Age differences in prefrontal cortical activity in working memory

Copyright 2001 by the American Psychological Association, Inc. 0882-7974/01/S5.00 DOI: 10.1037//0882-7974.16.3.371

Psychology and Aging 2001, Vol. 16, No. 3, 371-384

Age Differences in Prefrontal Cortical Activity in Working Memory Bart Rypma, Vivek Prabhakaran, John E. Desmond, and John D. E. Gabriel! Stanford University Working memory (WM) declines with advancing age. Brain imaging studies indicate that ventral prefrontal cortex (PFC) is active when information is retained in WM and that dorsal PFC is further activated for retention of large amounts of information. The authors examined the effect of aging on activation in specific PFC regions during WM performance. Six younger and 6 older adults performed a task in which, on each trial, they (a) encoded a 1- or 6-letter memory set, (b) maintained these letters over 5-s, and (c) determined whether or not a probe letter was part of the memory set. Comparisons of activation between the 1- and 6-letter conditions indicated age-equivalent ventral PFC activation. Younger adults showed greater dorsal PFC activation than older adults. Older adults showed greater rostral PFC activation than younger adults. Aging may affect dorsal PFC brain regions that are important for WM executive components.

Similar to older monkeys, older humans perform less well than younger humans on a variety of delayed response tasks, suggesting that similar cortical mechanisms may underlie the age-related performance changes in monkeys and humans. Disproportionate age-related prefrontal decline has been documented in anatomical and physiological studies of human and nonhuman primate brains. Structural magnetic resonance imaging (MRI) studies have found greater age differences in PFC than in other regions (e.g., Raz et al., 1997). Histological studies have found age-related differences in numbers of neurons (e.g., Brizzee, Ordy, & Bartus, 1980), susceptibility to amyloid plaques (e.g., Heilbroner & Kemper, 1990), loss of synapses (Huttenlocher, 1979), dendritic arborizations (Haug & Eggers, 1991), and white matter (e.g., Peters et al., 1996) in the PFC of older brains. Selective age-related declines in dopamine, a critical neurotransmitter for WM functions, have also been observed in PFC (Goldman-Rakic & Brown, 1981; Beal et al., 1991). These age-related physiological changes may be related to the age differences observed in WM performance (Gabrieli, 1996). Advancing age may have selective effects on different components of WM. The phonological loop, or verbal slave system, appears to be minimally affected by aging. For example, digit-span performance, in which participants recall a digit string immediately following presentation, often appears unaffected by healthy aging (Botwinick & Storandt, 1974; Bromley, 1958; Craik, 1968; Drachman & Leavitt, 1972; Friedman, 1974; Gilbert, 1941; Gilbert & Levee, 1971; Kriauciunas, 1968; Taub, 1973). In contrast to digit-span tasks, age-related effects are often seen when a delay is imposed between presentation and recollection. For instance, age-related WM performance declines are more often observed when delay intervals are increased in short-term memory tasks (e.g., Craik, 1977; Nielsen-Bohlman & Knight, 1995; Poon & Fozard, 1980; Smith, 1975). The amount of information that must be held in mind (i.e., memory load) also exacerbates agerelated differences in WM performance. A number of studies examining the effects of varying memory loads on delayed response task performance have shown greater age differences with higher than with lower memory loads (Anders, Fozard, & Lil-

Working memory (WM) can be defined as the cognitive apparatus that allows individuals to temporarily maintain and manipulate information in mind. Evidence from behavioral research indicates declines in WM with advancing age (e.g., Salthouse & Babcock, 1991). WM can be divided into separate components, including slave system buffers for the short-term retention of small amounts of information and a supervisory attentional system or central executive that controls allocation of attention and coordinates information held in the slave system buffers (Baddeley, 1986; Norman & Shallice, 1980). Behavioral studies with monkeys have indicated a critical role for prefrontal cortex (PFC) in mediating WM performance. Anatomical tracing and metabolic imaging studies with monkeys indicate that PFC mediates WM performance (e.g., Funahashi, Bruce, & Goldman-Rakic, 1989; Fuster & Alexander, 1971; Goldman-Rakic & Friedman, 1991; Kubota & Niki, 1971). Singlecell recording of monkey brains has shown persistent activity in dorsolateral PFC cells during the delay period of a detayed-matchto-sample task (Goldman-Rakic & Friedman, 1991). Moreover, monkeys with principal sulcus lesions show location-specific deficits in delayed response performance. Behavioral studies with older monkeys show performance deficits compared with their younger counterparts on similar delayed response WM tasks (e.g., Bachevalier et al., 1991; Bartus, Dean, & Fleming, 1979; Presty et al., 1987).

Bart Rypma and John E. Desmond, Department of Psychology, Stanford University; Vivek Prabhakaran, Program in Neuroscience, Stanford University; John D. E. Gabrieli, Department of Psychology and Program in Neuroscience, Stanford University. This work was supported by Grants AG055701, AG12995, and AG11121 from the National Institute on Aging. We thank Gary H. Glover for assistance in scanning and Margaret Zhao for assistance in data analysis. Correspondence concerning this article should be addressed to Bart Rypma, who is now at Department of Psychology, University of California, Berkeley, 3210 Tolman Hall, Berkeley, California 94720-1650. Electronic mail may be sent to [email protected].

371

372

RYPMA, PRABHAKARAN, DESMOND, AND GABRIELI

lyquist, 1972; Anders & Fozard, 1973; Eriksen, Hamlin, & Daye, 1973; Marsh, 1975). Anders et al. (1972), for instance, examined age-differential performance when participants had to remember various numbers of digits across an unfilled delay interval. They observed increasing age differences in performance with increasing memory load, indicating faster memory retrieval rates in younger than in older participants. However, other studies using similar designs have not observed age differences in memory retrieval rate (e.g., Boaz & Denney, 1993; Kirsner, 1972). The factors that mediate age-differential or age-equivalent WM performance are not yet clearly understood, but the observation of greater age differences as delay or load increases suggests that different components of WM may be differentially susceptible to the deleterious effects of advancing age. It may be that WM mechanisms that allow maintenance of lower memory loads are relatively unaffected by aging. Other memory mechanisms that allow maintenance of higher memory loads may be more available to younger adults than to older adults. The observation that aging has a greater effect on WM under conditions of increased mnemonic demand suggests that executive components may be differentially affected by advancing age. The disproportionate effect of increasing memory load on older peoples' WM performance may reflect a distinction between slave system components for the short-term maintenance of information and an executive component for manipulation of information in the service of optimizing short-term memory performance (cf. Craik & Jennings, 1992). The notion that rehearsal mechanisms mediate low memory-load performance whereas additional memory mechanisms must be recruited for high memory-load performance has been supported in a number of studies of short-term memory capacity (e.g., Baddeley & Hitch, 1974; Glanzer & Razel, 1974; Waugh & Norman, 1965). Waugh and Norman (1965), for instance, administered a cued-recall procedure in which participants were given lists of 16 digits and asked to rehearse only the currently presented digit. The last digit served as a probe and was a repetition of an earlier digit in the list. Participants were required to report the digit immediately prior to the earlier occurrence of the probe digit in the list. The probability of correctly recalling a list item diminished precipitously for those that occurred more than three list items prior to the final digit probe. They concluded that the number of items actually held in short-term storage is severely limited and that additional memory systems were necessary for the successful retention of long lists. On the basis of their data, Waugh and Norman developed an estimation procedure for determining the amount of information actually maintained in short-term storage (see also Murdock, 1967; Tulving & Colotla, 1970). Application of that procedure has consistently yielded estimates of between two and three items of information, although the size of the items (e.g., letters, words, or sentences) may vary considerably (Glanzer & Razel, 1974). Results from behavioral studies support the notion of executive involvement in WM maintenance tasks with above-capacity memory loads. Baddeley and Hitch (1974), for instance, required participants to comprehend prose passages while holding zero, three, or six letters in WM. When they compared comprehension in the zero- and three-letter memory-load conditions, performance did not change (73% and 70% accuracy, respectively). This result suggests that when participants must carry out a complex task while retaining a subcapacity memory load, resources can be

devoted entirely to the more demanding task. Significant decrements in prose comprehension were observed, however, when participants were required to hold six items in WM (60% accuracy). This result suggests that memory loads that approach the capacity of short-term storage may require involvement of executive functions. Neuroimaging studies suggest that PFC may be functionally subdivided in such a way as to support such a two-component WM system. Rypma, Prabhakaran, Desmond, Glover, and Gabrieli (1999) observed prefrontal cortical activation in a WM task in which participants were required to maintain one, three, or six letters for 5 s. When participants were required to maintain three letters in WM, relative to one letter, activation in frontal regions was limited to left ventral PFC corresponding to Brodmann's Area (BA) 44. When participants were required to maintain six letters, relative to one letter, additional activation of bilateral dorsolateral PFC, corresponding to BAs 9 and 46, was observed, suggesting that this brain region may mediate the involvement of additional memory mechanisms necessary for successful maintenance of larger memory loads. Thus, it may be that subcapacity (i.e., two to three items) maintenance of verbal information (which is not generally affected by aging) is supported by left ventrolateral regions of PFC (cf. Awh et al., 1996; Paulesu, Frith, & Frackowiak, 1993). Left and right dorsolateral PFC regions may be engaged selectively in WM to support supracapacity information maintenance, where age-related performance differences are often observed. These results are consistent with other neuroimaging studies indicating different roles for dorsolateral and ventrolateral PFC in WM tasks (e.g., D'Esposito et al., 1995; Petrides, 1996). The locations of the prefrontal activations (i.e., ventrolateral PFC under low memory-demand conditions and dorsolateral PFC under high memory-demand conditions) in Rypma et al.'s (1999) study are consistent with behavioral studies indicating only slave system involvement under low memory-demand conditions but additional executive involvement under high memory-demand conditions (e.g., Baddeley & Hitch, 1974). Further, the dorsolateral PFC activations observed by Rypma et al. are similar to those found in WM studies using more complex tasks (Cohen et al., 1994, 1997; Corbetta, Miezin, Dobmeyer, Shulman, & Petersen, 1991; D'Esposito et al., 1995; Petrides, Alivisatos, Evans, & Meyer, 1993; Petrides, Alivisatos, Meyer, & Evans, 1993; Prabhakaran, Smith, Desmond, Glover, & Gabrieli, 1997; Prabhakaran, Narayanan, Zhao, & Gabrieli, 2000; Prabhakaran, Rypma, & Gabrieli, 2001). It may be that suprathreshold maintenance of information requires cognitive operations similar to those required in these more complex tasks. These operations include monitoring the contents of WM (such as would be required in self-ordered tasks; Petrides, Alivisatos, Evans, & Meyer, 1993; Petrides, Alivisatos, Meyer, & Evans, 1993), updating of the contents of WM (such as would be required in n-back tasks; Cohen et al., 1994, 1997), coordination of slave system processes (such as would be required in dual tasks; e.g., D'Esposito et al., 1995), allocation of attention among multiple stimulus domains (such as would be required in divided attention tasks; e.g., Corbetta et al., 1991), organization and planning of behavior (such as would be required in high-level reasoning tasks, e.g., Baker et al., 1996; Gabrieli, 1996; Prabhakaran et al., 1997; Prabhakaran et al., 2001), and strategy shifting (Deiber et al., 1991; Jenkins, Brooks, Nixon, Frackowiak, & Passingham, 1994). These cognitive operations are

373

AGE DIFFERENCES IN PREFRONTAL CORTEX

among those that WM executive processes have been posited to perform. In the present study, we sought to test the possibility that age-related differences in processes related to dorsolateral frontal brain regions contribute to the age differences observed in performance on WM maintenance tasks. We compared cortical activity observed by Rypma et al. (1999) while younger participants maintained one or six letters with the cortical activity of older participants under the same conditions. In this comparison, younger and older participants were required to maintain the same amounts of information in the task. The question of interest was whether a supracapacity WM load would result in age-differential activation in prefrontal brain regions typically associated with tasks that require manipulation of supracapacity amounts of information held in WM, specifically, dorsolateral PFC. Method

Trial Sequence for 6- v. 1-Letter Condition K:

v

G ((:)

P

R

5s

1r

C

Participants Six older (3 men and 3 women, M age = 68.6 years, range = 62-73 years) and 6 younger people (2 men and 4 women, M age = 25.3 years, range = 22-29 years) participated in the experiment. Younger participants were recruited from the Stanford University Psychology Department. Older participants were recruited from respondents to a newspaper ad. All older participants were healthy community-dwelling individuals living in the San Francisco Bay Area. Younger and older participants scored equivalently on the National Adult Reading Test (Nelson & Willison, 1991; M younger = 39, M older = 40, t < 1; Table 1) and the Mini-Mental Status Exam (MMSE; Folstein, Folstein, & McHugh, 1975; M younger = 30, M older = 29, t < 1). The high scores of the older adults on the MMSE demonstrated their intact cognitive status. Listening Span (Daneman & Carpenter, 1983) performance tended to be higher for younger than for older participants (M younger = 4.5, M older = 3.3, f[10] = 1.78, p < .10). Younger participants completed more Digit-Symbol Substitutions from the Wechsler Adult Intelligence Scale (WAIS; Wechsler, 1981) than did older participants (M younger = 69, M older = 47, r[10] = 3.76, p < .003). Digit-span (also from the WAIS) performance tended to be higher for younger than for older participants (M younger = 7.8, M older = 7.2, /[10] = 1.86, p < .09).

Cognitive Task The cognitive task was modeled after the item-recognition task developed by Steinberg (1966). Each trial of the item-recognition task was composed of three phases (see Figure 1). Phase 1 (1,500 ms) was a target-presentation phase in which six uppercase consonant letters ap-

Table 1 Mean Standardized Test Scores and Standard Errors (in Parentheses)

(R K V) + (G C P)

5s

Phase 1: Target Presentation (Encoding)

Phase 2: Waiting Period (Maintenance)

ir P—

Phase 3: Probe (Retrieval)

Figure 1. Trial sequence and examples of stimuli in the one-letter and six-letter memory-load conditions.

peared on the computer screen. Participants encoded either one or six of these letters. Letters to be remembered were enclosed in parentheses (six letters always appeared on the screen in order to equate the perceptual demands of the two conditions). Phase 2 (5,000 ms) was an unfilled maintenance interval in which the participants viewed a blank screen. Phase 3 (2,000 ms) was a retrieval phase in which a single lowercase "probe" letter appeared among a series of dashes. On one half of the trials, the probe letter matched one of the letters shown in the target-presentation phase ("same" trials); on the other half of the trials, the probe letter did not match one of the letters in the target-presentation phase ("different" trials). On one half of the one-letter different trials were "catch trials" in which the probe letter was one of the five to-be-ignored letters that appeared in the presentation phase. On the other half of the one-letter different trials the probe letter was not one of the five to-be-ignored letters that appeared in the presentation phase. These catch trials were included so that we could determine whether or not participants noticed the periodic changes between one- and six-letter trials. Participants were to respond with a right-thumb button press if the probe letter matched one of the to-be-remembered letters. Each trial was followed by a 500-ms intertrial interval. The task alternated between blocks of one-letter trials and six-letter trials, and there were 4 trials per block. There were 24 trials in each memory-load condition. Each block was 36 s long. The entire task involved six alternating cycles and took 432 s.

MRI Scanning Procedure

Age group Abilities test

Older adults

Younger adults

National Adult Reading Test Mini-Mental Status Exam Listening Span Digit-Symbol Substitution Test Forward Digit Span

40(1.4) 29 (0.8) 3.3 (0.5) 47 (4.6) 7.2 (0.2)

39 (2.2) 30 (0.0) 4.5 (0.4) 69(3.1) a 7.8 (0.2)

* Indicates statistically significant difference between age groups.

Imaging was performed with a 1.5T whole-body MRI scanner (General Electric Medical Systems Signa, Rev. 5.6, Waukesha, WI). For functional MRI (fMRI), a prototype whole-head coil was used for signal amplification. Head movement was minimized with a bite bar formed with each participant's dental impression. A T2* sensitive gradient-echo spiral pulse sequence (Meyer, Hu, Nishimura & Macovski, 1992), which is relatively insensitive to cardiac pulsatility motion artifacts (Noll, Cohen, Meyer, & Schneider, 1995), was used for fMRI with parameters of TR = 720 ms, TE = 40 ms, and flip angle = 65°. Four interleaves were obtained for each image, with a total acquisition time (sampling interval) of 2.88 s per image and an inplane resolution of 2.35 mm X 2.35 mm. Tl-weighted, flowcompensated, spin-warp anatomy images (TR = 500 ms, minimum TE)

374

RYPMA, PRABHAKARAN, DESMOND, AND GABRIELI

were acquired for all sections that received functional scanning. Voxels found to be significantly activated during the functional scan were overlaid on these structural images. Eight 6-mm thick slices were acquired in the horizontal plane of the Talairach and Tournoux (1988) atlas starting from 7.5 mm below the anterior commissure (AC)-posterior commissure (PC) plane, with a 1 -mm interslice interval. Stimuli were generated from a computer and back-projected onto a screen located above the participant's neck by using a magnet-compatible projector. Visual images were viewed from a mirror mounted above the participant's head. The sequence of the presentations of the stimuli was synchronized with the imaging sequence of the scanner.

Data Analysis Image analysis was performed off-line by transferring the raw data to a Sun SparcStation (Sun Microsystems, Mountain View, CA). We used a gridding algorithm to resample the raw data into a Cartesian matrix prior to processing with 2-dimensional Fast Fourier Transform. Once individual images were reconstructed, the time series of each pixel was obtained and correlation methods that take advantage of periodically oscillating paradigms were used to analyze functional activation (Friston, Jezzard, & Turner, 1994). As described by Friston et al. (1994), the hypothesized neural activity was modeled with a square wave reference function that varied in time at the same frequency at which the task varied between high and low memory-load conditions. That is, the frequency of the square wave was computed from the number of task cycles divided by the total time of the experiment. Because the fMRI signal reflects hemodynamic activity that follows the hypothesized neural activity, the hypothesis reference function was convolved with an estimate of the hemodynamic response function derived from previous data collected at our site. For the experiments, one task cycle consisted of a control block and an experimental block each of equal duration. In each scan, there were six task cycles presented over a 432-s time period (frequency = 0.0139 Hz). To perform analyses in PFC regions, we drew regions of interest around inferior (ventrolateral) PFC, middle (dorsolateral) PFC, and superior (rostrolateral) PFC gyri on the basis of the anatomical definition of these gyri in a standard anatomical atlas (Duvernoy, 1991). To construct functional activation maps, the data were analyzed by using a cross-correlation method. Voxels that satisfied the criterion of z a 1.96 (representing a significance of p £ .025, one-tailed) were selected. Raw functional images were motion corrected and then spatially filtered by using a Gaussian filter of 8 mm full-width at half-maximum. This map was then processed with a median filter with a spatial extent of two voxels to emphasize spatially coherent patterns of activation. The filter was used on the assumption that voxels with spuriously high z values (i.e., false positives due to Type I errors) are less likely to occur in clusters than voxels with genuinely high z values, and thus clusters of voxels with high z values are more likely to reflect an active region. The resulting map was overlaid on the Tlweighted structural image. To obtain composite maps of activation over all participants, average functional activation maps were created by transforming each section from each participant to a corresponding standardized horizontal section (Talairach & Tournoux, 1988) at the same distance above and below the AC-PC plane (Desmond et al., 1995). This transformation was done individually for all horizontal sections. Following transformation, composite maps were computed by using a two-stage approach as described by Holmes and Friston (1998). For this approach, the first stage uses all of the images obtained from the scan session for each participant to compute a single contrast image at each slice. The second stage then assesses which voxels of the mean contrast image were significantly different from zero, using a degrees of freedom value that reflects the number of participants rather than the number of scans. Thus, in the first stage, the correlation between the reference waveform and the fMRI time series for a given

participant was summarized into one z image map at each slice, representing the high-versus low-load contrast for that participant. In the second stage, the null hypothesis that the mean z value at each voxel is zero was tested, and voxels that reached a statistical threshold corresponding to p < .005 or lower were displayed on each map.

Results Behavioral Results To examine age differences between older and younger participants, we conducted a 2 (younger vs. older group) X 2 (memoryload conditions) analysis of variance on reaction time (Figure 2) and accuracy data (Table 2). All participants responded more slowly with increasing WM load, F(l, 10) = 68.9, MSB = 8,023.3, p < .0001. Older participants showed a trend toward responding more slowly than younger participants, F(l, 10) = 3.1, MSE = 29,215.8, p < .09 (R2 = .25). The Memory Load X Age Group interaction was not significant (F < \;R2= .03), indicating that increases in reaction time with increases in WM load were equivalent between younger and older participants. Participants were less accurate with increasing WM loads, F(l, 10) = 13.2, MSE = 0.02, p < .005. Older participants tended toward lower accuracy than younger participants, F(l, 10) = 3.1, MSE = 0.03, p < .10 (R2 = .23). The Memory Load X Age Group interaction was not significant (F < 1; R2 = .04), indicating that decreases in accuracy with increases in WM load were equivalent between younger and older participants. Nonparametric tests of the relative differences between the one-letter and six-letter conditions in younger and older participants were also not significant. There were also no age-related differences in performance on catch trials.

fMRI Results The 6-load versus 1-load scan (referred to as the 6-1 scan from now on) yielded a number of activations greater for the 6-load than the 1-load task (see the top portion of Table 3 and Figure 3, top panel). For younger participants, major foci of activity occurred bilaterally in inferior frontal gyms (BAs 44, 45, and 47) and precentral gyrus (BA 6), more in spatial extent on the left than on

1100 1000

5 900 i=

-Older •Younger

800 700 600 500

6 Letters

1 Letter Memory Load

Figure 2. Reaction time results for older participants (diamonds) and younger participants (squares) in the one-letter and six-letter memory-load conditions.

AGE DIFFERENCES IN PREFRONTAL CORTEX Table 2 Mean Accuracy Rate and Standard Errors (in Parentheses) Memory load Group

1 letter

6 letters

Older adults Catch trials Younger adults Catch trials

0.89 (0.05) 0.94 (0.09) 0.97 (0.02) 1.00(0.00)

0.67 (0.09) 0.82 (0.06)

the right. Middle frontal gyms activity was also observed bilaterally (BAs 8, 9, and 46) but was greater (i.e., z values were greater), and extended more superiorly, on the right than on the left. Activation also occurred in bilateral superior frontal regions (BA 10) and was greater in dorsal-ventral extent on the right than on the left. Other regions of major activity in this scan were anterior cingulate (BA 32), medial frontal gyrus (BA 24), right caudate, and thalamus bilaterally. Minor foci of activity occurred bilaterally in middle temporal lobes (BA 21), left superior occipital gyrus (BA 19), and bilaterally in inferior parietal lobules (BA 40). For older participants (see the middle portion of Table 3 and Figure 3, middle panel) major foci of activity occurred in inferior frontal gyrus (BAs 44, 45, and 47), lateralized to the left hemisphere, and precentral gyrus (BA 6), more in spatial extent on the left than on the right. Major foci of activation also occurred in superior and medial frontal regions (BA 10), more on the left than on the right. Minor foci of activity occurred in middle frontal gyrus bilaterally (BAs 9 and 46). Other regions of activity in this scan included anterior cingulate (BA 32) and inferior parietal lobule (BA 40). Minor foci of activity occurred in middle and superior temporal lobe (BA 39), middle occipital gyrus, and cuneus (BA 19). Differences in activation between younger and older participants were compared directly by using random effects r-test comparisons between age groups (see the bottom portion of Table 3 and Figure 3, bottom panel). Results of these tests indicated activation that was greater for younger than older participants in left precentral gyrus (BA 6), middle frontal gyrus (BAs 9 and 46), right cingulate gyrus (BA 30), and left precuneus (BA 7), as well as left caudate, putamen bilaterally, and left insula. Activation that was greater for older participants than younger participants occurred in left superior and medial frontal gyri (B A 10). No differences between younger and older participants were observed in left inferior frontal regions (i.e., BAs 44, 45, and 47). To examine patterns of functional activation differences in distinct PFC regions, two further analyses were performed. First, the significant activations (the number of suprathreshold voxels in the composite maps) observed in the analyses of younger and older participants were summed according to hemisphere (left and right) and PFC region (ventral, dorsal, and rostral; see Figure 4). Younger participants showed greater activation (measured by the number of suprathreshold voxels) than older participants principally in dorsolateral PFC regions corresponding to BAs 9 and 46. Younger and older participants showed equivalent activation in left ventrolateral PFC corresponding to BAs 44, 45, and 47. Older participants showed greater activation than younger participants in left rostral PFC regions corresponding to BA 10.

375

The above analysis of suprathreshold voxels relies on use of a particular statistical threshold (z > 1.96 in this case). In a second analysis we examined age-differential dorsolateral activation and age-equivalent ventrolateral activation by age-group comparisons of PFC activation without assumptions inherent in the use of a z threshold. To examine age differences in PFC regional activation independent of a z threshold, we drew regions of interest around inferior (ventrolateral) PFC and middle (dorsolateral) PFC regions on the basis of their definition in a standard anatomical atlas (Duvemoy, 1991). We then computed for each participant a relative ratio of regional activity, defined as the regional, or ROI-wise, z-value increase (relative to baseline) to the corresponding z-value increase across the entire set of slices that contained the ROI. Examination of age differences in relative ratios in each region allowed us to compare, between younger and older participants, the extent of activation increase in a region, relative to increases that would be expected to occur in any random brain region, independent of any z threshold. Age differences in relative ratios of activity were assessed separately in each hemispheric region, between younger and older participants, with Mann-Whitney tests. Means of relative ratios for younger and older participants in middle and inferior frontal gyri are shown in Table 4. The largest difference between younger and older participants was in right middle frontal gyrus, and this difference was significant (p = .05). No other comparisons approached significance (all ps > .10).

Discussion This study compared activation in younger and older adults performing a WM maintenance task that included encoding, maintaining, and retrieving high or low loads of verbal information. In inferior frontal gyri (ventrolateral PFC) there was greater left- than right-hemisphere activity and no apparent age-related differences. In the middle frontal gyri (dorsolateral PFC) younger participants showed greater right- than left-hemisphere activity, and there was greater activity in younger participants than older participants. Older participants did, however, show greater activity relative to younger participants in left superior frontal gyri (rostrolateral PFC).

Interpreting Age-Related Differences

in JMR1 Results

Drawing inferences from random-effects tests regarding age differences in the neural substrates of cognitive processes from fMRI data rely on the assumption of an age-equivalent relationship between neural activity and the blood-oxygen-level-dependent hemodynamic response function (HRF) and adequate statistical power. The HRF-age-equivalence assumption has been tested in a number of studies (D'Esposito, Zarahn, Aguirre, & Rypma, 1999; Ross et al., 1997; Taoka et al., 1998) with results indicating age-related increases in noise components of the HRF but not in signal components (D'Esposito et al., 1999). D'Esposito et al. pointed out that, with findings of uniform age-related reductions in neuroimaging signal (as measured by t or z statistics, e.g.), attribution of the results to changes in hemodynamic coupling cannot be ruled out. If, however, age-related reductions in some regions, but age-related increases in others, are observed, it is unlikely that age-related changes in hemodynamic coupling could account for

376

RYPMA, PRABHAKARAN, DESMOND, AND GABRIELI

Table 3 Regions of Significant Activation in Younger and Older Participants Talairach Lobe

Region of activation

Hemisphere/ Brodmann's

X

Z score

y

z

53

1 20 12 32

Voxels"

Younger participants Frontal

Superior/middle Middle Middle/inferior

Inferior frontal

Inferior/premotor Inferior/insula Premotor

L10 L46/10 RIO R9 R6,8 L46 RIO R44,45,46,10 R45/46 R45,46,10 L45 L46 L44 R44 L44/6 L45 L6

-27 -37 28 29 41 -48 25 33 32 35 -49 -39 -52 41 -52 -38 -59

-51 Cingulate Parietal Temporal

Medial frontal Anterior Inferior parietal Supramarginal/angular Superior parietal Superior Middle Middle/inferior

Occipital

Lingual Cuneus Precuneus

Subcortex

Caudate/putamen/GP Caudate Globus pallidus Putamen Caudate/putamen/globus pallidus/claustrum Thalamus

B8 B24/32 L40

0 1 -49

L39

-30

L7 L22/42 L21 L21/37 R21 L37/21 R37/21 L18/19

-25 -66

-53 -57 43 -49

56

R R L L L

-18 12 5 20 0 16 14 -18 -13 -20

R B

-6

R19 R18 R31 L7

24

6

-4 1 1 24 20 45 8 20 20

2 11

-4 12

4.07 3.72 3.72 3.72 3.72 3.04 4.07 3.72 3.53 3.72 3.94 3.31 3.72 3.72 3.72 3.72 3.53 3.72 3.18 3.72 3.16 3.55 3.33 3.72 3.97 3.72 3.14 3.60 3.60 3.48 3.40 2.94 2.74 3.38 3.43 3.67 3.26 3.23 3.01

22 -22

1 12

3.82 3.72

616 2,923

46 42

20

3.04 3.04

2,711 2,526

3.28 2.42 2.89 3.23 2.33 2.47 3.04 3.04 2.52 2.08 2.23 3.28 3.04 3.04 2.33 2.94

434

29 35 31 1 43 53 12 23 28 26

35 5 5

2 22 4 -5 24 20 -34 -53 -63 -34 -30

-53 -50 -56 -45 -73 -59

-9 -69 -73 10 c

45 8 1

8 20 24 1 12 20 32 32 12 12 45 45 32 45 32 45 8 1

8 1 -4

398 2,001 1,730 2,162 2,390 86 1,795 4,064 1,733 5,006

239 335 549 2,162 5,648 1,348

389 1,102

292 4,413

560 753 444

388 891 603 101 309 290 604 139 452 129 471 1,000 1,028 469 87 251

Older participants Frontal

Superior/medial Superior/medial/cingulate

L10 L10.32

-24

Superior Superior/middle Middle

BIO R8 L10 R46 RIO R 10,46 R9.46 L9 R9 R6 L46 L47 L44 L6,44,45 L6 R6

-8 27 -24 47 42 26 38 -43 42 48 -47 -48 -48 -55 -49 48

Middle/inferior Inferior frontal Inferior frontal/premotor Premotor

-28

61 31

41 46 41 49 43

18 25 0

16 20 4 -3 4 -4

12 1 45 1

8 12 20 24 32 32 6 24 1 24 12 32 32

59 314 209 495 483 457 466 369 72 1,235 182 569 7,210 281 2,055

AGE DIFFERENCES IN PREFRONTAL CORTEX

377

Table 3 (continued) Talairach Lobe

Region of activation

Hemisphere/ Brodmann's

Z score

z

y

X

Voxels"

Older participants (continued) Cingulate

Anterior

Q O

29

38

L24 L32 R32

2 -11 -7 5

10 12 5

-4 20 32 45 45

R39 R40

37 44

-56 -55

24 45

L40

-33

-50

45

R39 L22 R42

40 -55 47

-58 8 -27

32 1 8

L22 L21 R21 L21 L39

-43 -52 59 -40 -50

-41 -14 -15 -52 -58

20 -4 -4 8 12

L19 R19

-30 8

-55 -93

1 24

L7

-20

-60

32

20 3

-93 14

1 8

19

-6 11 -29

-8 -29 2 18 -17 -13 -40

24 -4 -4 12 12 12 -4

20 22 -35

43 36 41

-4 16 1

31

47 17 31 22 17 10 29

1 16 32

L24 R24/32

Parietal

Inferior parietal

Temporal

Supramarginal Superior Middle

Occipital

Subcortex

Lingual Cuneus Precuneus Occipital Caudate Putamen Insula Thalamus Parahippocampal

R18 R

R L R L L R L19

23

18

-30

3.28 2.50 3.04 3.04 3.04 2.23 2.28 2.77 3.04 2.57 2.47 2.13 3.28 3.21 2.72 2.33 2.62 2.45 2.52 2.79 2.64 2.45 2.94 2.47 2.38 3.04 2.40 3.28

589 1,198

767 953 948 507 258

526 655 73 99

142 647 217 543 178 105 144

102 97 374

112 465 62 163 911

211 792

Young-old Frontal

L44

38 36 15 -32 -36 39 -58

Inferior/premotor Anterior

L44/6 L24 R32

-48 -3 12

-2 4 20

32 45 32

Posterior Inferior Inferior/superior Inferior Middle Caudate

R23 L40

9 -52

-38

-33

12 45

L40/7 R37 L19 L

-27 57 -48 -9

-64 -44 -77 2

45 -4 -4 -4

Insula

R R

-19 18 31

4 8 13

16 1 8

2.08 2.24 3.17 3.29 2.47 2.43 2.80 2.36 2.29 2.61 2.95 3.19 2.24 2.43 2.14 2.09 2.23 2.45 2.29 2.44 2.87 2.29 3.06

Superior Posterior Insula

L10 L23 L

-25 -20 -44

52 -63 3

16 16 1

1.97 1.98 1.97

Superior Superior/cingulate Middle/inferior

Inferior

RIO RIO/23 LI 0/44 RIO/44 R45/46 R9/46

R45 L45 L44/45 R44/45

Cingulate Parietal Temporal Occipital Subcortex

„ -1

8 8

20 20 45

53 249 351

901 151 94 667 311 241 145 420 529 48 205

72 58 77 158 30 171

363 89 321

Old-young Frontal Cingulate Subcortex

730 43 150

Note. Rostrolateral prefrontal cortex (PFC) = Brodmann's Area (BA) 10; dorsolateral PFC = BAs < and 46; ventrolateral PFC = BAs 44, 45, and 47. L = left; R = right; B = bilateral. a Suprathreshold.

378

RYPMA, PRABHAKARAN, DESMOND, AND GABRIELI

Figure 3. Functional magnetic resonance imaging (fMRI) results of the young (top panel), old (middle panel), and unmatched /-test comparison of young and old (bottom panel) superimposed on pictures of axial cuts at 12 mm (left column) and 32 mm (right column) above the anterior commissure-posterior commissure plane (Talairach & Tournoux, 1988). In the bottom panel, the yellow-to-red scale indicates significant age-differential activity favoring younger participants; white-to-blue scale indicates significant age-differential activity favoring older participants.

these results. Such a pattern of results was observed in the present study. Conclusions about similarities and differences between two groups depend on sufficient power to detect statistical effects. The relatively small number of participants could raise concerns about

power in the present study. There was sufficient power, however, to detect effects favoring the younger group in some regions (e.g., dorsolateral PFC) and the older group in other regions (e.g., rostrolateral PFC). This result indicates that there was sufficient power to detect age-related effects in the present study (e.g., Grady

379

AGE DIFFERENCES IN PREFRONTAL CORTEX

1400 i

o-

Left Rostral

Left Ventral

Right Ventral

Right Dorsal

Right Rostral

Figure 4. Numbers of suprathreshold voxels in left- and right-hemisphere prefrontal regions in younger and older participants.

et al., 1994). Further, the nearly identical levels of activation in left ventrolateral PFC (Figure 4) suggest that the lack of that age effect was not due to insufficient power. Another factor that complicates interpretation of age-differential patterns of activity is that, across studies, age-related differences in activation have not been consistently linked to age-related increases or decreases in cognitive performance. Regions of increased activity in older adults, relative to younger adults, have been observed in a number of studies (e.g., Cabeza et al., 1997; Grady et al., 1992; Madden et al., 1999; Reuter-Lorenz et al., in press). These age-related increases in activity have been accompanied by age-equivalent performance in some cases (Cabeza et al., 1997) and age-differential performance in others (ReuterLorenz et al., 2000). Jonides et al. (2000) have observed still a third pattern of activation-performance relationships, age-related reductions in positron-emission tomography (PET) activation associated with reduced cognitive performance in older persons relative to younger persons. Reuter-Lorenz et al. (2000) have reported age-related changes in the hemispheric asymmetry of PFC activation during WM perfor-

Table 4 Ratios of Regional Activity in Dorsolateral and Ventrolateral PFC Frontal region Ventrolateral

Dorsolateral

Group

Left

Right

Left

Right

Older adults Younger adults

1.45 1.72

1.60 1.32

1.17 1.11

1.30 1.72a

Note. PFC = prefrontal cortex. a Indicates statistically significant regional age difference.

mance. Younger adults demonstrated asymmetric PFC activation as a function of the material being maintained in WM. Thus, PFC activation was strongly left lateralized during a verbal WM task and right lateralized for a spatial WM task. Older adults, however, showed bilateral PFC activation during both verbal and spatial WM tasks because of age-related increases in activity in the nonspecialized hemisphere. In the present study, we found a contrasting influence of age on PFC laterality. Younger adults showed a right-hemisphere (dorsolateral PFC) activation increase in response to increasing verbal WM load, whereas older adults showed a left-hemisphere (rostrolateral PFC) activation increase in response to increasing verbal WM load. The variance between the results of the present study and those of Reuter-Lorenz et al. (2000) may be due to a number of factors. The older adults in the Reuter-Lorenz et al. study had significantly slower response times than the younger adults, whereas the response times of the older adults in the present study were more similar to those of the younger adults. Also, the prior study stressed WM with both memory-load and retention-interval demands, whereas the present study stressed WM only with memory load. Studies with younger adults that have specifically examined the relationship between activity and performance have shown decreased activation in dorsolateral PFC when their performance was fast and accurate relative to when it was slower and less accurate. Older adults, in contrast, showed increased activation in dorsolateral PFC when their performance was fast and accurate relative to when it was slower and less accurate (Rypma & D'Esposito, 1999, 2000). These results suggest that fast, accurate performance on WM tasks (such as would be observed in subcapacity WM conditions) may be associated with an optimal activation level. Deviations above or below this optimal level may be associated with slower and less accurate performance (such as would be observed in supracapacity WM conditions; cf. Kimberg & Farah, 1993;

380

RYPMA, PRABHAKARAN, DESMOND, AND GABRIELI

Kimberg, D'Esposito, & Farah, 1997; Servan-Schreiber, Printz, & Cohen, 1990). Such performance-activation relationships were also observed in the present study. First, when participants were slowed as a result of supracapacity WM conditions, younger participants showed greater activation than older participants in regions critical to WM performance (i.e., dorsolateral PFC), possibly reflecting cognitive processes that are more available to younger than to older adults. Second, age-related increases in activation were associated with age-equivalent performance, possibly reflecting the operation of compensatory strategies in the older participant group. Age-related increases in activation have been interpreted as reflecting functional compensation by a number of researchers (e.g., Cabeza et al., 1997; Grady et al., 1994, 1995; Reuter-Lorenz et al., in press). In one PET study, Cabeza et al. (1997) suggested that age-related activation increases observed in insular regions during episodic encoding and PFC during episodic retrieval may reflect functional compensation.

Ventrolateral Prefrontal Cortex The requirement to maintain a supracapacity WM load resulted in age-equivalent activation in ventrolateral PFC regions. Indeed, analyses of suprathreshold activation revealed nearly identical activation for younger and older groups in left ventrolateral PFC corresponding to BAs 44, 45, and 47. Posterior left ventral PFC regions have been implicated as an important constituent of the phonological loop or verbal slave system. For example, Broca's Area and adjacent speech-related regions may mediate maintenance of stored information through a process of verbal rehearsal (Awh et al., 1996; Paulesu et al., 1993; Rypma et al., 1999; Vallar & Baddeley, 1984). In one study, for instance, brief retention of phonological information resulted in activation associated with subvocal rehearsal in left inferior frontal cortex (Paulesu et al., 1993). The present result suggests that WM slave system processes, related to verbal rehearsal and supported by ventrolateral PFC, may be spared the deleterious effects of advancing age. Not all cognitive processes identified with inferior PFC appear to be so spared, however. Other neuroimaging studies have observed agerelated reduction in inferior regions of prefrontal cortex (Buckner, Sanders, Kelley, Snyder, & Morris, 1999; Stebbins et al., 1997) during encoding of semantic information. Further research is required to understand the circumstances under which agedifferential or age-equivalent activation of the ventrolateral PFC regions is observed.

Dorsolateral Prefrontal Cortex The requirement to maintain a supracapacity WM load resulted in far greater activation in younger than older participants in dorsolateral PFC. This pattern of age differences in activation suggests that cognitive operations required for optimal performance in supracapacity WM conditions may be more available to younger than to older adults. It may be that supracapacity WM performance requires manipulation of to-be-remembered information for-efficient storage and retrieval (Rypma et al., 1999; Rypma & D'Esposito, 1999, 2000).

Results from studies that explicitly require manipulation of stored information support the idea that age differences in high memory-load maintenance may be the result of an age-differential ability to effectively manipulate information in the service of optimizing memory performance. Results of several studies indicate large age differences in performance when participants must reorder stored information prior to retrieval (Botwinick & Storandt, 1974; Bromley, 1958; Craik, Morris, & Gick, 1990) compared with when participants are required to recall list items in the order that they were presented (Botwinick & Storandt, 1974; Bromley, 1958; Craik, 1968; Drachman & Leavitt, 1972; Friedman, 1974; Gilbert, 1941; Gilbert & Levee, 1971; Kriauciunas, 1968; Taub, 1973; Wiegersma & Meertse, 1990). Craik (1986), for instance, observed age-equivalent performance in a standard digitspan task but age-differential performance on an alpha-span task in which participants were given lists of words and were required to recall them in alphabetical order. These results suggest that manipulation processes that may be required for the successful maintenance of supracapacity WM loads (i.e., more than two to three items) are impaired by aging. Several lines of brain imaging research indicate involvement of dorsolateral PFC in such manipulation processes in WM tasks (Braver et al., 1997; Cohen et al., 1997; D'Esposito et al., 1995) and other cognitive tasks believed to require WM manipulation processes such as mental rotation tasks (e.g., Rypma et al., 1996; Zacks, Rypma, Gabrieli, Tversky, & Glover, 1999), divided attention tasks (e.g., Corbetta et al., 1991), and high-level reasoning tasks (e.g., Prabhakaran et al., 1997; Prabhakaran et al., 2001). Thus, age-related deficits in WM and other cognitive tasks may be tied to the disproportionate effect of aging on dorsolateral PFC. Differentially greater involvement of dorsolateral PFC by younger adults may reflect greater reliance on executive processes (e.g., mnemonic strategies), possibly at the time of encoding (Rypma & D'Esposito, 1999), in order to sustain performance in a high memory-load task (Rypma et al., 1999). Age-related differences in PFC activation could have resulted from age-group differences in the six-letter condition, the oneletter condition, or both. Because the one-letter condition served as the baseline in our experiment, it is not possible to determine with certainty whether there were age-related differences in the oneletter condition per se. Age-related attenuation of activity in the dorsolateral PFC could have resulted from increased activity in the one-letter condition in older adults as opposed to decreased activity in the six-letter condition in older adults. The possibility that there are age-related differences at low loads is suggested by results from previous studies that have indicated greater dorsolateral PFC activity in older than in younger adults at low memory loads in the presence of age-equivalent performance accuracy (four letters; Reuter-Lorenz et al., 2000). In contrast, however, other WM studies have shown age-equivalent performance with even lower memory loads that more closely approximate the one-letter condition of the present study (two letters; Rypma & D'Esposito, 2000). One reason to suspect an age-differential increase in activation in the one-letter condition is suggested by research indicating that older adults have greater difficulty in selective attention tasks that require participants to inhibit the encoding of irrelevant letters or words that are presented adjacent to to-be-attended letters (e.g., Hasher, Stoltzfus, Zacks, & Rypma, 1991). In the one-letter condition of the present study, the target letter was embedded in an

AGE DIFFERENCES IN PREFRONTAL CORTEX

array of six letters in order to equate the perceptual characteristics of the one- and six-letter encoding phases. This feature of the task may have introduced attentional selection demands into the oneletter memory-load condition (i.e., it may have required participants to inhibit encoding of the five irrelevant letters) that were differentially greater for older than for younger adults. Such agerelated task-demand differences could potentially lead to an attenuation of fMRI activation differences between the one- and sixletter conditions in the older group. Although we could not assess age-related differences in the one-letter condition of the present study (given that it was our baseline condition), analyses of performance on the catch trials did not suggest such an inhibition deficit in the older adults. That is, we did not observe age-related differences in accuracy between younger and older adults on one-letter trials where the correct response was different and the probe letter came from the set of to-be-ignored letters seen in the target presentation phase (i.e., the catch trials). This result suggests that the selection demands in the one-letter condition were not greater for older than for younger adults. It is still possible that effective resolution of the interference created by the distracting letters led to age-differential increases in activation in the oneletter condition. However, studies that have specifically examined age differences in interference resolution have implicated ventrolateral prefrontal cortex (e.g., Jonides et al., 2000), a region that showed age-equivalent activation in the present study. Further, the result reported by Jonides et al. (2000) was one of less activation in older adults compared with younger adults. Such a result would have enhanced the one versus six difference in the older adults. Thus, it is reasonable to assume age-equivalent activation in the one-letter condition, although more research is certainly required to understand the conditions under which age-equivalent and agedifferential activation is observed during WM maintenance tasks.

Rostrolateral Prefrontal Cortex In the present study, an age-related increase in left superior, or rostrolateral, PFC regions (BA 10) occurred in the presence of age-equivalent performance. This activation may reflect any of several psychological processes. One possibility is that this agerelated activation increase may reflect the processing of internal emotion associated with the high memory-demand condition. Rostrolateral PFC activation has been observed in younger adults (e.g., Lane et al., 1997) and older adults (e.g., Paradise et al., 1997) while viewing emotional stimuli. Alternatively, the age-related activation increase may reflect compensatory WM processes. Activation in rostrolateral PFC regions (BA 10) has also been observed during episodic memory task performance (Andreasen et al., 1995; Cabeza, 2001; Cabeza et al., 1997; Nolde, Johnson, & D'Esposito, 1998; Schacter, Savage, Alpert, Rauch, & Albert, 1996). These results, together with those of Rypma and D'Esposito (2000), suggest that older adults use alternative compensatory strategies to successfully perform supracapacity WM tasks. For instance, older adults may place relatively greater emphasis on the use of encoding operations. Increased activation in left prefrontal regions when memory tasks require additional encoding has been observed in previous brain imaging studies (e.g., Andreasen et al., 1995). Results from studies of episodic memory suggest that the greater left prefrontal activation observed in older adults reflects their greater reliance on the use of more deliberative memory

381

processes (Nolde et al., 1998). Thus, differentially greater involvement of left medial PFC by older participants may reflect greater reliance on compensatory strategies possibly related to a more deliberative analysis of memory (Nolde et al., 1998; Schacter et al., 1996) or age-related changes in encoding strategies (Andreasen et al., 1995). Further studies are needed to constrain the interpretation of the age-associated increase in left BA 10 activation.

Caudate Nucleus We also observed age-related reductions in activation in the head of caudate. Clinical, anatomical, and electrophysiological studies delineate a frontostriatothalamic loop in WM (e.g., Houk & Wise, 1993). One possible role of caudate nucleus in WM could be to mediate the sustained activity known to occur in dorsolateral PFC during the delay periods of trials in WM tasks (Cohen et al., 1997; Goldman-Rakic & Friedman, 1991). Houk and Wise (1993) have suggested that output from the striatum, whose cells code stimulus features relevant for WM, to the pallidum causes phasic inhibition of that structure and its inhibitory connection to the thalamus. As a result, resonating activity between the thalamus and dorsolateral prefrontal cortex could be expected to occur, given known bidirectional connections between these structures. Agerelated reduction in such a mechanism could mediate age differences in the availability of resources for WM maintenance and manipulation (Gabrieli, 1996). The age-related changes in sustained activity in caudate and PFC regions observed in our study and other studies are consistent with this hypothesis.

Conclusion Results from the present study suggest a number of conclusions regarding age-related differences in PFC regions during maintenance of information in WM. Left ventrolateral PFC showed age-equivalent activation in the present study. The age-equivalent activation observed in ventrolateral PFC may support the sparing of the phonological loop in advancing age. Older participants had reduced activation in right dorsolateral PFC and caudate. The functional compromise of these brain regions may account for executive WM and other cognitive deficits that are characteristic of normal aging. The age-related activation increase in rostrolateral PFC may reflect a compensatory mechanism that contributed to the nearly age-equivalent performance we observed in the present study.

References Anders, T. R., & Fozard, J. L. (1973). Effects of age upon retrieval from primary and secondary memory. Developmental Psychology, 9, 411415. Anders, T. R., Fozard, J. L., & Lillyquist, T. D. (1972). Effects of age upon retrieval from short-term memory. Developmental Psychology, 6, 214217. Andreasen, N. C., O'Leary, D. S., Arndt, S., Cizaldo, T., Hurtig, R., Rezai, K., Watkins, G. L., Boles Ponto, L. L., & Hichwa, R. D. (1995). Short-term and long-term verbal memory: A positron emission tomography study. Proceedings of the National Academy of Science, 92, 5111-5115. Awh, E., Jonides, J., Smith, E., Schumacher, E., Koeppe, R., & Katz, S. (1996). Dissociation of storage and rehearsal in verbal working memory: Evidence from PET. Psychological Science, 7, 25-31.

382

RYPMA, PRABHAKARAN, DESMOND, AND GABRIELI

Bachevalier, J. L., Landis, S., Walker, L. C., Brickson, M., Mishkin, M., Price, D. L., & Cork, L. C. (1991). Aged monkeys exhibit behavioral deficits indicative of widespread cerebral dysfunction. Neurobiology of Aging, 12, 99-111. Baddeley, A. (1986). Working memory. New York: Oxford University Press. Baddeley, A., & Hitch, G. J. (1974). Working memory. In G. Bower (Ed.), Recent advances in learning and motivation (Vol. VIII, pp. 47-90). New York: Academic Press. Baker, S. C., Rogers, R. D., Owen, A. M., Frith, C. D., Dolan, R. J., Frackowiak, R, S. J., & Robbins, T. W. (1996). Neural systems engaged by planning: A PET study of the Tower of London task. Neuropsychologia, 34, 515-526. Bartus, R. T., Dean, R. L. Ill, & Fleming, D. L. (1979). Aging in the rhesus monkey: Effects on visual discrimination learning and reversal learning. Journal of Gerontology, 34, 209-219. Beal, M. F., Walker, L. C., Storey, E., Segar, L., Price, D. L., & Cork, L. C. (1991). Neurotransmitters in neocortex of aged rhesus monkeys. Neurobiology of Aging, 12, 407-412. Boaz, T. L., & Denney, D. R. (1993). Speed of scanning in primary memory in persons with dementia of the Alzheimer type. Psychology and Aging, 2, 294-300. Botwinick, J., & Storandt, M. (1974). Memory, related functions, and age. Springfield, IL: Charles C Thomas. 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. Neurolmage, 5, 49-62. Brizzee, K. R., Ordy, J. M., & Bartus, R. T. (1980). Localization of cellular changes within multimodal sensory regions in aged monkey brain: Possible implications for age-related cognitive loss. Neurobiology of Aging, 1, 45-52. Bromley, J. (1958). Some effects of age on short-term learning and memory. Journal of Gerontology, 13, 12-21. Buckner, R. L., Sanders, A. L., Kelley, W. M., Snyder, A., & Morris, J. C. (1999). Functional anatomy of verbal and nonverbal encoding in young and old adults. Society for Neuroscience Abstracts, 25, 295. Cabeza, R. (2001). Functional neuroimaging of cognitive aging. In R. Cabeza & A. Kingstone (Eds.), Handbook of functional neuroimaging of cognition (pp. 331-377). Cambridge, MA: MIT'Press. Cabeza, R., Grady, C. L., Nyberg, L., Mclntosh, A. R., Tulving, E., Kapur, S., Jennings, J. M., & Craik, F. I. M. (1997). Age-related differences in neural activity during memory encoding and retrieval: A positron emission tomography study. Journal of Neuroscience, 17, 391-400. Cohen, J. D., Forman, S. D., Braver, T. S., Casey, B. J., Servan-Schreiber, D., & Noll, D. C. (1994). Activation of the prefrontal cortex in a nonspatial working memory task with functional MRI. Human Brain Mapping, 1, 293-304. Cohen, J. D., Perlstein, W. M., Braver, T. S., Nystrom, L. E., Noll, D. C., Jonides, J., & Smith, E. E. (1997, April). Temporal dynamics of brain activation during a working memory task. Nature, 386, 604-608. Corbetta, M., Miezin, F., Dobmeyer, S., Shulman, G., & Petersen, S. (1991). Selective and divided attention during visual discriminations of shape, color, and speed: Functional anatomy by positron emission tomography. Journal of Neuroscience, 11, 2383-2402. Craik, F. I. M. (1968). Short-term memory and the aging process. In G. A. Talland (Ed.), Human aging and behavior (pp. 131-168). New York: Academic Press. Craik, F. I. M. (1977). Age differences in human memory. In J. E. Birren & K. W. Schaie (Eds.), Handbook of the psychology of aging (pp. 384-420). New York: Van Nostrand Reinhold. Craik, F. I. M. (1986). A functional account of age differences in memory. In F. Klix & H. Hagendorf (Eds.), Human memory and cognitive capabilities, mechanisms and performances (pp. 409-422). Amsterdam: North Holland.

Craik, F. I. M., & Jennings, J. M. (1992). Human memory. In F. I. M. Craik & T. A. Salthouse (Eds.), The handbook of aging and cognition (pp. 51-110). Hillsdale, NJ: Erlbaum. Craik, F. I. M., Morris, R. G., & Gick, M. L. (1990). Adult age differences in working memory. In G. Vallar & T. Shallice (Eds.), Neuropsychological impairments of short-term memory (pp. 247-267). Cambridge, England: Cambridge University Press. Daneman, M., & Carpenter, P. A. (1983). Individual differences in integrating information between and within sentences. Journal of Experimental Psychology: Learning, Memory, and Cognition, 9, 561-584. Deiber, M. P., Passingham, R. E., Colebatch, J. G., Friston, K. J., Nixon, P. D., & Frackowiak, R. S. (1991). Cortical areas and the selection of movement: A study with positron emission tomography. Experimental Brain Research, 84, 393-402. Desmond, J. E., Sum, J. M., Wagner, A. D., Demb, J. B., Shear, P. K., Glover, G. H., Gabrieli, J. D. E., & Morell, M. J. (1995). Functional MRI measurement of language lateralization in Wada-tested patients. Brain, 118, 1411-1419. D'Esposito, M., Detre, J. A., Alsop, D. C., Shin, R. K., Atlas, S., & Grossman, M. (1995, November). The neural basis of the central executive system of working memory. Nature, 378, 279-281. D'Esposito, M., Zarahn, E., Aguirre, G. K., & Rypma, B. (1999). The effect of normal aging on the coupling of neural activity to the BOLD hemodynamic response. Neurolmage, 10, 6-14. Drachman, D., & Leavitt, J. (1972). Memory impairment in the aged: Storage versus retrieval deficit. Journal of Experimental Psychology, 93, 302-308. Duvernoy, H. (1991). The human brain. Wien, Austria: Springer-Verlag Wien. Eriksen, C. W., Hamlin, R. M., & Daye, C. (1973). Aging adults and rate of memory scan. Bulletin of the Psychonomic Society, 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. Friedman, H. (1974). Interrelation of two types of immediate memory in the aged. Journal of Psychology, 87, 177-181. Friston, K. J., Jezzard, P., & Turner, R. (1994). Analysis of functional MRI time-series. Human Brain Mapping, 1, 153-171. Funahashi, S., Bruce, C., & Goldman-Rakic, P. (1989). Mnemonic coding of visual space in the monkey's dorsolateral prefrontal cortex. Journal of Neurophysiology, 61, 331-349. Fuster, J. M., & Alexander, G. E. (1971, August). Neuron activity related to short-term memory. Science, 173, 652-654. Gabrieli, J. D. E. (1996). Memory systems analyses of mnemonic disorders in aging and age-related disease. Proceedings of the National Academy of Sciences of the United States of America, 93, 13534-13540. Gilbert, J. G. (1941). Memory loss in senescence. Journal of Abnormal and Social Psychology, 36, 73-86. Gilbert, J. G., & Levee, R. F. (1971). Patterns of declining memory. Journal of Gerontology, 26, 70-75. Glanzer, M., & Razel, M. (1974). The size of the unit in short-term storage. Journal of Verbal Learning and Verbal Behavior, 13, 114-131. Goldman-Rakic, P. S., & Brown, R. M. (1981). Regional changes of monoamines in cerebral cortex and subcortical structures of aging rhesus monkeys. Neuroscience, 6, 177-178. Goldman-Rakic, P. S., & Friedman, H. (1991). The circuitry of working memory revealed by anatomy and metabolic imaging. In H. S. Levin, H. M. Eisenberg, & A. Benton (Eds.), Frontal lobe function and dysfunction (pp. 72—91). New York: Oxford University Press. Grady, C. L., Haxby, J. V., Horwitz, B., Schapiro, M. B., Rapoport, S. I., Ungerleider, L. G., Mishkin, M., Carson, R. E., & Herscovitch, P. (1992). Dissociation of object and spatial vision in human extrastriate cortex: Age-related changes in activation of regional cerebral blood flow

AGE DIFFERENCES IN PREFRONTAL CORTEX measured with [15O] water and positron emission tomography. Journal of Cognitive Neuroscience, 4, 23-34. Grady, C. L., Maisog, J. M., Horwitz, B., Ungerleider, L. G., Mentis, M. J., Salerno, J. A., Pietrini, P., Wagner, E., & Haxby, J. V. (1994). Agerelated changes in cortical blood flow activation during visual processing of faces and location. The Journal of Neuroscience, 14, 1450-1462. Grady, C. L., Mclntosh, A. R., Horwitz, B., Maisog, J. M., Ungerleider, L. G., Mentis, M. J., Salerno, J. A., Pietrini, P., Wagner, E., & Haxby, J. V. (1995, July). Age-related reductions in human recognition memory due to impaired encoding. Science, 269, 218-221. Hasher, L., Stoltzfus, E. R., Zacks, R. T., & Rypma, B. (1991). Age and inhibition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 17, 163-169. Haug, H., & Eggers, R. (1991). Morphometry of the human cortex cerebri and corpus striatum during aging. Neurobiology of Aging, 12, 336-338. Heilbroner, P. L., & Kemper, T. L. (1990). The cytoarchitectonic distribution of senile plaques in three aged monkeys. Acta Neuropatholagica, 81, 60-65. Holmes, A. P., & Friston, K. J. (1998). Generalisability, random effects and population inference. Neuroimage: Abstracts of the 4th International Conference on Functional Mapping of the Human Brain, 7, S754. Houk, J. C., & Wise, S. P. (1993). Outline for a theory of motor behavior: Involving cooperative actions of the cerebellum, basal ganglia, and cerebral cortex. In P. Rudomin, M. A. Arbib, & F. Cervantes-Perez (Eds.), From neural networks to artificial intelligence (pp. 452-470). Heidelberg, Germany: Springer-Verlag. Huttenlocher, P. R. (1979). Synaptic density in human frontal cortex— Developmental changes and the effects of aging. Brain Research, 163, 195-205. Jenkins, I. H., Brooks, D. J., Nixon, P. D., Frackowiak, R. S. J., & Passingham, R. E. (1994). Motor sequence learning: A study with positron emission tomography. The Journal of Neuroscience, 14, 37743790. Jonides, J., Marshuetz, C., Smith, E. E., Reuter-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 Neuroscience, 12, 188-196. Kimberg, D. Y., & Farah, M. J. (1993). A unified account of cognitive impairments following frontal lobe damage: The role of working memory in complex, organized behavior. Journal of Experimental Psychology: General, 122, 411-428. Kimberg, D. Y., D'Esposito, M., & Farah, M. J. (1997). Effects of bromocriptine on human subjects depend on working memory capacity. NeuroReport, 8, 3581-3585. Kirsner, K. (1972). Developmental changes in short-term recognition memory. British Journal of Psychology, 63, 109-117. Kriauciunas, R. (1968). The relationship of age and retention interval activity in short term memory. Journal of Gerontology, 23, 169-173. Kubota, K., & Niki, H. (1971). Prefrontal cortical unit activity and delayed alternation performance in monkeys. Journal of Neurophysiology, 34, 337-347. Lane, R. D., Reiman, E. M., Ahern, G. L., Schwartz, G. E., & Davidson, R. J. (1997). Neuroanatomical correlates of happiness, sadness, and disgust. American Journal of Psychiatry, 154, 926-933. Madden, D. J., Turkington, T. G., Provenzale, J. M., Denny, L. L., Hawk, , T. C., Gottlob, L. R., & Coleman, R. E. (1999). Adult age differences in the functional neuroanatomy of verbal recognition memory. Human Brain Mapping, 7, 115—135. Marsh, G. R. (1975). Age differences in evoked potential correlates of a memory scanning process. Experimental Aging Research, 1, 3-16. Meyer, C. H., Hu, B. S., Nishimura, D. G., & Macovski, A. (1992). Fast spiral coronary artery imaging. Magnetic Resonance in Medicine, 28, 202.

383

Murdock, B. B. (1967). Recent developments in short-term memory. British Journal of Psychology, 58, 421-433. Nelson, H., & Willison, J. (1991). National Adult Reading Test: 2nd Edition. Windsor, England: NFER-Nelson. Nielsen-Bohlman, L., & Knight, R. T. (1995). Prefrontal alterations during memory processing in aging. Cerebral Cortex, 5, 541-549. Nolde, S. F., Johnson, M. K., & D'Esposito, M. (1998). The role of prefrontal cortex during tests of episodic memory. NeuroReport, 9, 3509-3514. Noll, D. C., Cohen, J. D., Meyer, C. H., & Schneider, W. (1995). Spiral k-space MRI of cortical activation. Journal of Magnetic Resonance Imaging, 5, 49-56. Norman, D., & Shallice, T. (1980). Attention to action: Willed automatic control of behavior (CHIP Report 99). San Diego: University of California. Paradise, S., Robinson, R. G., Andreasen, N. C., Downhill, J. E., Davidson, R. J., Kirchner, P. T., Watkins, G. L., Ponto, L. L., & Hichwa, R. D. (1997). Emotional activation of limbic circuitry in elderly normal subjects in a PET study. American Journal of Psychiatry, 154, 384-389. Paulesu, E., Frith, C., & Frackowiak, R. (1993, March). The neural correlates of the verbal component of working memory. Nature, 362, 342-345. Peters, A., Rosene, D. L., Moss, M. B., Kemper, T. L., Abraham, C. R., Tigges, J., & Albert, M. S. (1996). Neurobiological bases of age-related cognitive decline in the rhesus monkey. Journal of Neuropathology and Experimental Neurology, 55, 861-874. Petrides, M. (1996). Lateral frontal cortical contribution to memory. Seminars in the Neurosciences, S, 57-63. Petrides, M., Alivisatos, B., Evans, A., & Meyer, E. (1993). Dissociation of human mid-dorsolateral from posterior dorsolateral frontal cortex in memory processing. Proceedings of the National Academy of Sciences of the United States of America, 90, 873-877. Petrides, M., Alivisatos, B., Meyer, E., & Evans, A. (1993). Functional activation of the human frontal cortex during the performance of verbal working memory tasks. Proceedings of the National Academy of Sciences of the United States of America, 90, 878-882. Poon, L. W., & Fozard, J. L. (1980). Age and word frequency effects in continuous recognition memory. Journal of Gerontology, 35, 77—86. Prabhakaran, V., Narayanan, K., Zhao, Z., & Gabrieli, J. D. E. (2000). Integration of diverse information in working memory within the frontal lobe. Nature—Neuroscience, 3, 85-90. Prabhakaran, V., Rypma, B., & Gabrieli, J. D. E. (2001). Neural substrates of mathematical reasoning: A functional magnetic resonance imaging study of neocortical activation during performance of the Necessary Arithmetic Operations Test. Neuropsychology, 15, 115—127. Prabhakaran, V., Smith, J. A. L., Desmond, J. E., Glover, G. H., & Gabrieli, J. D. E. (1997). Neural substrates of fluid reasoning: An fMRI study of neocortical activation during performance of the Raven's Progressive Matrices Test. Cognitive Psychology, 33, 43-63. Presty, S. K., Bachevalier, J., Walker, L. C., Struble, R. G., Price, D. L., Mishkin, M., & Cork, L. C. (1987). Age differences in recognition memory of the rhesus monkey (Macaco mulatto). Neurobiology of Aging, 8, 435-440. Raz, N., Gunning-Dixon, F. M., Head, D., Dupuis, J. H., McQuain, J., Briggs, S. D., & Acker, J. D. (1997). Selective aging of the human cerebral cortex observed in vivo: Differential vulnerability of the prefrontal gray matter. Cerebral Cortex, 7, 268-282. Reuter-Lorenz, P. A., Jonides, J., Smith, E. E., Hartley, A., Miller, A., Marshuetz, C., & Koeppe, R. A. (2000). Age differences in the frontal lateralization of verbal and spatial working memory revealed by PET. Journal of Cognitive Neuroscience, 12, 174-187. Ross, M. H., Yurgelun-Todd, D. A., Renshaw, P. F., Maas, L. C., Mendelson, J. H., Mello, N. K., Cohen, B. M., & Levin, J. M. (1997).

384

RYPMA, PRABHAKARAN, DESMOND, AND GABRIELI

Age-related reduction in functional MRI response to photic stimulation. Neurology, 48, 173-176. Rypma, B., DeBell, M. A., Gabrieli, J. D. E., Prabhakaran, V., Zabinski, M. F., Desmond, J. E., & Glover, G. H. (1996). Functional MRI studies of mental rotation and object identification processes. Society for Neuroscience Abstracts, 22, 720. Rypma, B., & D'Esposito, M. (1999). The roles of prefrontal brain regions in components of working memory: Effects of memory load and individual differences. Proceedings of the National Academy of Sciences, 96, 6558-6563. Rypma, B., & D'Esposito, M. (2000). Isolating the neural mechanisms of age-related changes in human working memory. Nature-Neuroscience, 3, 509-515. Rypma, B., Prabhakaran, V., Desmond, J. E., Glover, G. H., & Gabrieli, J. D. E. (1999). Load-dependent roles of prefrontal cortical regions in the maintenance of working memory. Neurolmage, 9, 216-226. Salthouse, T. A., & Babcock, R. L. (1991). Decomposing adult age differences in working memory. Developmental Psychology, 27, 763776. Schacter, D. L., Savage, C. R., Alpert, N. M., Rauch, S. L., & Albert, M. S. (1996). The role of hippocampus and frontal cortex in age-related memory changes: A PET study. NeuroReport, 7, 1165-1169. Servan-Schreiber, D., Printz, H., & Cohen, J. D. (1990, August). A network model of catecholamine effects: Gain, signal-to-noise ratio, and behavior. Science, 249, 892-895. Smith, A. D. (1975). Aging and interference with memory. Journal of Gerontology, 30, 319-325. Stebbins, G. T., Gabrieli, J. D. E., Carrillo, M. C., Desmond, J. E., Masciari, F., Turner, D., & Glover, G. H. (1997). A fMRI study of frontal activation during semantic encoding and verbal working memory. Society for Neuroscience Abstracts, 23, 2110.

Steinberg, S. (1966, August 5). High-speed scanning in human memory. Science, 153, 652-654. Talairach, J., & Tournoux, P. (1988). A co-planar stereotaxic atlas of the human brain: An approach to medical cerebral imaging. New York: Thieme Medical Publishers. Taoka, T., Iwasaki, S., Uchida, H., Fukusumi, A., Nakagawa, H., Kichikawa, K., Takayama, K., Yoshioka, T., Takewa, M., & Ohishi, H. (1998). Age correlation of the time lag in signal change on EPI-fMRI. Journal of Computer Assisted Tomography, 22, 514-517. Taub, H. A. (1973). Memory span, practice and aging. Journal of Gerontology, 28, 335-338. Tulving, E., & Colotla, V. (1970). Free recall of trilingual lists. Cognitive Psychology, 1, 86-98. Vallar, G., & Baddeley, A. D. (1984). Fractionation of working memory: Neuropsychological evidence for a phonological short-term store. Journal of Verbal Learning and Verbal Behavior, 23, 151-161. Waugh, N. C., & Norman, D. A. (1965). Primary memory. Psychological Review, 72, 89-104. Wechsler, D. (1981). Manual for the Wechsler Adult Intelligence Scale— Revised. New York: The Psychological Corporation. Wiegersma, S., & Meertse, K. (1990). Subjective ordering, working memory and aging. Experimental Aging Research, 16, 73—77. Zacks, J., Rypma, B., Gabrieli, J. D. E., Tversky, B., & Glover, G. H. (1999). Imagined transformations of bodies: An fMRI investigation. Neuropsychologia, 37, 1029-1040.

Received December 30, 1999 Revision received June 16, 2000 Accepted October 18, 2000

Low Publication Prices for APA Members and Affiliates Keeping you up-to-date. All APA Fellows, Members, Associates, and Student Affiliates receive—as part of their annual dues—subscriptions to the American Psychologist and APA Monitor. High School Teacher and International Affiliates receive subscriptions to the APA Monitor, and they may subscribe to the American Psychologist at a significantly reduced rate. In addition, all Members and Student Affiliates are eligible for savings of up to 60% (plus a journal credit) on all other APA journals, as well as significant discounts on subscriptions from cooperating societies and publishers (e.g., the American Association for Counseling and Development, Academic Press, and Human Sciences Press). Essential resources. APA members and affiliates receive special rates for purchases of APA books, including the Publication Manual of the American Psychological Association, and on dozens of new topical books each year. Other benefits Of membership. Membership in APA also provides eligibility for competitive insurance plans, continuing education programs, reduced APA convention fees, and specialty divisions. More information. Write to American Psychological Association, Membership Services, 750 First Street, NE, Washington, DC 20002-4242.