close

Вход

Забыли?

вход по аккаунту

?

951

код для вставкиСкачать
r
Human Brain Mapping 5:389–409(1997)
r
Selective and Divided Visual Attention:
Age-Related Changes in Regional Cerebral
Blood Flow Measured by H215O PET
David J. Madden,1* Timothy G. Turkington,2 James M. Provenzale,2
Thomas C. Hawk,2 John M. Hoffman,3 and R. Edward Coleman2
1Center
for the Study of Aging and Human Development, and Department of Psychiatry and
Behavioral Sciences, Duke University Medical Center, Durham, North Carolina 27710
2Department of Radiology, Duke University Medical Center, Durham, North Carolina 27710
3Department of Neurology, Emory University School of Medicine, Atlanta, Georgia 30322
r
r
Abstract: Regional cerebral blood flow (rCBF) was measured using H215O and positron emission
tomography (PET) to test the hypothesis that age-related changes in the pattern of rCBF activation would
be greater under divided attention conditions than under selective attention conditions. Subjects were 24
right-handed men: 12 young adults (age 21–28 years), and 12 older adults (age 60–77 years). Measurement
of rCBF was obtained during performance of three visual search task conditions, each of which involved
viewing a series of nine-letter displays and making a two-choice button press response to each display.
Analyses of subjects’ mean reaction time and error rate confirmed that older adults’ search performance
was disproportionately impaired when it was necessary to divide attention among the display positions.
The rCBF data indicated that attending selectively to a target letter in a known (central) location was not
associated with cortical activation for either age group. The requirement to divide attention among the
display positions led to rCBF activation in occipitotemporal, occipitoparietal, and prefrontal cortical
regions. In the divided-attention condition, rCBF activation in the occipitotemporal pathway was relatively
greater for young adults; activation in prefrontal regions was relatively greater for older adults. These
differences in rCBF activation were related to search reaction time and suggest that, when attention was
divided, young adults’ performance relied primarily on letter identification processes, whereas older
adults required the recruitment of additional forms of task control. Hum. Brain Mapping 5:389–409,
1997. r 1997 Wiley-Liss, Inc.
Key words: neuroimaging; aging; vision; reaction time; information processing; cortical activation
r
r
INTRODUCTION
Current models of the functional neuroanatomy of
visual attention indicate that attentional processing is
Contract grant sponsor: National Institute on Aging; Contract grant
number: R01 AG11622.
*Correspondence to: David J. Madden, Ph.D., Box 2980, Duke
University Medical Center, Durham, NC 27710. E-mail:
[email protected]
Received for publication 23 July 1997; accepted 1 August 1997
r 1997 Wiley-Liss, Inc.
mediated by several interconnected cortical and subcortical areas [LaBerge, 1995; Posner, 1995]. Investigations
using positron emission tomography (PET) to measure
regional changes in cerebral blood flow (rCBF) and
metabolic rate for glucose (rCMRglc) have provided
important information regarding the brain regions
mediating different forms of attentional processing.
Selective attention appears to be expressed as an
enhancement of activity in the neural pathways relevant to task performance. Corbetta et al. [1990, 1991],
for example, reported that attending selectively to
r
Madden et al. r
either the shape or color of visually presented forms
led to rCBF activation in occipitotemporal regions,
whereas attending to velocity led to occipitoparietal
activation. Corbetta et al. [1990, 1991] also found that
dividing attention among several stimulus dimensions, rather than focusing on a single dimension,
activated regions outside of the visual system, in the
anterior cingulate gyrus and prefrontal cortex of the
right hemisphere. Activation of the prefrontal cortex
increases as a function of task difficulty and may
represent an increasing demand on attentional resources such as working memory [Grady et al., 1996].
When the task requires the spatial focus of attention to
be shifted to peripheral locations in the visual field,
activation is evident in the superior parietal cortex
[Corbetta et al., 1993, 1995]. Nobre et al. [1997] confirmed the association of parietal rCBF activation with
shifts of visuospatial attention, using a higher-resolution PET methodology that allowed the identification
of activation within individual subjects. These authors
proposed a large-scale neural system for visuospatial
attention comprising the posterior parietal cortex, anterior
cingulate cortex, and lateral and medial premotor cortex.
Behavioral research has established that changes in
attentional functioning occur during human aging
[Hartley, 1992; Madden and Plude, 1993; McDowd and
Birren, 1990]. A relatively consistent finding from this
research, in the domain of visual search and classification tasks, is that an age-related decline in performance
is magnified when attention must be divided among
several relevant stimulus inputs or tasks, as compared
to conditions in which attention can be focused on a
single task or input [Plude and Hoyer, 1986; Rabbitt,
1965]. Although an age-related slowing is evident in
virtually all stages of information processing [Myerson
et al., 1990; Salthouse, 1996], the age-related decline in
dividing attention is in some visual search tasks
greater than would be expected on the basis of a
generalized slowing. The age difference remains significant when the divided attention effect is expressed as a
relative (i.e., proportional) change, as well as an absolute one [Madden, 1986]. When a specific stimulus
dimension (such as target location) is specified in
advance, however, older adults often resemble young
adults in the ability to use the advance information to
improve search performance [Greenwood et al., 1993;
Hahn and Kramer, 1995; Hartley et al., 1990; Madden,
1992]: i.e., age-related performance declines are more
pronounced under divided attention conditions than
under selective attention conditions.
Tomographic measurement of rCBF and rCMRglc
has revealed age-related changes in neural function
that may underlie the changes in attention [Grady and
r
Haxby, 1995; Madden and Hoffman, 1997]. Under
resting-state conditions, an age-related decrease in
rCBF and rCMRglc has been observed to be more
pronounced for the prefrontal and parietal cortex
[Horwitz et al., 1986; Martin et al., 1991], which may
represent changes in attentional networks [Moeller et
al., 1996]. Activation studies, using subtraction methodology to measure changes in rCBF during performance
cognitive tasks, have documented age-related changes
in the pattern of rCBF activation. Grady et al. [1992,
1994] have developed visual discrimination tasks with
nonverbal stimuli that, when performed during PET,
lead to task-specific rCBF activation in the occipitoparietal and occipitotemporal pathways. Comparison of
young and older adults demonstrated that there was
an age-related decline in functional differentiation
between the tasks: i.e., rCBF activation outside of the
occipitoparietal pathway during a location-matching
task, and outside of the occipitotemporal pathway
during a face-matching task, was greater for older
adults than for young adults. These authors also found
that, in both tasks, areas associated with feature-level
processing (prestriate cortex) were relatively more
active for young adults. Areas in the prefrontal cortex
were relatively more active for older adults, especially
during the location-matching task. Grady et al. [1992,
1994] concluded that visual processing in the occipital
cortex was more efficient for young adults than for
older adults, and that the increased attentional demands of the location-matching task for older adults
led to the recruitment of processing mediated by the
prefrontal cortex. An age-related decline in rCBF activation in the occipitotemporal pathway has been
obtained with a visual word identification (lexical
decision) task [Madden et al., 1996].
Although PET investigations have yielded important data regarding the neural systems mediating
visual attention, and although age differences in the
pattern of rCBF activation appear to be related to
attentional demands, little information is currently
available on age-related changes in rCBF activation as
a function of specific forms of attentional processing.
In the present study we addressed this issue in a PET
investigation of adult age differences in rCBF during a
visual search task, with letter targets, under conditions
in which the location of the target was either constant or
varied across trials. In previous research with this type of
task, performance limitations occurring when the target
location is specified have been interpreted in terms of
failures of selective attention, whereas performance limitations associated with varying target location have been
interpreted in terms of the requirement to divide attention
among relevant stimulus inputs [Shiffrin, 1988].
390
r
r
Aging, Attention, and rCBF Activation r
The present task was a two-choice version of visual
search in which, on each of a series of trials, one of two
preassigned target letters was present in a display of
nine letters arranged in a 3 3 3 letter grid. Subjects
performed each of three task conditions during separate PET studies: a Central condition, in which the
target always occurred in the central display position,
a Divided condition, in which the target could occur in
any one of the nine display positions; and a Passive
condition, in which subjects viewed nine-letter displays that did not contain either of the target letters.
Subtractions between pairs of task conditions were
conducted to determine the pattern of rCBF activation
associated with attending selectively to a single display position (the Central condition) and dividing
attention among all the display positions (the Divided
condition). We also examined the correlation between
visual search reaction time and rCBF within each age
group, to determine the relation of the observed
changes in rCBF to subjects’ actual performance.
The results of research using behavioral measures of
cognitive performance led to the prediction that an
age-related decline in visual search performance will
be more pronounced under divided attention conditions than under selective attention conditions [Hartley, 1992; Madden and Plude, 1993]. The pattern of
rCBF activation for the Divided condition, which
requires processing of multiple display positions,
should involve the prefrontal and superior parietal
cortex [Corbetta et al., 1990, 1991, 1993, 1995; Nobre et
al., 1997]. The data of Grady et al. [1992, 1994] suggest
that regions outside the visual processing pathways,
especially in the prefrontal cortex, will be more active
for older adults than for young adults in the Divided
condition. The Central condition, in contrast, which
allows subjects to attend selectively to a single display
location, is more likely to be associated with rCBF
activation in the occipitotemporal pathway for both
age groups [Corbetta et al., 1990, 1991].
METHODS
Subjects
Twelve young adult men between age 21–28 years
(mean 5 24.33 years, SD 5 2.01), and 12 older adult
men between age 60–77 years (mean 5 65.50 years,
SD 5 5.20), participated. All subjects gave written
informed consent prior to participation, and the research procedures were approved by the Institutional
Review Board of the Duke University Medical Center.
All subjects were right-handed and had completed
some postsecondary school education. Subjects also
r
possessed corrected binocular visual acuity for near
point (40 cm) of 20/40 or better, as assessed from a
Keystone (Davenport, IA) Ophthalmic Telebinocular
Vision Tester. Subjects were free from major health
problems, as indicated by a screening questionnaire
[Christensen et al., 1992]. Each subject was also administered a neurologic screening examination by a physician. This included a brief clinical history, recitation of
a seven-digit numerical sequence, visual field testing
by confrontation, testing for abnormalities of motor
strength in face and limbs, sensory testing by light
touch, finger-to-nose and rapid alternating movement
examination, and evaluation for gait abnormalities.
Subjects completed several psychometric tests prior
to participation. All subjects scored 28 or higher (out of
30) on the Mini-Mental State Exam [Folstein et al.,
1975] and lower than 5 (out of 63) on the Beck
Depression Inventory [Beck, 1978]. The mean raw
score on the Vocabulary subtest of the Wechsler Adult
Intelligence Scale-Revised [WAIS-R; Wechsler, 1981], a
measure of verbal knowledge, was significantly higher
for older adults (mean 5 60.92, SD 5 4.98) than for
young adults (mean 5 55.67, SD 5 3.80) (t(22) 5 2.90,
P , .01). The mean raw score on the Digit Symbol
Substitution subtest of the WAIS-R, a measure of
perceptual-motor speed, was significantly higher for
young adults (mean 5 70.42, SD 5 9.10) than for older
adults (mean 5 47.08, SD 5 5.02) (t(22) 5 22.0,
P , .0001). The pattern of age differences in the WAIS
subtests is consistent with previous investigations
using these psychometric measures [Salthouse, 1982].
All subjects underwent MR imaging prior to PET
testing, and the MR images were examined by a
neuroradiologist for evidence of significant cerebral
atrophy or structural abnormality. Transaxial MRI
scans were acquired with a General Electric (Milwaukee, WI) 1.5 Tesla Signa System. Acquisitions were
made with a 3-mm slice thickness and no interslice
gap. T1-weighted images had a repetition time (TR ) of
600 msec and an echo delay time (TE ) of 20 msec, and
two excitations (NEX). T2-weighted images had a TR of
2,500 msec, TE values of 20 msec and 80 msec, and 1
NEX. The following exclusion criteria were applied to
both age groups: any signal abnormality indicating a
mass, ventricular enlargement, or atrophy atypical for
age, flow signal abnormality within intracranial vessels,
extraaxial fluid collection, and any focal signal abnormality within the caudate, putamen, globus pallidus, thalamus, brain stem, or cerebellum. For young adults, any
focal area of hyperintensity on T2-weighted images
was also an exclusion criterion, whereas for older
adults, the criterion was the presence of focal supraten-
391
r
r
Madden et al. r
Figure 1.
Examples of displays presented in the visual search task conditions. tion, none of the display letters was a target, and subjects
Subjects performed each of these three task conditions during two alternated responses across trials. In the Central and Divided
PET studies. During the PET study, subjects viewed a series of 162 conditions, each display contained one target letter and eight
nine-letter displays and made a two-choice button-press response nontargets. At display onset, subjects pressed the response button
to each display. Display duration was 1 sec, followed by an corresponding to the target. In the Central condition, the target
800-msec blank screen. For each subject, two letters were was always located in the central position of the display, whereas in
assigned as targets (one letter per response button), and this the Divided condition, the target could occur at any of the display
assignment remained constant across trials. In the Passive condi- positions.
torial hyperintense white matter signal abnormalities
greater than 3 mm.
Visual search task
During the PET studies, subjects performed a visual
search task that comprised three separate conditions:
Passive, Central, and Divided. There were six PET
studies per subject, and each subject performed a
single task condition during a PET study (i.e., each task
condition was performed twice). Each subject performed a total of 324 trials per task condition (162 in
each of two PET studies), and each trial involved
r
making a button-press response to a nine-letter display. Examples of the trial sequence for the task
conditions are presented in Figure 1.
Presentation of the displays and recording of subjects’ responses were controlled by a Tangent (Burlingame, CA) 386-processor microcomputer and Zenith
(St. Joseph, MI) ZCM-1490 video monitor. The monitor
was secured to a custom-made stand and positioned
directly above the PET bed, near the gantry opening.
The monitor screen faced downward at an approximately 45° angle, so that subjects could view the screen
while lying supine. Each nine-letter display was arranged as a 3 3 3 letter grid and subtended approxi-
392
r
r
Aging, Attention, and rCBF Activation r
mately 4.30° wide by 4.59° high at the viewing distance
of 40 cm. Each character space subtended approximately .86° square. The display items were presented
as white capital letters against a black background. The
duration of each display was 1 sec, followed by an
800-msec blank screen. Subjects responded by means
of a two-button computer mouse connected to the serial
port of the microcomputer. Subjects held the mouse in the
right hand throughout testing and used the first and
second fingers to press the response buttons. Reaction
time for the button-press response was measured from
the onset of the display.
In the Central and Divided conditions, each display
contained one of the two target letters and eight
different nontarget letters. Subjects were instructed in
these conditions to press one of two response buttons
at display onset, depending on which of two preassigned target letters was present in the display. In the
Central condition, subjects were informed that the
target would always occur in the central display
position. In the Divided condition, subjects were informed that the target could occur at any of the nine
display positions. The displays in the Passive condition contained only nontarget letters, and subjects
were instructed in this condition to press one of the
two response buttons at the onset of each display,
alternating responses across displays.
Each subject was assigned one of two stimulus lists.
In list 1, the target letters were K and P, and in list 2 the
target letters were F and R. For both lists, the nontarget
letters were B, C, D, G, H, J, L, M, N, Q, S, T, V, W, X, Y,
and Z. In the task condition associated with each PET
study, subjects viewed a series of 162 displays. In both
the Central and Divided conditions, each of the target
letters occurred 81 times, in a random sequence. Target
location was constant in Central condition; in the
Divided condition, each of the target letters occurred
nine times at each of the nine display positions. These
nine occurrences were also randomized. For each
display, nontarget letters were selected randomly, with
the two constraints that each letter occur only once per
display and that a particular letter not occur in the
same display position across two successive displays.
Within each age group, the order of the task conditions was counterbalanced so that, across the 6 subjects
assigned to each stimulus list, each of the task conditions occurred once at each of the six serial positions in
the sequence of PET studies. For each subject, the three
PET studies in each half of the testing sequence
included one instance of each of the three task conditions. Two instances of a particular task condition were
never presented consecutively. The assignment of tarr
get letters to response buttons was alternated across
the subjects assigned to each list.
After being positioned in the scanner, each subject
performed three blocks of practice trials that each included
10 displays for each of the three task conditions. The six
PET studies were administered after these practice blocks.
Positron emission tomography
The PET studies were conducted with a General
Electric Advance whole-body scanner containing 18
detector rings [DeGrado et al., 1994]. Data were acquired simultaneously from 35 imaging planes (18
direct planes and 17 crossplanes) separated by 4.25
mm. The axial field of view was 15.2 cm, and the
intrinsic in-plane and axial spatial resolutions were
approximately 5 mm FWHM. Data acquisition was
performed in the two-dimensional, high-sensitivity
mode (septa in).
At the beginning of the testing session, an intravenous catheter was placed in the subject’s left arm, for
radiotracer injection. The subject was positioned in the
tomograph with his head aligned in a plane approximately parallel to the glabella-inion line. Alignment
was conducted with the assistance of a low-power
laser. Prior to the emission scans, a 5-min transmission
scan was performed using a pair of 3–10 mCi 68Ge
rotating pin sources. Following the transmission scan,
subjects performed the three blocks of 10 practice trials.
For each emission scan, the radiotracer was administered
as an intravenous bolus injection of approximately 50 mCi
of H215O. Prior to each radiotracer injection, subjects were
reminded of the instructions for that particular task condition. Presentation of the visual displays began 1 min prior
to radiotracer injection and continued for approximately 5
min. PET data acquisition began automatically when the
radioactivity count rate exceeded a preset threshold of
100,000 counts/sec and continued for 1 min. Successive
scans were separated by 15 min. Reconstruction of the
PET image data was performed with filtered backprojection using a Hann filter, 128 3 128 pixel images
(35 slices) with 2 3 2 mm pixels. The data were
corrected for random coincidences, attenuation, scattered radiation, and dead time.
Data analysis
Cortical volume
Estimates of the volume of cortical gray matter were
obtained from each subject’s MRI, using methods
described by Madden et al. [1996]. Volumetric estimates were obtained separately for the total gray
matter of the left and right cerebral hemispheres
393
r
r
Madden et al. r
(excluding the basal ganglia), and for the left and right
cerebellar hemispheres.
Cerebral blood flow
The 1995 version of the Statistical Parametric Mapping (SPM) software [Friston et al., 1995] was used to
analyze the changes in rCBF between task conditions.
The scans for each subject were realigned using the
first scan in the Passive condition as a reference.
Following realignment, the scans were normalized and
transformed into a standard stereotaxic space [Talairach and Tournoux, 1988]. This procedure began with a
12-parameter affine transformation, followed by piecewise (contiguous transverse slices) nonlinear matching, constrained by a set of smooth basis functions
[Friston et al., 1996a]. The spatially normalized images
were smoothed with an additional 15-mm FWHM
isotropic Gaussian kernel.
The SPM analysis of the rCBF data used an analysis
of covariance (ANCOVA) to remove the effect of global
activity [Friston et al., 1990]. The ANCOVA was conducted on a voxel-by-voxel basis, and the resulting
adjusted rCBF values for each voxel were scaled to a
mean of 50 ml/100 g/min. To minimize reliance on the
assumption that the component processes of the task
remain constant in the context of additional task
demands [Friston et al., 1996b; Sergent et al., 1992], we
performed all three subtractions among task conditions: Central minus Passive, Divided minus Passive,
and Divided minus Central. A linear contrast was
constructed for each subtraction to represent rCBF
activation. The contrasts were reversed to yield the
corresponding decreases in rCBF. The resulting set of
voxel t values for each subtraction constitute the
statistical parametric map SPM5t6. The SPM5t6 values
were transformed to the unit normal distribution
SPM5Z6 and thresholded at 2.33 (or P 5 .01, uncorrected). The resulting foci were then characterized in
terms of the spatial extent (k) and peak height (u) of
local maxima. The local maxima were defined as
voxels with Z values greater than all voxels within 12
mm. Significance was estimated using distributional
approximations from the theory of Gaussian fields.
This characterization is in terms of the probability that
a region of the observed number of voxels or greater
would have occurred by chance (P[nmax . k]), and the
probability that the observed peak height of the local
maximum of activation is greater than would be
expected by chance (P[Zmax . u]). The estimates were
applied to the entire volume analyzed (i.e., a corrected
P value). Up to three local maxima are provided for
each region.
r
We also constructed composite MRI images of the
subjects in the present study, to provide an additional
anatomical reference for the PET data. Each subject’s
MRI was resliced and registered to the PET image
using a surface-fitting image registration technique
[Pelizzari et al., 1989; Turkington et al., 1995]. The
registered MRI data were taken through the stereotaxic
normalization based on the PET images, and a composite MRI was constructed for the young and older adult
groups by summing the images. Regions of rCBF
activation that were difficult to localize on the basis of
the atlas of Talairach and Tournoux [1988] were identified by superimposing the regions of rCBF activation
on the composite MRIs.
The potential age-related change associated with
each subtraction was assessed by including, in the
SPM analysis of the data for all subjects combined, two
simultaneous linear contrasts: one representing the
subtraction of interest and the other representing the
difference between the two age groups. The resulting
SPM5Z6 map for these interaction contrasts contained
those voxels that differed as a joint function of age
group and task condition.
The pattern of correlation between visual search
performance and rCBF was also examined within each
age group. In these analyses, each subject’s mean
reaction time for correct responses during each PET
study was included as a covariate of interest in SPM
analyses of each pair of task conditions (Central and
Passive; Divided and Passive; and Divided and Central). The regions of activation in these SPM5Z6 maps
consequently represented those regions for which
rCBF was correlated with reaction time during task
performance. Both positive correlations (representing
an increase in rCBF as a function of increasing reaction
time) and negative correlations (representing a decrease in rCBF as a function of increasing reaction
time) were examined.
RESULTS
Visual search performance
For each subject, the first and second sets of three
PET studies contained one instance of the three task
conditions. Preliminary analyses of mean reaction time
for correct responses and mean error rate (percentage
of incorrect responses) were conducted, in which the
first vs. second set of three studies was included as a
variable, in addition to age group and task condition.
For both the reaction time and error rate data, however, the first/second set variable did not yield a
significant main effect or interaction in an analysis of
394
r
r
Aging, Attention, and rCBF Activation r
Figure 2.
Mean reaction time of correct responses as a function of age group and task condition. See Figure 1
for a description of task conditions. Mean reaction time in the Divided condition is presented as a
function of the display position of the target.
variance (ANOVA). In subsequent analyses the data were
averaged over the two instances of each task condition.
Reaction time
The reaction time data for each age group are
presented in Figure 2. An ANOVA of these data,
including age group as a between-subjects variable
and task condition (Passive, Central, or Divided) as a
within-subjects variable, yielded significant main effects of age group (F(1,22) 5 4.89, P , .05), and condition (F(2,44) 5 442.75, P , .0001). The age group 3
condition interaction was also significant (F(2,44) 5
15.53, P , .0001). The age group main effect represents
a 44-msec higher reaction time for older adults (570
msec) than for young adults (526 msec). The condition
main effect represents a 245-msec increase from the
Passive condition (275 msec) to the Central condition
r
(520 msec), and a 329-msec increase from the Central
condition to the Divided condition (849 msec).
To examine the age group 3 condition interaction
further, two ANOVAs were conducted in which the
condition variable contained two levels, either Passive
and Central, or Central and Divided. The ANOVA of
the Passive and Central conditions yielded a significant main effect of condition (F(1,22) 5 585.26,
P , .0001), representing the 245-msec increase from
Passive to Central. The age group 3 condition interaction was also significant (F(1,22) 5 6.63, P , .05). The
increase in reaction time from Passive to Central was
greater for older adults (271 msec) than for young
adults (219 msec), and this increase was significant for
each age group (F[1,22] . 200.0, P , .0001, in each
case). The age group difference, however, was not
significant for either the Passive or Central condition
considered individually.
395
r
r
Madden et al. r
The ANOVA of the Central and Divided conditions
yielded significant main effects both for age group
(F(1,22) 5 12.10, P , .01), and for condition (F(1,22) 5
239.94, P , .0001). The age effect represents an 87-msec
higher reaction time for older adults (728 msec) than
for young adults (641 msec), and the condition effect
represents the 329-msec increase from Central to Divided noted in the analysis of all three conditions. The
age group 3 condition interaction was also significant
(F(1,22) 5 13.35, P , .001). The age-related increase in
reaction time was significant in the Divided condition
(165 msec) (F(1,22) 5 17.28, P , .001), but not in the
Central condition (10 msec). The increase associated
with the Divided condition, relative to the Central
condition, was significant for both age groups
(F(1,22) . 70.0, P , .0001, in each case), but was greater
in magnitude for older adults (407 msec) than for
young adults (252 msec).
As a result of the significant difference between age
groups in reaction time, the age difference in the
absolute magnitude of the difference between task
conditions may have reflected the same degree of
relative change for each group. To examine this possibility, the proportional increase in reaction time associated with the Divided condition, defined as (Divided 2
Central)/Central, was obtained for each subject, and
this value was also significantly greater for older
adults (.78) than for young adults (.51) (F(1,22) 5 7.63,
P , .01).
Target location varied in the Divided condition, and
previous research indicates that, in some tasks, the
increase in letter-identification time associated with
increasing target distance from fixation is greater for
older adults than for young adults [Scialfa et al., 1987].
Thus, even in the presence of comparable attentional
abilities, age differences in the Divided condition may
occur, as the result of an age-related change in the
effect of target eccentricity. In the Divided condition,
the target letters occurred an equal number of times at
each of the display positions. An analysis of the
Divided condition reaction-time data was conducted
in which target location (the nine display positions)
was included as a variable. In addition to the 165-msec
age difference, there was a significant effect of target
location (F(8,176) 5 17.73, P , .0001). As evident in
Figure 2, the location effect was due primarily to
higher reaction times for targets in the lower row of the
display (locations 7–9 in Fig. 2). The age group 3
location interaction was not significant, indicating that
the variation associated with target location was comparable for young and older adults.
As an additional control for the effects of target
location, we conducted an analysis of the Central and
r
Divided conditions in which the Divided condition
reaction times only included the data for targets
located in the central display position (location 5 in
Fig. 2). The data in this analysis consequently represent
the same target location for the two task conditions.
This ANOVA yielded a pattern of results comparable
to that obtained when all locations were included:
there were significant effects for age group
(F(1,22) 5 7.97, P , .01), condition (F(1,22) 5 148.96,
P , .0001), and the age group 3 condition interaction
(F(1,22) 5 8.05, P , .01). The increase in reaction time
associated with the Divided condition was greater in
magnitude for older adults (337 msec) than for young
adults (210 msec), and the age difference in the Divided condition was significant (F(1,22) 5 10.63,
P , .01). In this analysis of the data for display location
5, the proportional increase in reaction time from the
Central condition to the Divided condition was also
greater for older adults (.65) than for young adults (.43)
(F(1,22) 5 4.79, P , .05).
Error rate
Because responses in the Passive condition did not
involve target-letter identification, errors were calculated only for the Central and Divided conditions. The
mean percentage errors are presented for each age
group in Figure 3. An ANOVA of the error rates in the
Central and Divided conditions yielded significant
main effects of age group (F(1,22) 5 14.96, P , .001),
and condition (F(1,22) 5 25.66, P , .0001). The age
group 3 condition interaction was also significant
(F(1,22) 5 10.89, P , .01). The age group effect represents a higher error rate for older adults (9.84%) than
for young adults (3.24%), and the condition effect
represents a higher error rate for the Divided condition
(10.89%) than for the Central condition (2.19%). The
increase in error rate associated with the Divided
condition was greater for older adults (14.38%) than
for young adults (3.04%). The mean error rate in the
Central condition was not significantly different for
young adults (1.72%) and older adults (2.65%). In the
Divided condition, the error rate was significantly
higher for older adults (17.03%) than for young adults
(4.76%) (F(1,22) 5 13.59, P , .001).
An analysis of the Divided condition error rates was
also conducted in which target location was included
as a variable. In addition to the overall higher error
rate exhibited by the older adults, there was a significant main effect of location, (F(8,176) 5 5.82, P , .0001).
The trend of the location effect was less systematic in
396
r
r
Aging, Attention, and rCBF Activation r
Figure 3.
Mean percentage of incorrect responses (error rate) as a function of age group and task condition.
See Figure 1 for a description of task conditions. Mean error rate in the Divided condition is
presented as a function of the display position of the target.
the error data than in the reaction time data, although
as in the RT analyses, the results suggest increased
identification difficulty for targets in the lower row of
the display (locations 7–9 in Fig. 3). The age group 3
location interaction was not significant.
Errors in the Central and Divided conditions were
also examined in an analysis of the error rates that only
included data for targets in the center of the display
(location 5 in Fig. 3). The results were similar to those
of the analysis in which all locations were included:
There were significant main effects for age group
(F(1,22) 5 7.79, P , .01), and condition (F(1,22) 5 9.22,
P , .01), representing an age-related increase in error
rate of 5.44%, and a 6.03% higher error rate in the
Divided condition than in the Central condition. The
age group 3 condition interaction was also significant
(F(1,22) 5 5.16, P , .05), with the increase in error rate
associated with the Divided condition, relative to the
Central condition, being higher for older adults
r
(10.54%) than for young adults (1.52%). For the centrally located targets in the Divided condition, older
adults exhibited a significantly higher error rate
(13.19%) than did young adults (3.24%) (F(1,22) 5 6.68,
P , .05).
Cortical volume
Cortical volume data are presented in Table I. The
cerebellar data from one young subject were eliminated because of an artifact in the MR image. The
cortical gray matter data in Table I represent the
complete sample of 12 subjects in each age group.
Separate univariate ANOVAs were performed on the
cortical gray matter and cerebellar volumes, using age
group as a between-subjects variable and left vs. right
hemisphere as a within-subjects variable. No significant variation in the volumetric data was observed as a
function of either age group or hemisphere.
397
r
r
Madden et al. r
TABLE I. Volume of cortical gray matter (in cm3) as a
function of age group and cerebral hemisphere
Divided minus Passive
Hemisphere
Left
Young adults
Cortical gray
Cerebellum
Older adults
Cortical gray
Cerebellum
Right
Mean
SD
Mean
SD
353.96
65.03
35.51
7.08
361.49
65.52
32.22
7.55
367.27
66.19
42.47
10.58
365.61
65.79
37.72
10.0
Cerebral blood flow
The pattern of rCBF activation in the SPM5Z6 map is
presented in Figure 4. Local maxima of the areas of
increased rCBF are presented in Table II, and the local
maxima of the areas of decreased rCBF (obtained by
reversing the linear contrast for each subtraction) are
presented in Table III. A regional change was considered significant statistically, and included in the Tables,
if the corrected probability level of the peak height of
the local maximum was less than .05 and if the spatial
extent of the region was at least 50 voxels. The
simultaneous linear contrasts (i.e., age group comparisons) used an uncorrected probability level for Z, but
because each contrast was thresholded independently
at P 5 .01, the probability level for the age group
comparisons was P , .0001. The Tables also list the
coordinates of the local maxima in the standard stereotaxic space of Talairach and Tournoux [1988], the
Brodmann’s area (BA) designation, and the gyral
location. To provide information on regional changes
that fell short of the adopted significance level, voxels
in Figure 4 were thresholded at P , .001 uncorrected,
which corresponds to P , .65 corrected.
Increases in rCBF between task conditions
Central minus Passive
This subtraction represents processes involved in
identifying a single target letter in a predefined display
location and selecting a response. For the young
adults, rCBF activation in the anterior cingulate of the
left hemisphere was significant; no activation was
significant for the older adults. The age group comparison indicated that the anterior cingulate activation was
significantly greater for young adults than for older
adults.
r
This subtraction represents processes involved in
attending to multiple display positions, identifying the
target letter, and selecting a response. For the young
subjects, significant activation was present in the left
anterior cingulate and prefrontal cortex, and in the
occipitotemporal pathway (BA 18) bilaterally. In addition, local maxima were present for the young adults in
the left cerebellum and right thalamus. As did the
young adults, the older adults exhibited activation in
the anterior cingulate and prefrontal regions of the left
hemisphere. The older adults also exhibited activation
in BA 18 and the cerebellum, although for older adults
the BA 18 activation was located in the left hemisphere
rather than bilaterally as for young adults, and the
cerebellar activation for older adults was located on
the right rather than on the left as for young adults. In
contrast to the young adults, the older adults exhibited
activation in the superior parietal lobe bilaterally. The
comparison of the two age groups demonstrated that
activation in the fusiform gyri bilaterally (BA 18) of
both hemispheres was significantly greater for young
adults than for older adults, whereas activation in the
middle frontal gyrus of the left hemisphere was significantly greater for older adults than for young adults.
Divided minus Central
This subtraction also represents processes involved
in attending to multiple display positions, but subtracting the Central condition removes the process of
Figure 4.
Areas of increased rCBF between task conditions. See Figure 1 for
a description of task conditions. The Central minus Passive
subtraction represents the processes involved in identifying a
target letter in a known display location and selecting a response.
The Divided minus Passive subtraction represents the processes in
attending to multiple display positions, identifying the target letter,
and selecting a response. The Divided minus Central subtraction
represents the processes involved in dividing attention among
multiple display positions. The results are presented as SPM5Z6
voxel maps in the sagittal, coronal, and transverse planes, using the
stereotaxic space of Talairach and Tournoux [1988]. The images
are presented in standard neurologic orientation, so that the left
half of the coronal image and the upper half of the transverse image
correspond to the subject’s left hemisphere. The rCBF increases
for each age group are presented in the gray-scale values. Within
each age group, those voxels that differed significantly from those
of the other age group are presented in the color-scale values.
Cortical areas of significant rCBF increase are noted in Table II.
398
r
r
Aging, Attention, and rCBF Activation r
Figure 4.
r
399
r
r
Madden et al. r
TABLE II. Regions of rCBF increase as a function of age group and task condition
subtraction*
Region
size (k) P(nmax . k)
Z
P(Zmax . u)
x
y
z
BA
Location
Central minus Passive, rCBF increase for young adults
2,091
.003
5.01
4.83
.001
.003
212
210
26
10
28
44
32
32
Anterior cingulate
Anterior cingulate
32
32
10
Anterior cingulate
Anterior cingulate
Superior frontal gyrus
Central minus Passive, rCBF increase for older adults
No significant voxels
Central minus Passive, rCBF increase with young . older
276
4.58
4.58
2.92
.0001
.0001
.0001
210
212
218
8
28
36
44
24
16
Central minus Passive, rCBF increase with older . young
No significant voxels
Divided minus Passive, rCBF increase for young adults
1,335
.019
6.01
4.23
.0001
.032
28
242
22
24
3,939
.0001
733
.114
5.49
4.83
4.79
4.13
.0001
.003
.003
.045
24
242
210
8
296
274
280
218
32 32 Anterior cingulate
32 6/44 Sulcus between precentral
and inferior frontal gyri
212 18 Fusiform gyrus
216 18 Inferior occipital gyrus
224
Cerebellum
0
Thalamus
Divided minus Passive, rCBF increase for older adults
3,066
.0001
3,753
.0001
517
.231
6.08
5.22
4.97
6.03
6.01
4.41
5.52
.0001
.0001
.001
.0001
.0001
.016
.0001
26
246
210
224
2
232
22
10
48
2
32
28
24
264
40
274 228
288
4
266
40
6
6
32
7
18
7
Superior frontal gyrus
Precentral gyrus
Anterior cingulate
Superior parietal lobe
Cerebellum
Inferior occipital gyrus
Superior parietal lobe
18
18
Fusiform gyrus
Fusiform gyrus
Divided minus Passive, rCBF increase with young . older
98
67
4.16
3.73
.0001
.0001
24 298 216
218 298 216
Divided minus Passive, rCBF increase with older . young
55
3.09
.0001
250
16
32
9
Middle frontal gyrus
Divided minus Central, rCBF increase for young adults
3,196
.0001
4.83
4.80
4.52
.003
.003
.010
240 274 216
230 290 212
22 290 212
19
18
18
Fusiform gyrus
Inferior occipital gyrus
Lingual gyrus
7
6
6
32
Superior parietal lobe
Superior frontal gyrus
Middle frontal gyrus
Anterior cingulate
Cerebellum
Superior parietal lobe
Divided minus Central, rCBF increase for older adults
748
3,747
.109
.0001
443
521
.295
.228
5.15
4.94
4.70
4.47
4.91
4.86
.001
.002
.005
.012
.002
.002
224 262
40
26
12
48
224 22
52
210
18
32
4 278 228
22 264
40
r
400
r
7
r
Aging, Attention, and rCBF Activation r
TABLE II. (continued)
Region
size (k) P(nmax . k)
Z
P(Zmax . u)
x
y
z
BA
Location
18
18
18
Inferior occipital gyrus
Lingual gyrus
Lingual gyrus
32
32
6
9
Anterior cingulate
Anterior cingulate
Superior frontal gyrus
Middle frontal gyrus
Divided minus Central, rCBF increase with young . older
85
176
4.00
3.57
3.21
230 290 28
20 290 24
22 286 216
.0001
.0001
.0001
Divided minus Central, rCBF increase with older . young
650
57
3.53
3.33
2.95
3.44
216
216
222
28
.0001
.0001
.0001
.0001
12
24
28
30
44
28
48
24
*See Methods for a description of task conditions. Region size, number of voxels (k); P(nmax . k),
probability that a region of k voxels or greater would have occurred by chance; Z, t value of local
maximum of activation scaled to a unit normal distribution; P(Zmax . u), probability that the observed
peak height of local maximum of activation is greater than would be expected by chance; x, y, z,
coordinates in mm in the standard stereotaxic space of Talairach and Tournoux [1988]; x, right/left
hemisphere, negative indicates left hemisphere; y, anterior/posterior coordinate, negative indicates
posterior to the zero point (anterior commissure); z, superior/inferior coordinate, negative indicates
inferior to the AC-PC line; BA, Brodmann’s area. P levels of ,.0001 have been rounded to .0001.
identifying and responding to the target letter. For the
young subjects, the only significant activation was associated with the occipitotemporal pathway (BAs 18 and 19).
Local maxima for the young adults were located in the
fusiform and inferior occipital gyri of the left hemisphere,
and in the lingual gyrus of the right hemisphere. For the
older adults, the activation in both the left and right
superior parietal lobes was significant, as was activation in
the anterior cingulate and prefrontal cortex of the left
hemisphere, and right cerebellar hemisphere. Comparison
of the two age groups indicated that activation in BA 18
bilaterally was greater for young adults than for older
adults, whereas relatively greater activation for older
adults was present in the anterior cingulate and superior
frontal gyrus of the left hemisphere, and in the middle
frontal gyrus of the right hemisphere.
Decreases in rCBF between task conditions
Central minus Passive
For the young adults, the only significant reduction
in rCBF occurred in the middle frontal gyrus of the right
hemisphere; no reduction was significant for the older
adults. Comparison of the age groups indicated that the
young adults’ rCBF reduction was greater than that of the
older adults.
Divided minus Passive
Significant decreases in rCBF for the young adults
were located in the medial and inferior temporal
r
regions bilaterally, at the border of the medial and
inferior frontal gyri of the right hemisphere, and in the
left fusiform gyrus. Older adults exhibited rCBF decreases in the medial temporal cortex bilaterally, and in
the superior frontal and superior temporal gyri of the
right hemisphere. Comparison of the young and older
adults indicated that the pattern of rCBF decrease did
not vary significantly as a function of age group.
Divided minus Central
The young adults exhibited rCBF reduction in the
middle frontal gyrus of the left hemisphere and at the
border of the inferior parietal lobe and angular gyrus
of the right hemisphere. Reductions in the anterior
cingulate (at the midline) and left insula were significant for the older adults. In the comparison of the two
age groups, the rCBF reduction in the left middle
frontal gyrus was greater for young adults than for
older adults, whereas the reduction in the left insula
was relatively greater for the older adults.
Correlation between reaction time and rCBF
The correlation between visual search reaction time
and the pattern of rCBF activation was examined in
SPM analyses of the rCBF data in which mean reaction
time was included as a covariate of interest. In view of
the differences between the age groups that were
obtained in both the rCBF data and visual search
401
r
r
Madden et al. r
TABLE III. Regions of rCBF decrease as a function of age group and task condition subtraction*
Region
size (k)
P(nmax . k)
Z
P(Zmax . u)
x
y
z
BA
Location
Central minus Passive, rCBF decrease for young adults
3,176
.0001
5.39
4.50
4.44
.0001
.011
.014
36
48
52
14
14
20
48
36
20
8
8
9
Middle frontal gyrus
Middle frontal gyrus
Middle frontal gyrus
44
20
8
9
Middle frontal gyrus
Middle frontal gyrus
Central minus Passive, rCBF decrease for older adults
No significant voxels
Central minus Passive, rCBF decrease with young . older
226
4.69
3.94
.0001
.0001
40
52
16
20
Central minus Passive, rCBF decrease with older . young
No significant voxels
Divided minus Passive, rCBF decrease for young adults
4,459
.0001
5.73
5.19
5.11
.0001
.001
.001
1,106
.037
212
.624
4.91
4.44
4.17
.002
.014
.039
50
50
48
268
246
244
20
28
220
39
21/37
37
262
244
46
242
220
34
24
224
12
21
20
46/45
8
4
28
4
12
37
10
21
21
22
Medial temporal gyrus
Superior frontal gyrus
Medial temporal gyrus
Medial temporal gyrus
Superior temporal gyrus
40
20
8
39
Middle frontal gyrus
Inferior parietal lobe/angular gyrus
4
20
24
Anterior cingulate
Insula
44
8
Medial temporal gyrus
Sulcus between inferior and medial temporal gyri
Sulcus between fusiform and inferior temporal
gyri
Medial temporal gyrus
Fusiform gyrus
Sulcus between middle and inferior frontal gyri
Divided minus Passive, rCBF decrease for older adults
1,812
1,452
3,137
.006
.014
.0001
4.80
4.60
4.57
4.55
4.13
.003
.007
.008
.009
.045
248
4
44
60
56
260
60
224
242
232
Divided minus Passive, rCBF decrease with young . older
No significant voxels
Divided minus Passive, rCBF decrease with older . young
No significant voxels
Divided minus Central, rCBF decrease for young adults
371
348
.375
.405
4.75
4.22
.004
.033
226
48
28
270
Divided minus Central, rCBF decrease for older adults
860
1,558
.077
.011
4.24
4.16
.031
.040
0
232
24
28
Divided minus Central, rCBF decrease with young . older
62
3.98
.0001
226
28
Middle frontal gyrus
Divided minus Central, rCBF decrease with older . young
59
2.74
2.63
.0001
.0001
238
228
214
24
12
16
Insula
Insula
*See Methods for a description of task conditions. See footnote to Table II for a description of column headings.
r
402
r
r
Aging, Attention, and rCBF Activation r
TABLE IV. Regions of positive correlation between reaction time and rCBF as a
function of age group and task condition*
Region
size (k) P(nmax . k)
Z
P(Zmax . u)
x
y
z
BA
36
20
28
7
33
19
Location
Young adults: Central and Passive
No significant voxels
Older adults: Central and Passive
No significant voxels
Young adults: Divided and Passive
9,266
.0001
5.62
5.57
4.89
.0001
.0001
.002
224 268
28
20
240 280
Superior parietal lobe
Anterior cingulate
Inferior occipital gyrus
Older adults: Divided and Passive
7,803
1,736
733
202
.0001
.005
.105
.651
6.60
.0001
250
12
28
8/9
6.60
6.51
.0001
.0001
26
246
12
2
48
32
6
6
5.32
4.96
4.13
.0001
.002
.050
210 280 224
28 258
40
252 264 220
7
37
Sulcus between middle
and inferior frontal
gyri
Superior frontal gyrus
Sulcus between precentral and inferior frontal
gyrus
Cerebellum
Superior parietal lobe
Inferior temporal gyrus
Young adults: Divided and Central
600
1,233
.196
.031
4.91
4.21
.002
.034
226 270
236 276
4.15
.043
232 290
36
7 Superior parietal lobe
28 18/19 Sulcus between medial
and inferior occipital
lobes
4 18 Sulcus between medial
and inferior occipital
lobes
Older adults: Divided and Central
784
7,477
749
.061
.0001
5.69
4.85
.0001
.003
226 266
226 24
40
48
7
6
.069
4.47
4.38
4.39
.017
.025
.024
242
4
216
10
22 266
28
40
40
6
6
7
Superior parietal lobe
Sulcus between middle
and superior frontal
gyri
Precentral gyrus
Superior frontal gyrus
Superior parietal lobe
*See Methods for a description of task conditions. See footnote to Table II for a description of column
headings.
performance data, the reaction time-rCBF correlation
analyses were conducted within each age group. Local
maxima of the regions of significant correlation (at
P , .05, corrected) are listed in Table IV (positive
correlations) and Table V (negative correlations).
r
Positive correlations between reaction time and rCBF
The analysis of the young adults’ Central and
Passive conditions did not yield any positive correlation between reaction time and rCBF. (The left anterior
403
r
r
Madden et al. r
TABLE V. Regions of negative correlation between reaction time and rCBF as a
function of age group and task condition*
Region
size (k) P(nmax . k)
Z
P(Zmax . u)
x
y
z
BA
Location
12
48
42
8
Superior temporal gyrus
Middle frontal gyrus
4
48 24
50 244 216
10
37
28
22
48
246 216 224
8
20
Superior frontal gyrus
Sulcus between fusiform
and inferior temporal
gyri
Middle frontal gyrus
Inferior temporal gyrus
38 228 212
248 260
12
60 240
12
0
34
48
20
37
22
8
Medial temporal gyrus
Medial temporal gyrus
Superior temporal gyrus
Superior frontal gyrus
Young adults: Central and Passive
1,543
1,726
.012
.008
4.47
4.23
60 230
38
12
.013
.033
Older adults: Central and Passive
No significant voxels
Young adults: Divided and Passive
8,489
.0001
1,238
.033
5.22
5.00
.0001
.001
4.97
4.40
.001
.016
Older adults: Divided and Passive
10,408
.0001
69
.914
6.36
5.82
5.24
4.82
.0001
.0001
.0001
.003
Young adults: Divided and Central
1,021
.056
4.49
.012
4
54
12 9/10 Superior frontal gyrus
Older adults: Divided and Central
3,272
882
2,876
.0001
.042
.0001
4.99
4.74
4.29
4.25
4.22
.002
.006
.035
.041
.046
236
248
262
22
48
214
12
262
8
240
8
18 212
256
0
37
22
25
37
Insula
Medial temporal gyrus
Superior temporal gyrus
Anterior cingulate
Medial temporal gyrus
*See Methods for a description of task conditions. See footnote to Table II for a description of column
headings.
cingulate activation, which was evident in the young
adults’ Central minus Passive rCBF subtraction, was
marginally significant at the P , .07 level in the
correlational analysis.) The analysis of the young
adults’ Divided and Passive conditions indicated an
extensive region of correlation in the frontal, parietal,
and occipital cortex. The local maxima that were
significant were all located in the left hemisphere: the
superior parietal lobe, anterior cingulate, and inferior
occipital gyrus. Positive correlations were also present
for the young adults’ Divided and Central conditions.
The local maxima in this latter analysis were located in
the left superior parietal lobe and in the occipitotemporal pathway (BAs 18 and 19) of the left hemisphere.
The analysis of the older adults’ Central and Passive
conditions did not yield any regions of positive correlar
tion between reaction time and rCBF. For the older
adults’ Divided and Passive conditions, there was an
extensive region of positive correlation, with several
maxima in the frontal lobe of the left hemisphere.
Additional local maxima were present more posteriorly, in the right superior parietal lobe, left cerebellum,
and left inferior temporal cortex. For the Divided and
Central conditions, the older adults exhibited positive
correlations for the superior parietal lobe bilaterally,
and in BA 6 of the left frontal lobe.
Negative correlations between reaction time and rCBF
In the analysis of the young adults’ Central and
Passive conditions, there was a negative correlation
between reaction time and rCBF for the superior
404
r
r
Aging, Attention, and rCBF Activation r
temporal and middle frontal gyri of the right hemisphere. The young adults’ Divided and Passive conditions also exhibited negative correlations with local
maxima in the frontal lobe of the right hemisphere, and
in the inferior temporal lobe bilaterally. The data for
the young adults’ Divided and Central conditions
yielded a negative correlation for the right superior
frontal gyrus.
No negative correlations between reaction time and
rCBF were evident in the older adults’ Central and
Passive conditions. In the Divided and Passive conditions, there were negative correlations for the older
adults in the medial temporal gyrus bilaterally, in the
right superior temporal gyrus, and in the superior
frontal gyrus at the midline. The older adults’ Divided
and Central conditions yielded negative correlations
for the medial temporal gyrus bilaterally, and for the
insula, anterior cingulate, and superior temporal gyrus
of the left hemisphere.
DISCUSSION
Visual search performance
The visual search performance data confirmed previous findings indicating that age differences are greater
under conditions requiring divided attention than
under those requiring selective attention [Hartley,
1992; Madden and Plude, 1993]. As is apparent in
Figures 2 and 3, the age-related increase in reaction
time and error rate was more pronounced in the
Divided condition than in the Central condition, an
observation supported by the finding that, for both the
reaction time and error rate variables, the age difference was statistically significant in the Divided condition but not in the Central condition. The performance
of both age groups declined when the task required
dividing attention among all the display positions,
relative to when attention could be focused on a single
display position, but this decline was relatively greater
for older adults. This age difference was in addition
not simply an artifact of a proportional age-related
slowing that is constant for all task conditions, because
the age difference remained significant when expressed as the proportional increase in reaction time
for the Divided condition relative to the Central
condition.
In view of the fact that, relative to young adults,
older adults typically require additional eye movements to search complex displays [Scialfa et al., 1994],
and require additional processing time for individual
display characters [Ellis et al., 1996], especially as
retinal eccentricity increases [Scialfa et al., 1987], to
r
what extent can the age-related effects in the search
performance data be attributed to these types of
sensory variables rather than to attention per se?
Although subjects were encouraged to maintain visual
fixation centrally in the Divided condition, eye movements were not recorded, and the 1-sec display duration allowed saccades to occur. Analyses of the target
location effect for the Divided condition, however, led
us to believe that sensory-level variables were not
entirely responsible for the age differences we obtained. The change in reaction time and error rate as a
function of target location represented less efficient
processing of the lowest row in the display, which is
consistent with other investigations of display position
effects in visual search [Chaikin et al., 1962; Previc and
Blume, 1993]. This target location effect did not vary
significantly across age groups in the present experiment, suggesting that any performance changes associated with eye movements to peripheral targets were
comparable for young and older adults. In addition,
the disproportionate increase in the older adults’
reaction times in the Divided condition was evident
when the analysis was restricted to data for the Central
display position. Although age-related changes in the
efficiency of sensory processing undoubtedly contributed to age differences in performance, it is likely that
age-related limitations in the ability to divide attention
are also involved.
Cortical volume
The analysis of the cortical volume data demonstrated that the present young and older adults were
comparable in the volume of cortical gray matter and
cerebellum (Table I). Raz et al. [1997], in contrast,
reported significant age-related reduction in the volume of several cortical regions (especially prefrontal
gray matter) as estimated from MRI. Subjects in the
study of Raz et al. [1997] were screened for the
presence of neurological and other major disease, as
were the present subjects, and thus health status is not
an obvious basis for the discrepancy. However, there
was a substantially greater number of subjects (148) in
the study of Raz et al. [1997] than in the present study
(24), which would provide a higher level of statistical
power for the analysis of Raz et al. [1997]. The
comparable volume of cortical gray matter for the two
age groups in the present study minimizes the potential contribution of partial volume effects to age differences in rCBF.
405
r
r
Madden et al. r
Cerebral blood flow
In the Passive condition, subjects viewed displays
and pressed a response button on each trial, although
no specific decision was required. The Central minus
Passive subtraction in the SPM analyses consequently
represents the rCBF activation associated with identifying a target letter in a selectively attended display
position. Assuming that knowledge of the target’s
location facilitates its identification, the most probable
cortical region of activation for this subtraction would
be the occipitotemporal processing pathway [Corbetta
et al., 1990, 1991; LaBerge, 1995]. The only significant
activation that was obtained for the Central minus
Passive subtraction, however, was the left anterior
cingulate activation exhibited by the young adults.
Although the increase in reaction time from the Passive condition to the Central condition was statistically
significant, and greater than 200 msec, for both age
groups, the letter identification processes required by
the Central condition were apparently not sufficiently
demanding to engage the occipitotemporal pathway. It
is possible that subjects also identified the letters in the
displays presented during the Passive condition, thus
reducing the difference in the degree of letter processing that occurred during the Passive and Central
conditions.
The young adults’ anterior cingulate activation that
was obtained for the Central minus Passive subtraction is difficult to interpret. This cortical region has
been found to be involved in regulating attentional
allocation during task performance and in the selection of motor responses [Bench et al., 1993; Pardo et al.,
1990; Paus et al., 1993]. Because the age difference in
task performance, however, was not significant for
either the Central or Passive conditions considered
separately, we do not believe that this activation is
linked closely to the young adults’ task performance.
In addition, the correlation between the young adults’
reaction time and the anterior cingulate activation was
only marginally significant (P , .07). Activation of the
anterior cingulate has been observed in the absence of
task-specific processing and may represent cognitive
activity unrelated to task performance [McCarthy et
al., 1996; Murtha et al., 1996]. Madden et al. [1996] also
found that activation in the anterior cingulate of the
left hemisphere was greater for young adults than for
older adults, under conditions in which no substantial
age-related change in reaction time was apparent.
The Divided minus Passive subtraction represents a
variety of processes related to performance in this
search task, including dividing attention among multiple display locations (i.e., monitoring the locations
r
for target features), rejecting nontarget letters, shifting
the focus of attention between display positions, identifying the target letter, and selecting a response. The
rCBF activation for the Divided minus Passive subtraction is in agreement with previous PET studies of
visuospatial attention [Corbetta et al., 1990, 1991, 1993,
1995; Nobre et al., 1997], but the two age groups
differed in the pattern of activation, which was relatively occipitotemporal for the young adults and relatively prefrontal for the older adults. Activation in the
superior parietal lobe (BA 7) was present for the older
adults, consistent with previous findings for tasks
requiring shifts of spatial attention [Corbetta et al.,
1993, 1995; Nobre et al., 1997]. The parietal activation
was not significant statistically for the young adults,
but the interaction contrast for this subtraction indicated that the age difference in the magnitude of the
parietal activation was not significant. The interaction
analysis did indicate that activation in the fusiform
gyri bilaterally (BA 18) was greater for the young
adults than for the older adults, whereas a region in the
middle frontal gyrus of the right hemisphere exhibited
significantly more activation for the older adults than
for the young adults.
The Divided minus Central subtraction is particularly relevant to the issue of age differences in attentional processing, because the Central condition involves target identification and response processes.
Although there is no guarantee that identifying and
responding to the target occurs in exactly the same
way in the Divided and Central conditions [Friston et
al., 1996b; Sergent et al., 1992], the remaining cognitive
processes isolated by this subtraction should be related
closely to the attentional demands of processing multiple display positions. As in the Divided minus
Passive subtraction, the comparison of the two age
groups for the Divided minus Central subtraction
indicated that the young adults exhibited greater rCBF
activation than the older adults in the occipitotemporal
pathway bilaterally (BA 18), whereas the older adults
exhibited greater activation than the young adults in
the prefrontal regions in both hemispheres. In fact, the
age difference in the prefrontal regions was more
pronounced in the Divided minus Central subtraction
than in the Divided minus Passive subtraction (Table
II), which may be related to the decrease in prefrontal
rCBF exhibited by the young adults in the Divided
minus Central subtraction (Table III). As was the case
for the Divided minus Passive subtraction, the Divided
minus Central subtraction yielded significant activation in the superior parietal lobe (BA 7) for the older
adults but not for the young adults. The age difference
in parietal activation was not significant statistically,
406
r
r
Aging, Attention, and rCBF Activation r
however, because the young adults did exhibit a
low-level activation in this cortical region (Fig. 4).
The age difference in the pattern of rCBF activation
that occurred under conditions of dividing attention is
consistent with the age-related decline in occipitotemporal activation reported by Madden et al. [1996] for
visual word-identification performance. The present
data also support the observation of Grady et al. [1992,
1994] that older adults’ performance in visual discrimination tasks involves the recruitment of prefrontal
regions outside of the visual processing pathways. The
cognitive operations involved in this recruitment cannot be specified completely from the present data. One
possibility is that young adults are able to perform the
search task on the basis of letter-identification processes mediated primarily by the occipitotemporal
pathway. For older adults, a decline in ability to
process information from multiple display positions
simultaneously leads to reliance on additional input
from higher-order control processes, such as rehearsal
of the target letters and rechecking decisions prior to
responding. Regardless of the particular cognitive
processes involved, the present findings, as well as
those of Grady et al. [1992, 1994], indicate that aging is
associated with a change in the pattern in cortical
activation rather than with an overall decline in the
magnitude of activation.
It is also important to note that the specific pattern of
age effects in the present study differs in some respects
from that of Grady et al. [1992, 1994]. These authors
reported that the relatively greater activation in the
occipital cortex exhibited by young adults was primarily in medial, prestriate regions that may represent
feature-level processing. The young adults’ activation
in the present study was located in more lateral regions
of the association cortex that are presumed to mediate
higher-level visual identification processes. The older
adults’ prefrontal activation in the experiments of
Grady et al. [1992, 1994] primarily involved BAs 8, 10,
and 46, which led these authors to suggest that the
activation may have represented working memory
processes. The older adults’ prefrontal activation in the
present experiment was located more medially (BAs
32, 6, and 9), which may represent a more general level
of task control and response monitoring [Bench et al.,
1993; Pardo et al., 1990; Paus et al., 1993]. In addition,
the two age groups in the experiments of Grady et al.
[1992, 1994] performed at a comparable level of accuracy, whereas older adults performed the present
version of the letter search at a lower level of accuracy
than did the young adults (Fig. 3). Differences in the
pattern of age effects in rCBF may be to some extent a
result of these differences in performance level.
r
The reductions in rCBF associated with each subtraction (Table III) appear to represent the deactivation of
cortical regions, perhaps by an inhibitory process,
during performance of the search task. For both the
Divided minus Passive and Divided minus Central
subtractions, widespread decreases in rCBF throughout the temporal lobe were evident, which would be
expected if subjects were inhibiting auditory/verbal
processing while performing this visual task. The
pattern of decreases in rCBF for the subtractions
involving the Divided condition were in general similar for the two age groups, although a decrease in a left
prefrontal region was relatively greater for the young
adults, and a decrease in the left insular cortex was
relatively greater for the older adults.
Analyses of the correlation between reaction time
and rCBF (Tables IV and V) are important because they
establish the degree of relationship between the taskrelated changes in rCBF and subjects’ actual performance. In the present data, for example, the left
anterior cingulate exhibited by young adults for the
Central minus Passive subtraction did not reach the
level of statistical significance in the correlational
analysis that it did in the rCBF subtraction, suggesting
a relatively weak relationship between this activation
and search reaction time. In contrast, rCBF activation
in the superior parietal cortex (BA 7) of the left
hemisphere was correlated positively with the young
adults’ reaction time in the Divided and Central
conditions, but this cortical region did not exhibit a
significant level of activation in the Divided minus
Central subtraction. In general, however, the correlational analyses were consistent with the rCBF subtractions. The positive correlations between reaction time
and rCBF support the conclusion that the age difference in performance under divided attention conditions represents relatively greater processing in the
occipitotemporal pathway (BA 18) for young adults,
and relatively greater processing in prefrontal cortical
regions for older adults. Although most of the local
maxima for the positive correlations were located in
the left hemisphere, some of the activated regions were
spatially extensive and included homologous regions
of the right hemisphere. The negative correlations
between reaction time and rCBF were consistent with
the pattern of decreases in rCBF, obtained in the
subtraction analyses, which extended primarily
throughout the temporal lobes.
CONCLUSIONS
The reaction time and error rate data from the
present visual search task confirm previous reports of
407
r
r
Madden et al. r
a more pronounced age-related performance decline
under divided attention conditions than under selective attention conditions. The rCBF analyses indicated
that, relative to a passive viewing condition, identifying a single target letter in a known display position is
not associated with task-related rCBF activation for
either young adults or older adults. When it was
necessary to divide attention among multiple display
positions, however, rCBF activation was evident in
cortical regions emphasized by current models of the
functional neuroanatomy of visual attention, especially the occipitotemporal processing pathway, prefrontal cortex, and superior parietal cortex. In addition,
significant age differences in rCBF activation were
observed under divided attention conditions, reflecting relatively greater processing in the occipitotemporal pathway for young adults, and relatively greater
involvement of prefrontal regions for older adults.
These patterns of rCBF activation were related to
subjects’ reaction time and suggest that young adults’
performance in this task is based primarily on letter
identification processes, whereas older adults rely on
additional forms of task control.
ACKNOWLEDGMENTS
This research was supported by research grant R01
AG11622 from the National Institute on Aging. We are
grateful to Richard Frackowiak and Karl Friston for
providing the Statistical Parametric Mapping software, and to Sharon Hamblen, Mary Traynor, Collin
McKinney, Laura Denny, Amy Harris, and Corri Oenbring for technical assistance.
REFERENCES
Beck AT (1978): The Beck Depression Inventory. New York: Psychological Corporation.
Bench CJ, Frith CD, Grasby PM, Friston KJ, Paulesu E, Frackowiak
RSJ, Dolan RJ (1993): Investigations of the functional anatomy of
attention using the Stroop test. Neuropsychologia 31:907–922.
Chaikin JD, Corbin HH, Volkmann J (1962): Mapping a field of
short-time visual search. Science 138:1327–1328.
Christensen KJ, Moye J, Armson RR, Kern TM (1992): Health
screening and random recruitment for cognitive aging research.
Psychol Aging 7:204–208.
Corbetta M, Miezin FM, Dobmeyer S, Shulman GL, Petersen SE
(1990): Attentional modulation of neural processing of shape,
color, and velocity in humans. Science 248:1556–1559.
Corbetta M, Miezin FM, Dobmeyer S, Shulman GL, Petersen SE
(1991): Selective and divided attention during visual discrimination of shape, color, and speed: Functional anatomy by positron
emission tomography. J Neurosci 11:2383–2402.
Corbetta M, Miezin FM, Shulman GL, Petersen SE (1993): A PET
study of visuospatial attention. J Neurosci 13:1202–1228.
r
Corbetta M, Shulman GL, Miezin FM, Petersen SE (1995): Superior
parietal cortex activation during spatial attention shifts and
visual feature conjunction. Science 270:802–805.
DeGrado TR, Turkington TG, Williams JJ, Stearns CW, Hoffman JM,
Coleman RE (1994): Performance characteristics of a whole-body
PET scanner. J Nucl Med 35:1398–1406.
Ellis RD, Goldberg JH, Detweiler MC (1996): Predicting age-related
differences in visual information processing using a two-stage
queuing model. J Gerontol 51:155–165.
Folstein MF, Folstein SE, McHugh PR (1975): Mini-mental state: A
practical method for grading the cognitive state of patients for
the clinician. J Psychiatr Res 12:189–198.
Friston KJ, Frith CD, Liddle PF, Dolan RJ, Lammertsma AA,
Frackowiak RSJ (1990): The relationship between global and local
changes in PET scans. J Cereb Blood Flow Metab 10:458–466.
Friston KJ, Holmes AP, Worsley KJ, Poline JP, Frith CD, Frackowiak
RSJ (1995): Statistical parametric maps in functional imaging: A
general linear approach. Hum Brain Mapping 2:189–210.
Friston KJ, Ashburner J, Poline JB, Frith CD, Heather JD, Frackowiak
RSJ (1996a): Spatial realignment and normalization of images.
Hum Brain Mapping 3:165–189.
Friston KJ, Price CJ, Fletcher P, Moore C, Frackowiak RSJ, Dolan RJ
(1996b): The trouble with cognitive subtraction. Neuroimage
4:97–104.
Grady CL, Haxby JV (1995): Functional neuroimaging of the aging
brain. In: Boller F, Grafman J (eds): Handbook of Neuropsychology, Volume 10. Amsterdam: Elsevier, pp 467–481.
Grady CL, Haxby JV, Horwitz B, Schapiro MB, Rapoport SI,
Ungerleider LG, Mishkin M, Carson RE, Herscovitch P (1992):
Dissociation of object and spatial vision in human extrastriate
cortex: Age-related changes in activation of regional cerebral
blood flow measured with [15O] water and positron emission
tomography. J Cogn Neurosci 4:23–34.
Grady CL, Maisog JM, Horwitz B, Ungerleider LG, Mentis MJ,
Salerno JA, Pietrini P, Wagner E, Haxby JV (1994): Age-related
changes in cortical blood flow activation during visual processing of faces and location. J Neurosci 14:1450–1462.
Grady CL, Horwitz B, Pietrini P, Mentis MJ, Ungerleider LG,
Rapoport SI, Haxby JV (1996): Effect of task difficulty on cerebral
blood flow during perceptual matching of faces. Hum Brain
Mapping 4:227–239.
Greenwood PM, Parasuraman R, Haxby JV (1993): Changes in
visuospatial attention over the adult lifespan. Neuropsychologia
31:471–485.
Hahn S, Kramer AF (1995): Attentional flexibility and aging: You
don’t need to be 20 years of age to split the beam. Psychol Aging
10:597–609.
Hartley AA (1992): Attention. In: Craik FIM, Salthouse TA (eds): The
Handbook of Aging and Cognition. Hillsdale, NJ: Erlbaum, pp
3–50.
Hartley AA, Kieley JM, Slabach EH (1990): Age differences and
similarities in the effects of cues and prompts. J Exp Psychol
[Hum Percept] 16:523–537.
Horwitz B, Duara R, Rapoport SI (1986): Age differences in intercorrelations between regional cerebral metabolic rates for glucose.
Ann Neurol 19:60–67.
LaBerge D (1995): Computational and anatomical models of selective attention in object identification. In: Gazzaniga MS (ed): The
Cognitive Neurosciences. Cambridge, MA: MIT Press, pp 649–
663.
Madden DJ (1986): Adult age differences in the attentional capacity
demands of visual search. Cogn Dev 1:335–363.
408
r
r
Aging, Attention, and rCBF Activation r
Madden DJ (1992): Selective attention and visual search: Revision of
an allocation model and application to age differences. J Exp
Psychol [Hum Percept] 18:821–836.
Madden DJ, Hoffman JM (1997): Application of positron emission
tomography to age-related cognitive changes. In: Krishnan KRR,
Doraiswamy PM (eds): Brain Imaging and Clinical Psychiatry.
New York: Marcel Dekker, pp 575–613.
Madden DJ, Plude DJ (1993): Selective preservation of selective
attention. In: Cerella J, Rybash J, Hoyer W, Commons ML (eds):
Adult Information Processing: Limits on Loss. San Diego: Academic Press, pp 273–300.
Madden DJ, Turkington TG, Coleman RE, Provenzale JM, DeGrado
TR, Hoffman JM (1996): Adult age differences in regional cerebral
blood flow during visual word identification: Evidence from
H215O PET. Neuroimage 3:127–142.
Martin AJ, Friston KJ, Colebatch JG, Frackowiak RSJ (1991): Decreases in regional cerebral blood flow with normal aging. J
Cereb Blood Flow Metab 11:684–689.
McCarthy G, Puce A, Constable RT, Krystal JH, Gore JC, GoldmanRakic P (1996): Activation of human prefrontal cortex during
spatial and nonspatial working memory tasks measured by
functional MRI. Cereb Cortex 6:600–611.
McDowd JM, Birren JE (1990): Aging and attentional processes. In:
Birren JE, Schaie KW (eds): Handbook of the Psychology of
Aging, 3rd Edition. San Diego: Academic Press, pp 222–233.
Moeller JR, Ishikawa T, Dhawan V, Spetsieris P, Mandel F, Alexander
GE, Grady C, Petrini P, Eidelberg D (1996): The metabolic
topography of normal aging. J Cereb Blood Flow Metab 16:385–
398.
Murtha S, Chertkow H, Beauregard M, Dixon R, Evans A (1996):
Anticipation causes increased blood flow to the anterior cingulate cortex. Hum Brain Mapping 4:103–112.
Myerson J, Hale S, Wagstaff D, Poon LW, Smith GA (1990): The
information loss model: A mathematical theory of age-related
cognitive slowing. Psychol Rev 97:475–487.
Nobre AC, Sebestyen GN, Gitelman DR, Mesulam MM, Frackowiak
RSJ, Frith CD (1997): Functional localization of the system for
visuospatial attention using positron emission tomography. Brain
120:515–533.
Pardo JV, Pardo PJ, Janer KE, Raichle ME (1990): The anterior
cingulate cortex mediates processing selection in the Stroop
attentional conflict paradigm. Proc Natl Acad Sci USA 87:256–
259.
Paus T, Petrides M, Evans AC, Meyer E (1993): Role of the human
anterior cingulate cortex in the control of oculomotor, manual,
r
and speech responses: A positron emission tomography study. J
Neuropsychol 70:453–469.
Pelizzari CA, Chen GTY, Spebring DR, Weichselbaum RR, Chen CT
(1989): Accurate three-dimensional registration of CT, PET, and/or
MR images of the brain. J Comput Assist Tomogr 13:20–26.
Plude DJ, Hoyer WJ (1986): Age and the selectivity of visual
information processing. Psychol Aging 1:4–10.
Posner MI (1995): Attention in cognitive neuroscience: An overview.
In: Gazzaniga MS (ed): The Cognitive Neurosciences. Cambridge, MA: MIT Press, pp 615–624.
Previc FH, Blume JL (1993): Visual search asymmetries in threedimensional space. Vision Res 33:2697–2704.
Rabbitt P (1965): Age and discrimination between complex stimuli.
In: Welford AT, Birren JE (eds): Behavior, Aging, and the Nervous
System. Springfield, IL: Charles C. Thomas, pp 35–53.
Raz N, Gunning FM, Head D, Dupuis JH, McQuain J, Briggs SD,
Loken WJ, Thornton AE, Acker JD (1997): Selective aging of the
human cerebral cortex observed in vivo: Differential vulnerability
of the prefrontal gray matter. Cereb Cortex 7:268–282.
Salthouse TA (1982): Adult Cognition: An Experimental Psychology
of Human Aging. New York: Springer-Verlag.
Salthouse TA (1996): The processing speed theory of adult age
differences in cognition. Psychol Rev 103:403–428.
Scialfa CT, Kline DW, Lyman BJ (1987): Age differences in target
identification as a function of retinal location and noise level:
Examination of the useful field of view. Psychol Aging 2:14–19.
Scialfa CT, Thomas DM, Joffe KM (1994): Age differences in the
useful field of view: An eye movement analysis. Optom Vision
Sci 71:736–742.
Sergent J, Zuck E, Levesque M, MacDonald B (1992): Positron
emission tomography study of letter and object processing:
Empirical findings and methodological considerations. Cereb
Cortex 2:68–80.
Shiffrin RM (1988): Attention. In: Atkinson RC, Herrnstein RJ,
Lindzey G, Luce RD (eds): Stevens’ Handbook of Experimental
Psychology, 2nd Edition, Volume 2: Learning and Cognition.
New York: Wiley, pp 739–811.
Talairach J, Tournoux P (1988): Co-Planar Stereotaxic Atlas of the
Human Brain. Stuttgart: Thieme.
Turkington TG, Hoffman JM, Jaszcazk RJ, MacFall JR, Harris CC,
Kilts CD, Pelizzari CA, Coleman RE (1995): Accuracy of surface
fit registration for PET and MR brain images using full and
incomplete brain surfaces. J Comput Assist Tomogr 19:117–124.
Wechsler D (1981): Wechsler Adult Intelligence Scale—Revised. New
York: Psychological Corporation.
409
r
Документ
Категория
Без категории
Просмотров
4
Размер файла
396 Кб
Теги
951
1/--страниц
Пожаловаться на содержимое документа