Abstract

It is controversial whether the effects of aging on various cognitive functions have the same common cause or several different causes. To investigate this issue, we scanned younger and older adults with functional magnetic resonance imaging (fMRI) while performing three different tasks: working memory, visual attention and episodic retrieval. There were three main results. First, in all three tasks, older adults showed weaker occipital activity and stronger prefrontal and parietal activity than younger adults. The occipital reduction is consistent with the view that sensory processing decline is a common cause in cognitive aging, and the prefrontal increase may reflect functional compensation. Secondly, older adults showed more bilateral patterns of prefrontal activity than younger adults during working memory and visual attention tasks. These findings are consistent with the Hemispheric Asymmetry Reduction in Older Adults (HAROLD) model. Finally, compared to younger adults, older adults showed weaker hippocampal formation activity in all three tasks but stronger parahippocampal activity in the episodic retrieval task. The former finding suggests that age-related hippocampal deficits may have a global effect in cognition, and the latter is consistent with an age-related increase in familiarity-based recognition. Taken together, the results indicate that both common and specific factors play an important role in cognitive aging.

Introduction

As we age, we become less able to manage several pieces of information at the same time, to filter distracting internal and external stimuli, and to remember what we did and what we have to do. Behavioral studies have carefully described and measured these age-related deficits in working memory, attention, episodic memory, and other cognitive functions (for reviews, see Craik and Salthouse, 2000). A basic issue is whether these various kinds of age-related cognitive deficits have the same common cause or several different causes. Thus far, this issue has been investigated primarily using statistical analyses of behavioral data. In the present study, we approached this problem from the new perspective provided by functional neuroimaging techniques.

A long-standing controversy in the cognitive aging literature is the one between common and specific factor theories. According to common factors theories, age-related sensory and cognitive changes can be accounted for by a common, biologically based factor (Lindenberger and Baltes, 1994; Anstey et al., 2001; Christensen et al., 2001). Common factor theories are attractive because they provide a parsimonious account of a large amount of empirical findings. At the same time, it has been noted that common causes cannot easily account for all manifestations of sensory and cognitive aging (Lindenberger and Baltes, 1994; Li and Lindenberger, 2002). There is much evidence that age effects can vary considerably across tasks (see Light, 1991), which may be more in line with specific factor theories.

From a cognitive neuroscience perspective, it is reasonable to assume that both common and specific factors are necessary for a full account of aging on sensory and cognitive functioning. On one hand, some forms of cerebral aging are likely to affect most cognitive tasks regardless of their nature (task-independent age effects). For example, age-related deficits in white matter integrity (for a review, see Raz, 2000) may reduce conduction speed and slow down responses in most cognitive tasks (Salthouse, 1996). On the other hand, other forms of cerebral aging are likely to affect only certain kinds of tasks (task-specific age effects). For instance, age-related deficits in dorsolateral prefrontal cortex (PFC) function are likely to have a greater impact on tasks that are highly dependent on executive functions (for reviews, see Moscovitch and Winocur, 1992; West, 1996), such as context memory (e.g. Cabeza et al., 2000). Thus, rather than trying to decide between common and specific factor theories, our more modest goal was to distinguish between task-independent and task-dependent age effects on brain activity using functional neuroimaging (cf. Esposito et al., 1999; Grady, 2002).

If one reviews functional neuroimaging studies of cognitive aging (e.g. Cabeza, 2001a,b), one can identify at least two consistent task-independent age effects. First, several studies have found an age-related decrease in occipital activity across a variety of tasks, including face matching (Grady et al., 1994), lexical decision (Madden et al., 1996), word-pair encoding (Cabeza et al., 1997) and retrieval (e.g. Anderson et al., 2000), temporal-order memory (e.g. Cabeza et al., 2000) and letter working memory (e.g. Rypma et al., 2001). Age-related decreases in occipital cortex activity have been attributed to inefficient sensory processing in the ventral (occipito-temporal) pathway (Grady et al., 1994; Madden et al., 1996). Second, several studies have found an age-related increase in PFC activity across a variety of tasks, such as face matching (e.g. Grady et al., 1994), lexical decision (Madden et al., 1996), word recognition (Madden et al., 1999), verbal working memory (e.g. Rypma and D’Esposito, 2000) and problem-solving (Esposito et al., 1999). Age-related increases in PFC activity have been typically attributed to functional compensation (e.g. Grady et al., 1994; Madden et al., 1996; Rypma and D’Esposito, 2000). Relating both types of effects, Grady et al. (1994) suggested that during perception older adults might compensate for deficits in sensory processes mediated by occipital regions by recruiting strategic processes mediated by PFC regions. Similarly, Li and Lindenberger (2002) noted that the results from a number of studies suggest that older adults use cognitive processes to compensate for compromised sensory information.

When reviewing functional neuroimaging studies of cognitive aging, one can also identify task-specific age effects. For example, many studies have found an age-related contralateral PFC recruitment (for a review, see Cabeza, 2002), that is, older adults show greater PFC activity in the side of PFC that was less activated in younger adults. PFC lateralization differs across tasks, and age-related contralateral PFC recruitment does too. For example, in tasks associated with left PFC activity in young adults, such as episodic encoding, older adults show contralateral recruitment in right PFC (e.g. Logan and Buckner, 2001; Morcom et al., 2003), whereas in tasks associated with right PFC activity in younger adults, older adults shows contralateral recruitment of left PFC (Bäckman et al., 1997; Cabeza et al., 1997; Madden et al., 1999; Grady et al., 2002).

The main problem with the foregoing list of task-independent and task-specific age effects is that it is based on cross-experiment comparisons. In the case of task-independent age effects, it is unclear if the putative common region was actually the same area of the brain in all studies. In the case of task-specific age effects, it is uncertain if the different age effects reflected the nature of the tasks employed or other methodological differences in the methods employed (participants, stimuli, procedures, etc.). In order to clearly distinguish between task-independent and task-specific age effects on brain activity, it is critical to compare age-related differences on brain activity on several different tasks, within-subjects and under similar conditions. This was the overall goal of the present study.

Extending, to aging, a cross-function approach (Cabeza and Nyberg, 2002) that we have previously investigated in young adults (Cabeza et al., 2002a, 2003b; Nyberg et al., 2002, 2003), we employed event-related fMRI to directly compare the effects of aging on the neural correlates of three cognitive tasks: working memory (WM), visual attention (VA) and episodic retrieval (ER). As illustrated by Figure 1, trials had two phases in all tasks. During the first phase of WM trials participants encoded and sustained a memory set of four words, and during the second phase they indicated if a probe word was part of the memory set or a new word. During the first phase of VA trials participants sustained attention to a symbol on the screen to determine if it blipped once, twice, or never during a 12 s interval, and during the second phase they indicated the presence or absence of blips. In most VA trials there was no blip during Phase 1, and only these no-blip trials were included in the fMRI analyses. Given that in the trials analyzed participants stared at the screen but nothing happened, the working memory component of the VA condition is minimal. During the first phase of ER trials participants generated the mental set of ER (retrieval mode; Tulving, 1983), and during the second phase they made the decision about whether or not a cue word was studied before the scanning session.

The analyses of fMRI data identified task-independent and task-specific age effects. To identify task-independent age effects, we used a conjunction procedure that isolated age-related differences in activity that occurred in each and every cognitive task. On the basis of prior findings, we predicted that all three tasks would show age-related decreases in occipital activity and age-related increases in PFC activity. To identify task-specific age effects, we used a masking procedure that excluded from the age-related differences in activations of each task those previously identified as task-independent (conjunction analysis). Given that in younger adults, PFC activity tends to be left lateralized for verbal WM (for a review, see Smith and Jonides, 1999) and right lateralized for sustained VA (for reviews, see Coull, 1998; Sarter et al., 2001), we predicted age-related contralateral recruitment in right PFC for WM but in left PFC for VA.

Materials and Methods

Subjects

The subjects were 20 younger adults (13 males) and 20 older adults (10 males). All participants were healthy, right-handed, native English speakers, with no history of neurological or psychiatric episodes. No participant was taking medications or had a medical condition that could affect cerebral blood flow (e.g. hypertension). All participants gave informed consent to a protocol approved by Duke University Institutional Review Board. Younger adults were Duke University students and staff with a mean age of 22.6 years (SD = 3.7). Older adults were community dwelling high-functioning individuals with a mean age of 70.3 years (SD = 6.3). Older adults were selected from a pool of senior research volunteers at Duke University and screened for health problems and conditions that could affect blood flow (e.g. hypertension, certain medications) using a questionnaire, and for signs of mild cognitive impairment or dementia using a test battery. The selected group of older participants performed within the normal range in the Mini-mental Status Exam (mean = 28.9) and in all other tasks in the battery.

Materials

The critical materials were concrete words selected from the MRC Psycholinguistic Database (http://www.psy.uwa.edu.au/mrcdatabase/mrc2.html). The words were four to six letters in length and of moderate frequency. Half of the words referred to living things and half to nonliving things.

Procedure

After completing health and MRI screening questionnaires and practicing the tasks to be performed in the scanner, subjects were placed in the scanner and anatomical scans were conducted. Following the anatomical scans and before the functional scans, subjects studied a list of 40 words (36 targets, two primacy fillers, two recency fillers), presented at a rate of 3 s/word. Subjects made a living/nonliving decision for each word and were also instructed to remember the words for a subsequent memory test. In the scanner (1.5 T GE LX Nvi), all stimuli were projected using an LCD projector to a screen located ∼70 cm behind the subjects’ crown, which subjects could see via an angled mirror. When necessary, vision was corrected using MRI-compatible lenses that matched the prescription used by the participant. Responses were recorded using a three-button MR-compatible response box. During functional scanning, subjects performed WM, VA and ER trials in random order. Each trial lasted 30 s and had two phases, each consisting of a stimulus (3 s) followed by fixation (12 s). In all three tasks, subjects made a three-choice response to the second stimulus, which was always a single letter string. They were encouraged to respond while the word was on the screen (3 s) and trials with RTs beyond this interval were excluded from the analyses.

In WM trials, the first stimulus was a memory set of four words, presented in two columns, and the second stimulus was a probe word which subjects recognized as being in one of the two columns in the memory set or as a new word. In VA trials, the first stimulus was the instruction to perform the VA task, and the second stimulus was a response screen. Following the instruction screen, subjects had to stare continuously at the fixation symbol (a letter ‘B’) in order to determine if it ‘blipped’ (a brief disappearance) once, twice, or never during the 12 s interval. When the response screen was presented, subjects entered their response (once, twice, never). Only no-blip trials were included in the analyses. In ER trials, the first stimulus was the instruction to perform the ER task, and the second stimulus was the cue word. Subjects responded to the cue word by indicating whether they remembered having read the word in the study list before scanning (‘Remember’ response), whether they believed the word was in the study list but could not retrieve any specific detail about its occurrence within the list (‘Know’ response), or whether they thought the word was not included in the study list (‘New’ response). The Remember–Know paradigm was not included to compare Remember and Know trials, which was precluded by the total number of old words scanned (36), but to encourage subjects to use a recollection-based retrieval strategy.

Functional scanning consisted of 12 runs, and each run included nine critical trials: three WM trials with words from the memory set, three VA trials with no blips, and three ER trials with studied words. Additionally, each run contained several ‘catch trials’ which were not included in the analyses: an average of 0.5 WM trials with new words, 0.5 VA trials with blips, and 1.0 ER trial with new words. Across the 12 runs, there were a total of 36 old-word WM trials, 36 no-blip VA trials, and 36 old-word ER trials. Only trials in which the word was correctly classified as part of the intra-trial memory set (WM) trials, the pre-scanning study list (ER trials), or in which the absence of blips was correctly noted (VA trials) were included in the analyses. With accuracy levels of ∼86–97%, subjects contributed ∼30–35 trials to each condition.

fMRI Methods

Anatomical Scanning

A T1-weighted sagittal localizer series was first acquired. The anterior (AC) and posterior commissures (PC) were identified in the mid-sagittal slice, and 34 contiguous oblique slices were prescribed parallel to the AC–PC plane. High-resolution T1-weighted structural images were acquired with a 450 ms TR (repetition time), a 9 ms TE (echo time), a 24 cm FOV (field of view), a 256 × 256 matrix, and a slice thickness of 3.75 mm. A second series of 46 oblique T1-weighted images perpendicular to the AC-PC was then acquired using the same imaging parameters.

Functional Scanning

Thirty-four contiguous gradient-echo echoplanar images (EPIs) sensitive to blood-oxygen level dependent (BOLD) contrast were acquired parallel to the AC–PC plane, using the same slice prescription described above for the near-axial structural images. The EPIs were acquired with a 3 s TR, 40 ms TE, one radio frequency excitation, 24 cm FOV, 64 × 64 image matrix, and a 90° flip angle. Slice thickness was 3.75 mm, resulting in cubic 3.75 mm3 isotropic voxels.

Image Preprocessing

All image preprocessing and statistical analyses were performed using SPM99 (http://www.fil.ion.ucl.ac.uk/spm/). Time series were corrected for differences in slice acquisition times, and realigned to correct for motion artifacts. Anatomical images were coregistered with the first functional images for each subject, and then both anatomical and functional images were spatially normalized to a standard stereotactic space, using the Montreal Neurological Institute (MNI) templates implemented in SPM99. The coordinates were later converted to Talairach and Tournoux (1988) space using software available online (http://www.mrc-cbu.cam.ac.uk/imaging/mnispace/html). Subsequently, the functional images were spatially smoothed using an 8 mm isotropic Gaussian kernel. They were proportionally scaled to the whole-brain signal, which was not significantly correlated with any of the activations identified by subsequent statistical contrasts.

Statistical Analyses

Phase 1 and Phase 2 were modeled separately. For each subject, task-related activity was identified by a convolving vector of the onset times of the stimuli with a synthetic hemodynamic response function (HRF) and its temporal derivative. The general linear model, as implemented in SPM99, was used to model the effects of interest and other confounding effects (e.g. head movement, and magnetic field drift). We assessed statistical parametric maps (SPMs) for each individual subject by applying linear contrasts to the parameter estimates for the events of interest, resulting in a t-statistic for every voxel. Trials associated with correct and incorrect responses were modeled separate, and only the results for the former are reported. Group averages and group comparisons were calculated for each condition by employing a one-sample t-test without constant term (random effects) on the resulting contrast images, in which the variances of the young and old groups were pooled. This type of analysis allows inclusive masking and conjunction analyses over the effects of interest (see below).

Age-independent Activations

These were first identified for ER, VA, and WM by performing conjunction analyses over the voxels that were activated in both groups as compared to an implicit baseline, which were thresholded at P < 0.0005, uncorrected with a cluster size > 5.

Task-independent Age Effects

These were assessed at an uncorrected threshold of P < 0.001. First, we calculated group differences for each task at a threshold of P < 0.1, uncorrected. Next, to make sure that these group differences were not due to deactivations, we masked them (inclusively) with the voxels that were activated in each group compared to the baseline at P < 0.05, uncorrected. Finally, the conjunction of the group differences (global age effects) was calculated by multiplying the resulting t-maps. The probability of finding a voxel that is independently activated in each and all contrasts, i.e. the joint probability, can be estimated by multiplying the probabilities for each contrast (e.g. Allan et al., 2000): 0.1 × 0.1 × 0.1 = P < 0.001, uncorrected.

Task-specific Age Effects

Thesewere assessed by thresholding the group differences at P < 0.001, uncorrected with a cluster size > 5. Next, in order to remove the regions that showed common age differences across tasks, we not only masked them (inclusively, P < 0.05, uncorrected) with the voxels that were activated in each group relative to baseline, but additionally, we used the t-map comprising the conjunction of group differences as an exclusive mask.

Results

Behavioral Results

The proportion of correct responses in WM, VA, and ER tasks are listed in Table 1. A 2 (group: younger, older) × 3 (task: WM, VA, ER) analysis of variance (ANOVA) on accuracy data, yielded significant effects of task [F(2,38) = 11.1, P < 0.0001], but no reliable main effect of age [F(1,38) = 3.0, P > 0.05] or age × task interaction [F(2,38) = 2.6, P > 0.05]. The task effect reflected greater accuracy in the VA task than in the WM task [t(39) = 3.7, P < 0.001] and ER tasks [t(39) = 4.2, P < 0.0001], with no difference between WM and ER tasks (t < 1). In sum, accuracy was greater in the VA task, but similar for young and older adults.

For the ER task, we conducted additional analyses on corrected recognition scores and Remember/Know responses. The proportion of false alarms was 0.26 in younger adults and 0.22 in older adults. Neither false alarms nor corrected recognition scores (hits–false alarms) differed across groups (t < 1 for both). The proportions of Remember and Know responses in the ER task were 0.77 and 0.23, respectively, for younger adults, and 0.65 and 0.35, respectively, for older adults. Thus, older adults showed less Remember responses and more Know responses than younger adults [t(38) = 2.0, P < 0.05]. This result suggests that during ER older adults relied less on recollection and more on familiarity than younger adults.

RTs in WM, VA, and ER tasks are listed in Table 1. A 2 (group: younger, older) × 3 (task: WM, VA, ER) ANOVA on RTs yielded significant effects of age [F(1,38) = 13.929, P < 0.001] and task [F(2,38) = 208.0, P < 0.0001], as well as a reliable age × task interaction [F(2,38) = 15.8, P < 0.0001]. The effect of age reflected faster responses for younger than for older adults, and the effect of task, faster responses in the VA task than in the WM task [t(39) = 12.9, P < 0.0001] and ER tasks [t(39) = 13.8, P < 0.0001], and faster responses in the WM task than in the ER task [t(39) = 6.965, P < 0.0001]. The age × task interaction reflected a more pronounced age effect for ER than for WM and VA, and for WM than for VA. In sum, older adults were slower than younger adults, particularly in the ER task.

Imaging Results

Age-independent Activations

Table 2 shows brain regions where both younger and older adults showed significant activations in each of the three tasks. The role of each of these regions in each of the three tasks has been previously discussed (Cabeza et al., 2002a, 2003b) and is beyond the scope of the present report, which focuses on the effects of aging. However, it is important to emphasize the despite the existence of significant age-related differences in activation, activation patterns in healthy older adults were remarkably similar to those displayed by young adults. As shown by Table 2, the two groups showed activations in many of the same brain regions. In addition to supporting the reliability of our data, this finding suggests that the effects of healthy aging on the neural correlates of cognition are relatively modest, leaving untouched a large proportion of the neural activity underlying fundamental cognitive functions such as WM, VA, and ER.

Task-independent Age Effects

Table 3 lists brain regions that showed significant age-related differences in activity in all three tasks. Consistent with our predictions, older adults showed weaker activity in occipital regions and stronger activity in PFC regions in all tasks (see Fig. 2). Age-related decreases in occipital activity were found in lateral (BA 19, 18) and medial (BA 31) extrastriate cortex, and age-related increases in PFC activity were found both left (BA 6) and right (BA 44/45) prefrontal regions. Beyond our predictions, interesting task-independent age effects were found in the hippocampal formation (subicular region), which was more activated in younger than in older adults, and in parietal regions, which were more activated in older than in younger adults.

Task-specific Age Effects

Table 4 lists brain regions that showed significant age-related differences in activity that were specific to WM, VA, or ER tasks. Consistent with our predictions, age-related contralateral PFC recruitment was found in right PFC for WM and in left PFC for VA (see Fig. 3). The age-related increase in left PFC activity during VA was identified using a more lenient threshold (P < 0.01). Beyond our predictions, an interesting task-specific age effect was in bilateral parahippocampal regions, which during ER were more activated in older than in younger adults (see Fig. 4). Other task-specific age effects included age-related decreases in striate cortex (BA 17) activity during WM, and in posterior cingulate activity during ER.

Discussion

The study yielded three main findings. First, consistent with our predictions, task-independent age effects included an age-related reduction in occipital activity coupled with an age-related increase in PFC activity. Secondly, also consistent with our predictions, task-specific age effects included age-related contralateral recruitments in right PFC during WM and in left PFC during VA. Finally, there was a dissociation between two medial temporal lobe (MTL) regions: the hippocampal formation, which showed weaker activity in older adults across all tasks, and the parahippocampal gyrus, which showed stronger activity in older adults during ER. Below, we discuss these three main results.

Age-related Decreases in Occipital Activity versus Age-related Increases in PFC

As illustrated by Figure 2, task-independent age effects included age-related decreases in occipital activity coupled with age-related increases in PFC activity. The co-occurrence of these two age effects extends to verbal working memory, visual sustained attention, and verbal episodic recognition memory tasks a finding originally reported by Grady et al. (1994) using a face matching task (see also Grady, 2002). Importantly, the present study demonstrates that exactly the same occipital and PFC regions can display identical age-related changes in activity across different cognitive tasks.

These task-independent findings are consistent with common factor theories of aging, although as discussed below, inconsistent task-specific age effects were also observed. Weaker visual cortex activity in older adults across all tasks fits well with the theory that age-related sensory processing plays a major role in cognitive aging (Baltes and Lindenberger, 1997). As noted before, this theory is supported by evidence of strong correlations between age-related sensory/perceptual decline and age-related cognitive decline (e.g. Lindenberger and Baltes, 1994; Baltes and Lindenberger, 1997). There are three main explanations for the close link between sensory and cognitive deficits in older adults (Lindenberger and Baltes, 1994; Baltes and Lindenberger, 1997; Schneider and Pichora-Fuller, 2000): (i) sensory deficits cause cognitive deficits (sensory deprivation hypothesis); (ii) cognitive deficits cause sensory deficits (cognitive load on perception hypothesis); and (iii) sensory and cognitive deficits are both caused by a third underlying mechanism (common cause). The age-related decrease in visual cortex activity seems more consistent with the sensory deprivation hypothesis (Lindenberger and Baltes, 1994), but predicting activation data from the three views is not straightforward.

As for task-independent age-related increases in PFC activity, they suggest that some forms of functional compensation in older adults may be common across tasks. Given that lateral PFC regions are recruited by a variety of cognitive tasks (Cabeza and Nyberg, 2000, 2002), it is not surprising to find task-independent age effects in these areas. As mentioned before, Grady et al. (1994) suggested that age-related increases in PFC activity could compensate for age-related decreases in visual cortex activity (Grady et al., 1994; Madden et al., 1996). Although we did not find significant correlations between PFC activations and older adults’ performance, we did find a reliable negative correlation between the effect size of age-related decrease in occipital activity (x, y, z = –23, –84, –8) and age-related increases in PFC activity (x, y, z = –45, –6, 38) in the WM task (r = –0.5, P < 0.02). In other words, in the WM condition, those older adults who showed the weakest occipital activations also showed in some cases the strongest PFC activations. This finding is consistent with Grady et al.’s hypothesis.

As shown in Table 3, older adults showed task-independent increases in activity not only in PFC but also in parietal regions. Age-related increases in parietal activity have been previously found during episodic memory (Anderson et al., 2000; Grady et al., 2002), but in the present study they were also found during WM and VA. Given that the recruitment of parietal regions occurred in combination with weaker activity in visual occipital regions, the results suggest an age-related shift from the ventral (occipitotemporal) stream to the dorsal (occipitoparietal) stream, which are pathways respectively associated with object and spatial processing (Ungerleider and Mishkin, 1982). Like PFC recruitment, parietal recruitment in older adults could reflect compensatory mechanisms. Another alternative is that the ventral/dorsal differentiation is less pronounced in older adults, an idea supported by behavioral (Chen et al., 2002) and functional neuroimaging evidence (Grady et al., 1994) and consistent with the notion of age-related dedifferentiation (Li and Lindenberger, 1999).

Age-related Contralateral PFC Recruitment

Not all age-related increases in PFC activity were task-independent; some were task-specific, including the recruitment of contralateral PFC regions. As illustrated by Figure 3, in the case of WM, older adults showed greater activity than younger adults in right PFC, whereas in the case of VA, older adults showed greater activity in left PFC during VA. Since PFC activity during ER in younger adults was largely bilateral, contralateral recruitment could not be observed. In contrast, in the case of WM and VA some PFC regions were lateralized in young adults, and contralateral recruitment yielded a more bilateral pattern of activity in older than in younger adults (see Fig. 3). This pattern, known as ‘Hemispheric Asymmetry Reduction in Older Adults’ or HAROLD, has been found in variety of cognitive tasks, including perception, episodic memory, semantic memory, working memory, language, and inhibition tests (for a review, see Cabeza, 2002). These studies show that when PFC activity is left lateralized in younger adults, older adults may recruit additional activity in right PFC and when PFC activity is right lateralized in younger adults, older adults may recruit additional activity in left PFC. As illustrated by Figure 3, we found both forms of contralateral recruitment in the same participants and within the same experiment (see also, Reuter-Lorenz et al., 2000).

Age-related contralateral PFC recruitment may be beneficial for cognitive performance (compensation view, Cabeza et al., 1997, 2002b; Reuter-Lorenz et al., 2000; Rosen et al., 2002), or it may reflect a difficulty in recruiting specialized neural mechanisms (dedifferentiation view, Li and Lindenberger, 1999; Logan et al., 2002). Recent studies comparing brain activity in high- and low-performing older adults have provided evidence supporting the compensation view (Cabeza et al., 2002b; Rosen et al., 2002). If one assumes this view, one may ask about the differential contributions of left versus right PFC recruitment to older adults’ cognitive performance. In the episodic retrieval domain, we have recently proposed that left PFC is differentially more involved than right PFC in semantically guided production processes, whereas right PFC is more involved than left PFC in monitoring processes (Cabeza et al., 2003a). If the production-monitoring distinction can be extended beyond the episodic retrieval domain — and there is preliminary evidence that it can (Cabeza et al., 2002a, 2003b), then the results in Figure 3 would suggest that older adults compensate for production deficits in left PFC by recruiting monitoring processes in right PFC, and vice versa.

Age-related Decrease in Hippocampal Activity and Increase in Parahippocampal Activity

The third main finding of the study was a dissociation between two MTL regions (see Fig. 4): the hippocampal formation, which showed decreased activity in older adults across all tasks, and the parahippocampal gyrus, which showed increased activity in older adults during ER. Age-related decreases in hippocampal activity have been previously found for episodic memory (e.g. Grady et al., 1995; Daselaar et al., 2003; Morcom et al., 2003) but not across several different cognitive functions. The present task-independent decrease suggests that a decline in hippocampal integrity during healthy (Raz et al., 1997) and pathological (e.g. Kohler et al., 1998; Du et al., 2001) aging could affect other cognitive functions besides episodic memory, including working memory and attention. Three recent fMRI studies found significant hippocampal activity during working memory tasks (Ranganath and D’Esposito, 2001; Stern et al., 2001; Cabeza et al., 2002a), suggesting that the indexing function of the hippocampus applies to both long-term and short-term memory representations (Cabeza et al., 2002a). Although it may be argued that hippocampal activity during WM reflected episodic memory encoding, this idea cannot easily account for the involvement of the hippocampus in the VA task. Given that only no-blip trials were included in the fMRI analyses, the memory component of the VA task was minimal. Participants had to maintain the goal of the task in working memory but this kind of sustained activity is unlikely to account for the event-related activations observed. In contrast, the hippocampal activation during VA fits very well with the idea that the hippocampus is involved in attentional orienting (e.g. Williams et al., 2000; Oswald et al., 2002) and novelty detection (e.g. Tulving et al., 1994; Knight, 1996). Thus, the age-related decrease in hippocampal activity across the three tasks could reflect an age-related deficit in attentional orienting and/or novelty detection processes.

Whereas the hippocampal formation showed weaker activity in older than in younger adults, a left parahippocampal region showed stronger activity in older than in younger adults (see Fig. 4). This dissociation is consistent with the combination of two types of evidence: evidence that these two MTL regions are differentially involved in recollection and familiarity, and evidence that recollection and familiarity are differentially affected by aging. The results of lesion and functional neuroimaging research (for reviews, see Aggleton and Brown, 1999; Yonelinas, 2002) have linked the hippocampus to recollection (retrieval of specific details) and the surrounding perirhinal/parahippocampal cortex to familiarity (feeling of knowing). In general, older adults tend to be impaired in recollection but not in familiarity (Parkin and Walter, 1992; Jennings and Jacoby, 1993; Mantyla, 1993; Java, 1996; Searcy et al., 1999). Moreover, in the Remember/Know paradigm, older adults have been found to show less ‘Remember’ responses (reduced recollection) but more ‘Know’ responses (increased familiarity) than younger adults (Parkin and Walter, 1992). This finding was replicated in the present study. Thus, it is reasonable to speculate that the present MTL dissociation reflected an age-related reduction in hippocampal-based recollection coupled with an age-related increase in parahippocampal-based familiarity. To test this idea, we calculated the correlation between the effect size of the age-related increase in parahippocampal activity (x, y, z = –26, –34, –11) and the number of Know responses in older adults. Consistent with the familiarity account, the correlation was positive and significant (R = 0.50, P < 0.05). Although further research is obviously required, this is probably the first evidence directly linking age-related differences in MTL activity to differential effects of aging on recollection and familiarity.

Conclusions

In summary, the study yielded three main results. First, consistent with our predictions, task-independent age effects included an age-related reduction in occipital activity coupled with an age-related increase in PFC. The former finding is consistent with the common factor view that age-related cognitive deficits are in great part due to a decline in sensory processing, and the latter finding suggests that some forms of compensatory PFC recruitment are common across tasks. Secondly, also consistent with our predictions, task-specific age effects included age-related contralateral recruitments in left PFC during WM and in right PFC during VA. This result suggests that older adults may compensate for deficits in production processes by recruiting monitoring processes, and vice versa. Finally, there was a dissociation between two MTL regions: the hippocampal formation, which showed decreased activity in older adults across all tasks, and the parahippocampal gyrus, which showed increased activity in older adults during ER. The task-independent decrease in hippocampal activity suggests that age-related deficits in hippocampal function may account for age-related cognitive decline beyond the episodic memory domain. The age-related increase in parahippocampal activity could reflect older adults’ reliance on familiarity-based responding. Thus, the present results indicate that both common factors and specific factors play an important role in cognitive aging.

We thank Amber Baptiste for participant recruitment, Dana Torpey for neuropsychological testing, the personnel at the Brain Imaging and Analysis Center of Duke University for assistance with data collection and analyses, and Dave Madden and Gregory McCarthy for support and encouragement. This research was supported by National Institute on Aging of the National Institutes of Health (1R01AG19731-01).

Address correspondence to Roberto Cabeza, Center for Cognitive Neuroscience, Duke University, Box 90999, LSRC Bldg, Rm B203, Durham, NC 27708, USA. Email: cabeza@duke.edu.

Figure 1. Behavioral paradigm. WM was investigated with a word delayed-response test, VA with a sustained attention task, and ER with a word recognition task.

Figure 1. Behavioral paradigm. WM was investigated with a word delayed-response test, VA with a sustained attention task, and ER with a word recognition task.

Figure 2. Task-independent age effects. Compared to younger adults, older adults showed weaker activity in occipital cortex but greater activity in frontal regions. The time-course plots are from left BA19 (x, y, z = –23, –84, –8) and right BA45 (x, y, z = 41, 12, 21).

Figure 2. Task-independent age effects. Compared to younger adults, older adults showed weaker activity in occipital cortex but greater activity in frontal regions. The time-course plots are from left BA19 (x, y, z = –23, –84, –8) and right BA45 (x, y, z = 41, 12, 21).

Figure 3. Task-specific age effects. Older adults showed contralateral recruitment in right PFC during WM and in left PFC during VA. As a result of these changes, activity in certain PFC regions was more bilateral in older than in younger adults (‘Hemispheric Asymmetry Reduction in Older Adults’ or HAROLD).

Figure 3. Task-specific age effects. Older adults showed contralateral recruitment in right PFC during WM and in left PFC during VA. As a result of these changes, activity in certain PFC regions was more bilateral in older than in younger adults (‘Hemispheric Asymmetry Reduction in Older Adults’ or HAROLD).

Figure 4. Dissociation between two medial temporal lobe regions. The hippocampal formation, bilaterally, was more activated in younger than in older adults during all tasks (task-independent age effect), whereas the parahippocampal gyrus, bilaterally, was more activated in older than in younger adults during the ER task (task-specific age effect). The time-course plots are from the left hippocampal formation (x, y, z = –19, –29, –2) and the left parahippocampal gyrus (x, y, z = –26, –34, –11).

Figure 4. Dissociation between two medial temporal lobe regions. The hippocampal formation, bilaterally, was more activated in younger than in older adults during all tasks (task-independent age effect), whereas the parahippocampal gyrus, bilaterally, was more activated in older than in younger adults during the ER task (task-specific age effect). The time-course plots are from the left hippocampal formation (x, y, z = –19, –29, –2) and the left parahippocampal gyrus (x, y, z = –26, –34, –11).

Table 1


 Behavioral results in WM, VA and ER tasks

 Younger  Older 
 Mean SD  Mean SD 
Accuracy (proportion correct)      
Working memory 0.92 0.1  0.91 0.9 
Visual attention 0.97 0.6  0.96 0.7 
Episodic memory 0.93 0.8  0.86 0.7 
RTs (ms)      
Working memory 1485 263  1776 338 
Visual attention  993 199   969 180 
Episodic memory 1711 250  2241 405 
 Younger  Older 
 Mean SD  Mean SD 
Accuracy (proportion correct)      
Working memory 0.92 0.1  0.91 0.9 
Visual attention 0.97 0.6  0.96 0.7 
Episodic memory 0.93 0.8  0.86 0.7 
RTs (ms)      
Working memory 1485 263  1776 338 
Visual attention  993 199   969 180 
Episodic memory 1711 250  2241 405 
Table 2


 Activations during WM, VA, and ER tasks shared by younger and older adults

Region of activation  Lat BA Ph Coordinates  t-value 
     x y z  Younger Older 
Working memory           
Occipital ctx Extrastriate 19 –49 –63  –9   5.5  3.3 
Striate ctx 17  15 –91  –2   5.4  4.3 
  17  11 –91  –2  12.0 16.0 
PFC  Ventrolateral 47 –34  23   6   4.0  4.6 
  47  49  18  –8   5.4  3.5 
   9 –45   5  31   3.5  6.1 
 Dorsolateral  6  –4   7  55   8.9  5.9 
   6  –4   4  56   5.0  4.7 
Parietal ctx  40 –41 –41  47   6.8  8.9 
Cerebellum    30 –56 –22   8.3  5.2 
   –38 –53 –23   6.3  5.9 
Thalamus   –11 –18   8  11.0  7.2 
    11 –14   8   5.7  6.7 
Lateral temporal ctx  22 –55   8  –4   5.9  5.9 
Visual attention           
Occipital (extrastriate) ctx  18  34 –83   1  10.6 11.1 
  19 –30 –87   8  10.7  9.9 
  18  34 –84  –3  11.6  8.1 
  18 –34 –88  –2   9.5  9.0 
PFC  Dorsolateral  –4  –1  49  12.1  9.6 
   45   2  45   7.3  6.5 
 Anterolateral 10  34  45  19   4.7  4.4 
Thalamus  – –15 –18   5  10.9  6.6 
  –  11 –14   8   7.4  5.5 
Caudate nucleus  – –11   4   7   4.5  4.4 
Episodic retrieval           
Occipital ctx Extrastriate 18 –30 –88  –2   8.6  5.6 
  18  30 –88  –2   8.9  6.4 
 Striate 17  11 –88  –2  11.0 14.0 
PFC  Ventrolateral 47  34  22  –8   5.5  6.3 
  47 –34  22  –4   5.4  5.1 
 Dorsolateral  6  –4  13  48   9.5  7.9 
   9 –49   9  35   5.0  6.5 
  46  49  24  26   3.4  4.7 
 Anterolateral 10 –34  55   7   4.7  3.9 
Thalamus  –  –8 –18   5   7.6  6.7 
Parietal ctx  40 –41 –31  43   4.6  7.0 
  –38 –60  48   4.3  7.4 
Cerebellum  –  34 –63 –25   6.0  4.5 
  – –34 –59 –22   6.9  3.6 
Parahippocampal gyrus  27 –23 –29  –5   8.3  3.3 
Region of activation  Lat BA Ph Coordinates  t-value 
     x y z  Younger Older 
Working memory           
Occipital ctx Extrastriate 19 –49 –63  –9   5.5  3.3 
Striate ctx 17  15 –91  –2   5.4  4.3 
  17  11 –91  –2  12.0 16.0 
PFC  Ventrolateral 47 –34  23   6   4.0  4.6 
  47  49  18  –8   5.4  3.5 
   9 –45   5  31   3.5  6.1 
 Dorsolateral  6  –4   7  55   8.9  5.9 
   6  –4   4  56   5.0  4.7 
Parietal ctx  40 –41 –41  47   6.8  8.9 
Cerebellum    30 –56 –22   8.3  5.2 
   –38 –53 –23   6.3  5.9 
Thalamus   –11 –18   8  11.0  7.2 
    11 –14   8   5.7  6.7 
Lateral temporal ctx  22 –55   8  –4   5.9  5.9 
Visual attention           
Occipital (extrastriate) ctx  18  34 –83   1  10.6 11.1 
  19 –30 –87   8  10.7  9.9 
  18  34 –84  –3  11.6  8.1 
  18 –34 –88  –2   9.5  9.0 
PFC  Dorsolateral  –4  –1  49  12.1  9.6 
   45   2  45   7.3  6.5 
 Anterolateral 10  34  45  19   4.7  4.4 
Thalamus  – –15 –18   5  10.9  6.6 
  –  11 –14   8   7.4  5.5 
Caudate nucleus  – –11   4   7   4.5  4.4 
Episodic retrieval           
Occipital ctx Extrastriate 18 –30 –88  –2   8.6  5.6 
  18  30 –88  –2   8.9  6.4 
 Striate 17  11 –88  –2  11.0 14.0 
PFC  Ventrolateral 47  34  22  –8   5.5  6.3 
  47 –34  22  –4   5.4  5.1 
 Dorsolateral  6  –4  13  48   9.5  7.9 
   9 –49   9  35   5.0  6.5 
  46  49  24  26   3.4  4.7 
 Anterolateral 10 –34  55   7   4.7  3.9 
Thalamus  –  –8 –18   5   7.6  6.7 
Parietal ctx  40 –41 –31  43   4.6  7.0 
  –38 –60  48   4.3  7.4 
Cerebellum  –  34 –63 –25   6.0  4.5 
  – –34 –59 –22   6.9  3.6 
Parahippocampal gyrus  27 –23 –29  –5   8.3  3.3 

Lat, lateralization; BA, Brodmann Area; Ph, phase; coordinates from Talairach and Tournoux (1988).

Table 3


 Age-related differences common across WM, VA and ER tasks

Region of activation  Lat BA Ph Coordinates  t-values 
     x y z  WM VA ER 
Younger > older            
Occip. (extrastriate) ctx Lateral 19 –45 –77  –5  4.3 2.8 3.1 
  18 –23 –84  –8  2.8 4.9 3.4 
  18  23 –81  –9  3.1 3.8 3.0 
 Medial 31  19 –62  10  4.9 3.9 4.7 
Anterior cingulate  32   0  20  30  1.8 4.1 2.5 
Precuneus    0 –16  46  2.7 2.4 1.5 
Hippocampal formation (subiculum)   –19 –29  –2  6.6 5.0 4.0 
    19 –29   1  4.8 4.8 3.4 
Motor ctx  –38 –23  53  5.4 4.9 3.8 
Cerebellum (hemisphere)    23 –48 –17  3.3 2.7 2.9 
Older > younger            
PFC  6/9 –45  –6  38  6.7 3.0 2.7 
  44/45  41  12  21  2.7 2.0 2.0 
Inferior parietal ctx  40 –45 –35  37  2.7 2.9 2.0 
Parieto-occipital ctx  19 –26 –75  42  2.0 2.3 2.0 
  19  30 –71  45  3.1 2.7 2.2 
Brain stem     4 –26  –8  2.3 3.7 3.2 
Cerebellum    15 –82 –21  2.4 2.2 2.7 
   –26 –56 –22  1.6 2.7 2.3 
Region of activation  Lat BA Ph Coordinates  t-values 
     x y z  WM VA ER 
Younger > older            
Occip. (extrastriate) ctx Lateral 19 –45 –77  –5  4.3 2.8 3.1 
  18 –23 –84  –8  2.8 4.9 3.4 
  18  23 –81  –9  3.1 3.8 3.0 
 Medial 31  19 –62  10  4.9 3.9 4.7 
Anterior cingulate  32   0  20  30  1.8 4.1 2.5 
Precuneus    0 –16  46  2.7 2.4 1.5 
Hippocampal formation (subiculum)   –19 –29  –2  6.6 5.0 4.0 
    19 –29   1  4.8 4.8 3.4 
Motor ctx  –38 –23  53  5.4 4.9 3.8 
Cerebellum (hemisphere)    23 –48 –17  3.3 2.7 2.9 
Older > younger            
PFC  6/9 –45  –6  38  6.7 3.0 2.7 
  44/45  41  12  21  2.7 2.0 2.0 
Inferior parietal ctx  40 –45 –35  37  2.7 2.9 2.0 
Parieto-occipital ctx  19 –26 –75  42  2.0 2.3 2.0 
  19  30 –71  45  3.1 2.7 2.2 
Brain stem     4 –26  –8  2.3 3.7 3.2 
Cerebellum    15 –82 –21  2.4 2.2 2.7 
   –26 –56 –22  1.6 2.7 2.3 

See Table 2 for abbreviations.

Table 4


 Age-related differences in brain activity specific to WM, VA, or ER tasks

Region of activation Lat BA Ph Coordinates Z-value 
    x y z  
Working memory        
Younger > older        
Occipital (striate) ctx 17 –11 –80  11 3.7 
Older > younger        
PFC   6 –26   6  51 4.0 
 46 –45  41  19 3.7 
  9  45  24  26 4.0 
Parietal ctx 40  41 –45  44 4.8 
  7  30 –74  45 3.8 
  7/40 –45 –56  48 3.6 
Parieto-occipital ctx 19 –34 –79  39 4.4 
Premotor ctx  8  23  13  44 4.8 
Somatosensory ctx  1/2/3 –45 –17  36 3.6 
Visual attention        
Younger > older        
No reliable differences        
Older > younger        
PFC 46  49  30  20 3.9 
 45 –34  23   6 2.4* 
Parietal ctx 40 –34 –42  37 3.8 
Episodic Retrieval        
Younger > older        
Posterior cingulate gyrus 23/31   0 –28  29 4.0 
Older > younger        
Parahippocampal gyrus 36  19 –37  –7 4.0 
 36 –26 –34 –11 4.2 
Lateral temporal ctx 21 –45  –9 –19 4.7 
 21  45  –5 –19 3.9 
Parieto-occipital ctx 19 –30 –71  31 5.0 
Thalamus – –11 –32  12 3.8 
Cerebellum –  –8 –41 –20 5.3 
Region of activation Lat BA Ph Coordinates Z-value 
    x y z  
Working memory        
Younger > older        
Occipital (striate) ctx 17 –11 –80  11 3.7 
Older > younger        
PFC   6 –26   6  51 4.0 
 46 –45  41  19 3.7 
  9  45  24  26 4.0 
Parietal ctx 40  41 –45  44 4.8 
  7  30 –74  45 3.8 
  7/40 –45 –56  48 3.6 
Parieto-occipital ctx 19 –34 –79  39 4.4 
Premotor ctx  8  23  13  44 4.8 
Somatosensory ctx  1/2/3 –45 –17  36 3.6 
Visual attention        
Younger > older        
No reliable differences        
Older > younger        
PFC 46  49  30  20 3.9 
 45 –34  23   6 2.4* 
Parietal ctx 40 –34 –42  37 3.8 
Episodic Retrieval        
Younger > older        
Posterior cingulate gyrus 23/31   0 –28  29 4.0 
Older > younger        
Parahippocampal gyrus 36  19 –37  –7 4.0 
 36 –26 –34 –11 4.2 
Lateral temporal ctx 21 –45  –9 –19 4.7 
 21  45  –5 –19 3.9 
Parieto-occipital ctx 19 –30 –71  31 5.0 
Thalamus – –11 –32  12 3.8 
Cerebellum –  –8 –41 –20 5.3 

See Table 2 for abbreviations. *P < 0.01.

References

Aggleton JP, Brown MW (
1999
) Episodic memory, amnesia, and the hippocampal-anterior thalamic axis.
Behav Brain Sci
 
22
:
425
–489.
Allan K, Dolan RJ, Fletcher PC, Rugg MD (
2000
) The role of the right anterior prefrontal in episodic retrieval.
Neuroimage
 
11
:
217
–227.
Anderson ND, Iidaka T, McIntosh AR, Kapur S, Cabeza R, Craik FIM (
2000
) The effects of divided attention on encoding- and retrieval-related brain activity: a PET study of yournger and older adults.
J Cogn Neurosci
 
12
:
775
–792.
Anstey KJ, Luszcz MA, Sanchez L (
2001
) A reevaluation of the common factor theory of shared variance among age, sensory function, and cognitive function in older adults.
J Gerontol B Psychol Sci Social Sci
 
56
:
3
–11.
Bäckman L, Almkvist O, Andersson J, Nordberg A, Windblad B, Rineck R, Lågström B (
1997
) Brain activation in young and older adults during implicit and explicit retrieval.
J Cogn Neurosci
 
9
:
378
–391.
Baltes PB, Lindenberger U (
1997
) Emergence of a powerful connection between sensory and cognitive functions across the adult life span: a new window to the study of cognitive aging?
Psychol Aging
 
12
:
12
–21.
Cabeza R (
2001
) Functional neuroimaging of cognitive aging. In: Handbook of functional neuroimaging of cognition (Cabeza R, Kingstone A, eds), pp.
331
–377. Cambridge, MA: MIT Press.
Cabeza R (
2001
) Cognitive neuroscience of aging: contributions of functional neuroimaging.
Scand J Psychol
 
42
:
277
–286.
Cabeza R (
2002
) Hemispheric asymmetry reduction in old adults: the HAROLD model.
Psychol Aging
 
17
:
85
–100.
Cabeza R, Nyberg L (
2000
) Imaging cognition II: an empirical review of 275 PET and fMRI studies.
J Cogn Neurosci
 
12
:
1
–47.
Cabeza R, Nyberg L (
2002
) Seeing the forest through the trees: the cross-function approach to functional neuroimaging. In: The cognitive electrophysiology of mind and brain (Zani A, Proverbio AM, eds), pp.
41
–68. San Diego, CA: Academic Press.
Cabeza R, Grady CL, Nyberg L, McIntosh AR, Tulving E, Kapur S, Jennings JM, Houle S, Craik FIM (
1997
) Age-related differences in neural activity during memory encoding and retrieval: a positron emission tomography study.
J Neurosci
 
17
:
391
–400.
Cabeza R, Anderson ND, Houle S, Mangels JA, Nyberg L (
2000
) Age-related differences in neural activity during item and temporal-order memory retrieval: a positron emission tomography study.
J Cogn Neurosci
 
12
:
1
–10.
Cabeza R, Dolcos F, Graham R, Nyberg L (
2002
) Similarities and differences in the neural correlates of episodic memory retrieval and working memory.
Neuroimage
 
16
:
317
–330.
Cabeza R, Anderson ND, Locantore JK, McIntosh AR (
2002
) Aging gracefully: compensatory brain activity in high-performing older adults.
Neuroimage
 
17
:
1394
–1402.
Cabeza R, Locantore JK, Anderson ND (
2003
) Lateralization of prefrontal cortex activity during episodic memory retrieval: evidence for the production-monitoring hypothesis.
J Cogn Neurosci
 
15
:
249
–259.
Cabeza R, Dolcos F, Prince S, Rice H, Weissman D, Nyberg L (
2003
) Attention-related activity during episodic memory retrieval: a cross-function fMRI study.
Neuropsychologia
 
41
:
390
–399.
Chen J, Myerson J, Hale S (
2002
) Age-related dedifferentiation of visuospatial abilities.
Neuropsychologia
 
40
:
2050
–2056.
Christensen H, Mackinnon AJ, Korten A, Jorm AF (
2001
) The ‘common cause hypothesis’ of cognitive aging: evidence for not only a common factor but also specific associations of age with vision and grip strength in a cross-sectional analysis.
Psychol Aging
 
16
:
588
–599.
Coull JT (
1998
) Neural correlates of attention and arousal: Insights from electrophysiology, functional neuroimaging and psychoparmacology.
Progr Neurobiol
 
55
:
343
–361.
Craik FIM, Salthouse TA (
2000
) Handbook of aging and cognition II. Mahwah, NJ: Erlbaum.
Daselaar SM, Veltman DJ, Rombouts SA, Lazeron RH, Raaijmakers JG, Jonker C (
2003
) Neuroanatomical correlates of episodic encoding and retrieval in young and elderly subjects.
Brain
 
126
:
43
–56.
Du AT, Schuff N, Amend D, Laakso MP, Hsu YY, Jagust WJ, Yaffe K, Kramer JH, Reed B, Norman D, Chui HC, Weiner MW (
2001
) Magnetic resonance imaging of the entorhinal cortex and hippocampus in mild cognitive impairment and Alzheimer’s disease.
J Neurol Neurosurg Psychiatryry
 
71
:
441
–447.
Esposito G, Kirby GS, Van Horn JD, Ellmore TM, Faith Berman K (
1999
) Context-dependent, neural system-specific neurophysiological concomitants of ageing: mapping PET correlates during cognitive activation.
Brain
 
122
:
963
–979.
Grady CL (
2002
) Age-related differences in face processing: a meta-analysis of three functional neuroimaging experiments.
Can J Exp Psychol
 
50
:
208
–220.
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, McIntosh AR, Horwitz B, Maisog JM, Ungerleider LG, Mentis MJ, Pietrini P, Schapiro MB, Haxby JV (
1995
) Age-related reductions in human recognition memory due to impaired encoding.
Science
 
269
:
218
–221.
Grady CL, Bernstein LJ, Beig S, Siegenthaler AL (
2002
) The effects of encoding strategy on age-related changes in the functional neuroanatomy of face memory.
Psychol Aging
 
17
:
7
–23.
Java RI (
1996
) Effects of age on state of awareness following implicit and explicit word-association tasks.
Psychol Aging
 
11
:
108
–111.
Jennings JM, Jacoby LL (
1993
) Automatic versus intentional uses of memory: aging, attention, and control.
Psychol Aging
 
8
:
283
–293.
Knight R (
1996
) Contribution of human hippocampal region to novelty detection.
Nature
 
383
:
256
–259.
Kohler S, Black SE, Sinden M, Szekely C, Kidron D, Parker JL, Foster JK, Moscovitch M, Winocour G, Szalai JP, Bronskill MJ (
1998
) Memory impairments associated with hippocampal versus parahippocampal-gyrus atrophy: an MR volumetry study in Alzheimer’s disease.
Neuropsychologia
 
36
:
901
–914.
Li KZH, Lindenberger U (
2002
) Relations between aging sensory/sensorimotor and cognitive functions.
Neurosci Biobehav Rev
 
26
:
777
–783.
Light LL (
1991
) Memory and aging: four hypotheses in search for data.
Annu Rev Psychol
 
42
:
333
–376.
Li S-C, Lindenberger U (
1999
) Cross-level unification: a computational exploration of the link between deterioration of neurotransmitter systems dedifferentiation of cognitive abilities in old age. In: Cognitive neuroscience of memory (Nilsson L-G, Markowitsch HJ, eds), pp.
103
–146. Seattle, WA: Hogrefe & Huber.
Lindenberger U, Baltes PB (
1994
) Sensory functioning and intelligence in old age: a strong connection.
Psychol Aging
 
9
:
339
–355.
Logan JM, Buckner RL (
2001
) Age-related changes in neural correlates of encoding. In: Eighth Annual Meeting of the Cognitive Neuroscience Society. New York.
Logan JM, Sanders AL, Snyder AZ, Morris JC, Buckner RL (
2002
) Under-recruitment and nonselective recruitment: dissociable neural mechanisms associated with aging.
Neuron
 
33
:
827
–840.
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.
Madden DJ, Turkington TG, Provenzale JM, Denny LL, Hawk TC, Gottlob LR, Coleman RE (
1999
) Adult age differences in functional neuroanatomy of verbal recognition memory.
Hum Brain Mapp
 
7
:
115
–135.
Mantyla T (
1993
) Knowing but not remembering: adult age differences in recollective experience.
Mem Cognit
 
21
:
379
–388.
Morcom AM, Good CD, Frackowiak RS, Rugg MD (
2003
) Age effects on the neural correlates of successful memory encoding.
Brain
 
126
:
213
–229.
Moscovitch M, Winocur G (
1992
) The neuropsychology of memory and aging. In: The handbook of aging and cognition (Craik FIM, Salthouse TA, eds), pp.
315
–372. Hillsdale, NJ: Erlbaum.
Nyberg L, Forkstam C, Petersson KM, Cabeza R, Ingvar M (
2002
) Brain imaging of human memory systems: between-systems similarities and within-system differences.
Cogn Brain Res
 
13
:
281
–292.
Nyberg L, Marklund P, Persson J, Cabeza R, Forkstam C, Petersson KM, Ingvar M (
2003
) Common prefrontal activations during working memory, episodic memory, and semantic memory.
Neuropsychologia
 
41
:
371
–377.
Oswald CJ, Yee BK, Rawlins JN, Bannerman DB, Good M, Honey RC (
2002
) The influence of selective lesions to components of the hippocampal system on the orienting response, habituation and latent inhibition.
Eur J Neurosci
 
15
:
1983
–1990.
Parkin AJ, Walter BM (
1992
) Recollective experience, normal aging, and frontal dysfunction.
Psychol Aging
 
7
:
290
–298.
Ranganath C, D’Esposito M (
2001
) Medial temporal lobe activity associated with active maintenance of novel information.
Neuron
 
31
:
865
–873.
Raz N (
2000
) Aging of the brain and its impact on cognitive performance: Integration of structural and functional findings. In: Handbook of aging and cognition II (Craik FIM, Salthouse TA, eds), pp. 
1
–90. Mahwah, NJ: Erlbaum.
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.
Reuter-Lorenz P, Jonides J, Smith ES, Hartley A, Miller A, Marshuetz C, Koeppe RA (
2000
) Age differences in the frontal lateralization of verbal and spatial working memory revealed by PET.
J Cogn Neurosci
 
12
:
174
–187.
Rosen AC, Prull MW, O’Hara R, Race EA, Desmond JE, Glover GH, Gabrieli JDE (
2002
) Variable effects of aging on frontal lobe contributions to memory.
Neuroreport
 
13
:
2425
–2428.
Rypma B, D’Esposito M (
2000
) Isolating the neural mechanisms of age-related changes in human working memory.
Nat Neurosci
 
3
:
509
–515.
Rypma B, Prabhakaran V, Desmond JD, Gabrieli JDE (
2001
) Age differences in prefrontal cortical activity in working memory.
Psychol Aging
 
16
:
371
–384.
Salthouse TA (
1996
) The processing speed theory of adult age differences in cognition.
Psychol Rev
 
103
:
403
–428.
Sarter M, Givens B, Bruno JP (
2001
) The cognitive neuroscience of sustained attention: where top-down meets bottom-up.
Brain Res Rev
 
35
:
146
–160.
Schneider BA, Pichora-Fuller MK (
2000
) Implications of perceptual deterioration for cognitive aging research. In: Handbook of cognitive aging II (Craik FIM, Salthouse TA, eds), pp. 155–219. Mahwah, NJ: Erlabum.
Searcy JH, Bartlett JC, Memon A (
1999
) Age differences in accuracy and choosing in eyewitness identification and face recognition.
Mem Cognit
 
27
:
538
–552.
Smith EE, Jonides J (
1999
) Storage and executive processes in the frontal lobes.
Science
 
283
:
1657
–1661.
Stern CE, Sherman SJ, Kirchhoff BA, Hasselmo ME (
2001
) Medial temporal and prefrontal contributions to working memory tasks with novel and familiar stimuli.
Hippocampus
 
11
:
337
–346.
Talairach J, Tournoux P (
1988
) A co-planar sterotactic atlas of the human brain. Stuttgart: Thieme.
Tulving E (
1983
) Elements of episodic memory. Oxford: Oxford University Press.
Tulving E, Markowitsch HJ, Kapur S, Habib R, Houle S (
1994
) Novelty encoding networks in the human brain: positron emission tomography data.
Neuroreport
 
5
:
2525
–2528.
Ungerleider LG, Mishkin M (
1982
) Two cortical visual systems. In: Analysis of visual behavior (Ingle DJ, Goodale MA, Mansfield RJW, eds), pp.
549
–589. Cambridge, MA: MIT Press.
West RL (
1996
) An application of prefrontal cortex function theory to cognitive aging.
Psychol Bull
 
120
:
272
–292.
Williams LM, Brammer MJ, Skerrett D, Lagopolous J, Rennie C, Kozek K, Olivieri G, Peduto T, Gordon E (
2000
) The neural correlates of orienting: an integration of fMRI and skin conductance orienting.
Neuroreport
 
11
:
3011
–3015.
Yonelinas AP (
2002
) The nature of recollection and familiarity: a review of 30 years of research.
Mem Lang
 
46
:
441
–517.