42, 333-376.

Lustig, C., Tonev, S. T., & Hasher, L. (2000). Visual distraction and processing
speed 1. Presented at the Cognitive Aging Conference, Atlanta, Georgia.

May, C. P., Hasher, L. & Kane, M., (1999). The role of interference in memory
span. Memory and Cognition. 27. 759-767.

May, C. P., Kane, M., & Hasher, L. (1995). Determinants of negative priming.
Psychological Bulletin. 118. 35-54.

Milner, B. (1964). Some effects of ﬁontal lobectomy in man. The frontal granular
cortex and behavior. New York: McGraw-Hill.

Nettelbeck, T. & Rabbitt, P. M. A. (1992). Aging, cognitive performance, and
mental speed. Intelligence, 16, 189-205.

Neumann, E. & DeSchepper, B. G. (1992). An inhibition-based fan effect:
evidence for an active suppression mechanism in selective attention. Canadian Journal of

Psychology, 46, 1-40.

Osterweil, D., Mulford, P., Syndulko, K., and Martin, M. (1994). Cognitive
ﬁrnction in old and very old residents of a residential facility: Relationship to age,
education and dementia. Journal of the America; Geriatrics SocietyL42. 766-773.

90

Pietrini, P. & Rapoport, S. I. (2000). Cerebral blood ﬂow and glucose metabolism
in healthy human aging. In C. E. Coﬂ‘ey & J. L. Cummings (Eds) Textbook of Geriatric
Neuropsychiatiy. Washington: American Psychiatric Press, Inc.

Persad, C. C., Abeles, N., lacks, R. T., & Denburg, N. (in press). Inhibitory
changes after age 60 and their relationship to measures of attention and memory. Journal

of Gerontology: Pachological Sciences.

 

Salthouse, T. A (1991). Mediation of adult age differences in cognition by
reductions in working memory and speed of processing. Paychological Sci_ep_ce- 2. 179-
183.

Salthouse, T. A. (1993). Speed mediation of adult age diﬁ‘erences in cognition.
Developmental Review, 14, 413-437.

Salthouse, T. A. (1994). The aging of working memory. Neuropaychology, 8, 535-
543.

Salthouse, T. A. (1994). Aging associations: Inﬂuence of speed on adult age
differences in associative learning. Journal of Experimental Psychology: Learning,
Memory. and Cogaition. 20, 1486-1503.

Salthouse, T. A. (1996). The processing-speed theory of adult age differences in
cognition. Psychological Review, 103, 403-428.

Salthouse, T. A & Coon, V. E. (1993). Inﬂuence of task-speciﬁc processing

speed on age differences in memory. Journal of Gerontology: Psychological Sciencea, 48.
P245-P255.

Salthouse, T. A. & Coon, V. E. (1994). Interpretation of differential deﬁcits: The

case of aging and mental arithmetic. Journal of Experimental Psychology: Leaming,
Memogy and Cogpition. 20. 1172-1182.

Salthouse, T. A. & Meinz, E]. (1995). Aging, inhibition, working memory, and
speed. Journal of Gerontology: Pachological Sciences, 50B, P297-P306.

Schooler, C., Neumman, E., Caplan, L.J. & Roberts, B. R. (1997). Continued
inhibitory capacity throughout adulthood: Conceptual negative priming in younger and
older adults. Pachology and Agm' g, 12, 667-674.

Schumacher, E. H., Lauber, E., Awh, E., Jonides, J ., Smith, E. E., & Koeppe, R.

A. (1996). PET evidence for an amodal verbal working memory system. Neuroimage, 3,
79-88.

91

See, S. T. K. & Ryan, E. B. (1995). Cognitive mediation of adult age differences
in language performance. Psychology and Aging, 10, 458-468.

Shipley, W. C. (1946). Institute of Living Scale. Los Angeles: Western
Psychological Services.

Shimamura, A. P. & Jurica, P. J. (1994). Memory interference effects and aging:
Findings from a test of frontal lobe function. Neuropsychology, 8, 408-412.

Smith, E. E. & Jonides, J. (1997). Working Memory: A view from neuroimaging.
Cogrp'tive Psychology, 33, 5-42.

Smith, E. E., Jonides, J. & Koeppe, R. A. (1996). Dissociating verbal and spatial
working memory using PET. Cerebral Cortea, 6, 11-20.

Speiler, D. H., Balota, D. A., & Faust, M., E. (1996). Stroop performance in
healthy younger and older adults and in individuals with dementia of the Alzheimer’s type.

Journal of Experimental Psychology: Human Perception and Performance, 22, 461-479.

Stemberg, S. (1966). High speed scanning in human memory. Science. 153. 652-
654.

Stuss, D. T. & Benson, D. F. (1984). Neuropsychological studies of the frontal
lobes. Psychological Bulletin. 95. 3-28.

Sullivan, M. P. & Faust, M. E. (1993). Evidence for identity inhibition during
selective attention in old adults. Psychology and Aging, 8, 589-598.

Tipper, SP. (1985). The negative priming eﬂ‘ect: Inhibitory priming by ignored
objects. Qtaarterly Journal of Experimental Psychology, 37A, 571-590.

Tipper, S. R, Weaver, B., & Houghton, G. (1994). Behavioural goals determine

inhibitory mechanisms of selective attention. The Quarterly Journal of Experimental
Psychology . 47A. 809-840.

Tonev, S. T., Lustig, C., & Hasher, L. (2000). Visual distraction and processing
speed H. Presented at the Cogm’tive Aging Conference. Atlanta, Georgia.

Verhaegen, P. & Salthouse, T. A. (1997). Meta-analyses of age-cognition relations
in adulthood: Estimates of linear and nonlinear age effects and structural models.

Psychological Bulletin, 122, 231-249.

Veroff, A. E. (1980). The neuropsychology of aging. Psychological Research, 41,
259-268.

92

West, R., & Baylis, G. C. (1998). Effects of increased response dominance
and contextual disintegration on the Stroop interference effects in older adults.

Psychology and Agm’ g, 13, 206-217.

Wickens, DD. (1972). Characteristics of word encoding. In A.W. Melton &
E. Martin (Eds) Coding Propasses in human memogi. Washington, DC : Winston.

Wixted, J. T. & Rohrer, D. (1993). Proactive interference and the dynamics of
free recall. Journal of Experimenaal Psychology: Raining. Memory and Cognition.
_1_9, 1024-1039.

Zacks, R. T. & Hasher, L. (1994). Directed ignoring: Inhibitory regulation of
working memory. In D. Dagenbach & T. H. Carr (Eds) iahibitog Processes in
Attention. Memory and Lapguage. San Diego: Academic Press.

Zacks, R. T, Radvansky, G., & Hasher, L. (1996). Studies of directed

forgetting in older adults. Journal of Eap_erimental Psychology: Learning, Memog
and Cogpitiop, 22, 143-156.

93

"Illllllllllllilllllilllls