Sci. Aging Knowl. Environ., 29 January 2003
Vol. 2003, Issue 4, p. pe2
[DOI: 10.1126/sageke.2003.4.pe2]

PERSPECTIVES

The Contribution of Functional Brain Imaging to Our Understanding of Cognitive Aging

Adam Gazzaley, and Mark D'Esposito

The authors are at the Henry H. Wheeler, Jr. Brain Imaging Center at the Helen Wills Neuroscience Institute and the Department of Psychology at the University of California, Berkeley, CA 94720, USA. E-mail: despo{at}socrates.Berkeley.edu (M.D.)

http://sageke.sciencemag.org/cgi/content/full/sageke;2003/4/pe2

Key Words: cognitive aging • functional brain imaging • PET • fMRI • neuroimaging

Introduction

Normal aging is associated with a decline in cognitive abilities in multiple domains such as memory, attention, and perception (1, 2). Determining which abilities decline and how such changes occur has been a major area of research for many years. Two prominent, although not necessarily mutually exclusive, hypotheses have emerged. One hypothesis proposes that all cognitive decline in normal older individuals can be explained by a single mechanism such as a decrease in information processing speed (3). Another possibility is that certain specific cognitive abilities decline with age, whereas others do not (4). The search for the neural mechanisms that produce cognitive deficits has involved a variety of experimental approaches, including neurophysiological, neurochemical, and neuroanatomical methodologies. This combination of multiple approaches has been valuable to our understanding of the neural mechanisms underlying cognitive aging.

One possibility is that the age-associated decline in cognitive function might be at least partially brought about by changes in neuron structure. For example, multiple dendrites extend from each neuronal cell body, a formation known as the dendritic arbor, and several anatomical studies have revealed subtle age-related structural changes, such as alterations in the pattern of dendritic arborization and spine count. Spines are small protuberances located on dendrites that are the primary site of excitatory synapses (5, 6). However, many other studies have revealed no such changes and additionally indicate that neuronal number is preserved during aging (7-10). Despite the lack of overwhelming evidence for structural alterations, there has been a steady accumulation of studies that reveal physiological changes (11, 12) and changes in neurotransmitter concentrations with aging (13). These observations have led to an emerging principle: Perhaps changes in neural activity and connectivity, rather than alterations in neuron structure, account for age-related cognitive deficits.

Functional Brain Imaging Techniques

This principle has been a guiding force for the application of functional brain imaging techniques to the study of age-related memory alterations, which has occurred relatively recently. Such methods include positron emission tomography (PET) and functional magnetic resonance imaging (fMRI), which allow an assessment of age-related changes in brain function while subjects are performing cognitive tasks. PET and fMRI both measure changes in cerebral blood flow, an indirect but highly localized correlate of increases in neural activity. PET requires the injection of radiotracers into study subjects in order to measure their cerebral blood flow, whereas fMRI is a noninvasive tool that measures relative changes in the concentration of oxygen in the blood, based on the magnetic properties of hemoglobin. Both techniques offer a powerful means of exploring activity changes in the functioning human brain, but fMRI has emerged as the dominant technique in use today because of the widespread availability of MRI scanners as compared to cyclotrons (which are required for PET), the lack of a need for injection of radiotracers, and improved temporal and spatial resolution (14).

More recently, the application of novel multivariate statistical analysis to fMRI data has furthered the scope of conclusions that can be derived from imaging studies by allowing the analysis of brain region interactions in distributed cortical networks (15). Distributed cortical networks refer to the collection of multiple brain regions that collectively participate in a neural process. These regions interact with each other via the well-described neurophysiological mechanism of neuronal signal transduction. It is the complex pattern of communication between brain regions that generates the diverse array of cognitive processes that we command. Multivariate statistical analysis allows researchers to analyze these network interactions by studying the degree of correlation between simultaneous activities in multiple brain regions within the confines of their established structural connections.

Possible Study Outcomes

One approach of functional brain imaging studies of aging has been to compare brain activation patterns in a group of healthy young individuals (usually in the age range of 18 to 25 years old) and a group of healthy older individuals (usually in the range of 65 to 85 years old) during the performance of a task that taps a cognitive process or ability that shows age-related decline in behavioral tests. An example of such a task would be one that tested the ability of a subject to remember a series of visual images, such as faces, over time. Three possible age-related differences in patterns of brain activation are conceivable:

Scenario 1: A similar pattern of activation between groups but (i) lower or (ii) greater activation in older individuals in all (or some) of the same brain regions. This activation pattern might indicate that the reduced cognitive performance in older individuals relative to younger individuals is a result of a failure in neural systems that are necessary for a specific cognitive operation [in the case of possibility (i)] or that neural systems supporting certain cognitive abilities are "working harder" as behavioral performance fails [in the case of (ii)]. This second possibility might reflect a compensatory mechanism.

Scenario 2: A different pattern of activation between groups; that is, (i) new brain regions are activated in older individuals that are not active in young individuals during performance of the same cognitive task, or (ii) some brain regions are no longer active in older individuals. Pattern (i) might also represent a compensatory mechanism; either (i) or (ii) might represent some fundamental change in functional organization with advanced age.

Scenario 3: A similar pattern of activation between groups, but a modification of the network connectivity (the interactions among neural regions) during the performance of cognitive tasks as determined by multivariate statistical analysis. For example, there might not be a difference in the mean level of activation in a particular brain region between younger and older individuals, but the relationship of the activity in one region to the activity in other anatomically connected regions might differ between groups during particular cognitive tasks (16). This result might represent a structural reorganization of cognitive pathways or a physiological change in their connectivity.

Of course, any combination of these scenarios is also possible, and indeed most functional brain imaging studies reveal mixed results. As an example, a recent study by Cabeza et al. (17) revealed both decreases (scenario 1i) and increases (1ii) in the activity of brain regions of old adults that were also activated in young adults, as well as a loss of activity in certain brain regions (2ii) and newly identified active regions (2i) in the older adults.

Compensation and Dedifferentiation Hypotheses

Two primary hypotheses to explain the alterations in activity and connectivity observed in older individuals have been developed [reviewed in (18)]: the compensation hypothesis (19) and the dedifferentiation hypothesis (20). The compensation hypothesis proposes that observed age-related neural alterations occur to counteract cognitive deficits. This conclusion is usually reached in studies in which changes in neural activity and/or connectivity are observed in the setting of preserved cognitive performance in older subjects. The dedifferentiation hypothesis proposes that some pathological process causes an age-related difficulty in performing a particular cognitive function. This difficulty in turn leads to the recruitment of a less specialized neural mechanism to perform that function, as opposed to the highly specialized mechanism formerly in place. Such decreased specialization manifests as an increased correlation among diverse cognitive measures (21) and/or anatomical patterns of activity (22). The exact brain mechanism responsible for dedifferentiation is currently unknown, but it is considered an example of a primary degenerative event associated with aging rather than a compensatory mechanism. Compensation and dedifferentiation are not mutually exclusive processes, and both might be occurring and influencing one another (23).

The changes described in scenario 1, especially age-related reductions in regional cerebral blood flow, are common findings in imaging studies (17, 24, 25). These results, however, must be interpreted with caution and not automatically attributed to decreased neural activity. Older individuals are more likely to experience cerebrovascular disease that might lead to decreased cerebral blood flow and neurovascular decoupling, resulting in the misinterpretation of a reduced signal in PET or fMRI experiments as representing a reduction in neural activity (see further discussion of this issue below). Increases in PET and fMRI signals, especially when correlated with changes in behavioral performance, might be easier to accept as genuine neural correlates of cognitive aging. Such patterns have not been frequently observed in brain imaging studies, however. Most studies that describe "increases" in activity in older subjects are actually reporting novel areas of activation that represent alterations in the pattern of activity. However, the relationship between the extent of neural activation and behavioral performance might be complex. For example, age-related changes in activity have been described in which better behavioral performance was associated with decreased brain activation in young individuals and with increased brain activation in older individuals (25).

Scenario 2, a change in the pattern of activity in older individuals, has been described in multiple brain imaging studies. A common finding in PET studies (19, 26-28), which has also been observed in an fMRI study (29), is that activity in the prefrontal cortex seems to be more bilateral in older subjects, as compared to young subjects, in a variety of cognitive domains. For example, young individuals often manifest asymmetric prefrontal activity, so that left prefrontal activity occurs during memory encoding and right prefrontal activity occurs during memory retrieval (2), whereas older subjects display more equivalent right and left prefrontal activity during both of these processes. Memory encoding occurs when a subject is first presented with a stimulus (such as a face). Retrieval refers to the subsequent memory stage that occurs after a delay and involves the subject selecting from a series of similar stimuli with the goal of identifying the correct initial stimulus (for example, selecting the correct face from a series of faces). A model based on these findings, Hemispheric Asymmetry Reduction in Old Adults, or HAROLD (19), has been proposed and has generated a great deal of interest in whether this alteration in the pattern of activity represents a compensatory change or dedifferentiation. Several studies have suggested that the decrease in asymmetry represents a compensatory change, because it often occurs in conjunction with preserved behavioral performance in the older age group (26, 27). The most convincing argument, however, was presented in a recent paper (17). In this study, prefrontal cortical activity during a recall memory task was measured in three study groups: young adults, low-performing older adults, and high-performing older adults (who exhibited performance equivalent to that of the young). The results of this study support the proposal that HAROLD represents a compensatory change, in that both young and low-performing older adults displayed asymmetric, right prefrontal cortex activation, whereas the high-performing older adults had a bilateral prefrontal cortex pattern (Fig. 1). These findings suggest that the low-performing older adults used the same region of the prefrontal cortex less efficiently than did young adults, and that the high-performing older adults were additionally able to recruit the left prefrontal cortex (leading to the bilateral pattern of activity) in order to maintain cognitive function.



View larger version (44K):
[in this window]
[in a new window]
 
Fig. 1. A PET study revealing that activity in the prefrontal cortex was right-lateralized in both the young and low-performing older subjects, but bilateral in the older higher-performing subjects. [Reprinted from (17) with permission from Elsevier Science]

 
When considering scenario 3 (network reorganization), it is important to remember that brain regions do not operate in isolation but generate a broad array of cognitive abilities by their interactions with other regions. For example, working memory involves communication between the prefrontal cortex (which has an important role in attention) and visual association areas (presumed site of visual memory storage). The recent use of multivariate statistical analyses, such as partial least squares analysis and structural equation modeling, has served as a powerful means of investigating modifications in regional connectivity associated with cognitive aging. Several multivariate analyses of fMRI (30) and PET data (15, 16, 31, 32) have revealed that aging is not merely associated with regional changes in neural activity but also with more widespread alterations in large-scale neural networks.

Network analysis of face perception (in which individuals make judgments about faces presented simultaneously) has revealed that despite relative preservation of perceptual abilities, as noted by equivalent behavioral performance, there is a reorganization of the interregional connectivity during aging. For example, alterations in connections between executive centers (areas of the brain that are necessary for the overriding control and execution of many cognitive processes) and primary sensory areas have been observed. Specifically, older adults demonstrated a similar pattern of activation in the brain during face perception, but displayed stronger feedback influences (which constitute the communication of executive brain regions with primary sensory regions) from the frontal cortex, an executive center, to the occipital cortex, a primary sensory area where visual stimuli are actually perceived. This result suggests that certain brain functions in older individuals might require additional monitoring by executive centers (33, 34). An increase in feedback influences signifies a greater dependence on the influence of these control areas on the primary areas during perception in the older adults (Fig. 2). Increased prefrontal feedback to the hippocampus, a region of the brain important for long-term memory consolidation, has also been observed in memory tasks, revealing that additional reliance on executive centers in aging is also associated with memory processes (15).



View larger version (40K):
[in this window]
[in a new window]
 
Fig. 2. A proposed model illustrating an increase in the influence of an executive center (prefrontal cortex) on a primary sensory area (occipital cortex) in older adults.

 
Another age-related modification in network connectivity that has been observed is a ventral/dorsal shift of the hippocampal-cortical network. Hippocampal activity during object and word encoding correlates with more ventral cortices in young adults (in such a way that activity in the ventral cortices and hippocampus change together) and with more dorsal cortices in older adults. These changes are also correlated with better recognition memory performance in older adults (15). In other words, adults who display greater correlation between the activities of the hippocampus and the dorsal cortices also display better recognition memory performance. This ventral/dorsal shift, as well as a memory performance correlation, was also observed for the encoding of memories for faces (27), suggesting that this shift represents a memory modification that occurs in response to a variety of stimuli. These findings suggest a shift from the interaction of the hippocampus with perceptually based regions in young adults [regions such as those localized to ventral cortices (e.g., the extrastriate cortex, a visual area)] to interaction with more executive-dependent regions, such as those localized to dorsal cortices (e.g., the prefrontal cortex, an executive area) in older adults. This conclusion was based on the ventral/dorsal shift in activity correlation and further suggests a greater reliance on executive control in aging adults. Furthermore, these results support the compensation hypothesis, because this change in connectivity correlates with performance, and no overall performance differences were observed between age groups (15).

Conclusions

Despite the possibilities that might lie ahead in the use of functional brain imaging methods to study the mechanisms of cognitive aging, several methodological concerns must be considered. The principal concern is that PET and fMRI are measures of blood flow, which is an indirect correlate of neural activity. If aging affects regional cerebral blood flow as well as neural activity, age-related differences in patterns of "brain activation" found in functional brain imaging measures might be misleading. For example, decreases in activation or changes in the pattern of activation might be a result of vascular and not neural differences between young and old brains, given that there is significant evidence for vascular changes in normal aging (35).

Several groups have begun to ask whether there are age-related differences in the coupling between neural activity and imaging signals (36-38). In other words, how well an imaging signal represents neural activity might vary depending on age. For example, our laboratory has examined this coupling in the primary motor cortex during the performance of a simple motor task, which presumably taps an age-invariant ability (36). Importantly, significant neuronal loss in the primary motor cortex during aging has not been demonstrated (39), and similar movement-related electrical potentials have been observed in young and older individuals during the performance of simple motor tasks (40). Thus, we presumed that any changes in fMRI signal that we observed between young and older individuals in the motor cortex would be a result of vascular, and not neural, changes that occur during normal aging. In our study, we found a significantly decreased signal-to-noise ratio in the fMRI signal in older individuals as compared to young individuals. This result was attributed to a greater level of noise in the older individuals and suggests that there is some property of the coupling between neural activity and fMRI signal that changes with age. The source of this difference has not yet been determined, but our results suggest that it is a result of a hemodynamic change (the response of the vascular system to neural activity), and not a neural change, in normal aging. This finding cautions against simple interpretations of results of functional brain imaging studies that compare young and older individuals. Moreover, new methods must be developed to account for this potential methodological confound.

Nevertheless, with the advancement of our understanding of functional brain imaging methods and the development of other complementary methods, new answers to the question of how brain-behavior relationships change with age are likely to emerge.


January 29, 2003
  1. S. A. Small, Age-related memory decline: current concepts and future directions. Arch. Neurol. 58, 360-364 (2001).[CrossRef][Medline]
  2. C. L. Grady, F. I. Craik, Changes in memory processing with age. Curr. Opin. Neurobiol. 10, 224-231 (2000).[CrossRef][Medline]
  3. T. A. Salthouse, The processing-speed theory of adult age differences in cognition. Psychol. Rev. 103, 403-428 (1996).[CrossRef][Medline]
  4. A. D. Fisk, W. A. Rogers, Toward an understanding of age-related memory and visual search effects. J. Exp. Psychol. Gen. 120, 131-149 (1991).[CrossRef][Medline]
  5. B. Anderson, V. Rutledge, Age and hemisphere effects on dendritic structure. Brain 119 (Pt 6), 1983-1990 (1996).[Abstract/Free Full Text]
  6. J. M. de Brabander, R. J. Kramers, H. B. Uylings, Layer-specific dendritic regression of pyramidal cells with ageing in the human prefrontal cortex. Eur. J. Neurosci. 10, 1261-1269 (1998).[CrossRef][Medline]
  7. A. H. Gazzaley, M. M. Thakker, P. R. Hof, J. H. Morrison, Preserved number of entorhinal cortex layer II neurons in aged macaque monkeys. Neurobiol. Aging 18, 549-553 (1997).[CrossRef][Medline]
  8. M. J. West, P. D. Coleman, D. G. Flood, J. C. Troncoso, Differences in the pattern of hippocampal neuronal loss in normal ageing and Alzheimer's disease. Lancet 344, 769-772 (1994).[CrossRef][Medline]
  9. J. H. Morrison, P. R. Hof, Life and death of neurons in the aging brain. Science 278, 412-419 (1997).[Abstract/Free Full Text]
  10. A. Peters, Structural changes in the normally aging cerebral cortex of primates. Prog. Brain Res. 136, 455-465 (2002).[Medline]
  11. S. Shankar, T. J. Teyler, N. Robbins, Aging differentially alters forms of long-term potentiation in rat hippocampal area CA1. J. Neurophysiol. 79, 334-341 (1998).[Abstract/Free Full Text]
  12. L. L. Chao, R. T. Knight, Prefrontal deficits in attention and inhibitory control with aging. Cereb. Cortex 7, 63-69 (1997).[Abstract/Free Full Text]
  13. A. H. Gazzaley, S. J. Siegel, J. H. Kordower, E. J. Mufson, J. H. Morrison, Circuit-specific alterations of N-methyl-D-aspartate receptor subunit 1 in the dentate gyrus of aged monkeys. Proc. Natl. Acad. Sci. U.S.A. 93, 3121-3125 (1996).[Abstract/Free Full Text]
  14. M. D'Esposito, Functional neuroimaging of cognition. Semin. Neurol. 20, 487-498 (2000).[CrossRef][Medline]
  15. C. L. Grady, A. R. McIntosh, F. I. Craik, Age-related differences in the functional connectivity of the hippocampus during memory encoding. Hippocampus, in press.
  16. R. Cabeza, A. R. McIntosh, E. Tulving, L. Nyberg, C. L. Grady, Age-related differences in effective neural connectivity during encoding and recall. Neuroreport 8, 3479-3483 (1997).[Medline]
  17. R. Cabeza, N. D. Anderson, J. K. Locantore, A. R. McIntosh, Aging gracefully: compensatory brain activity in high-performing older adults. Neuroimage 17, 1394-1402 (2002).[CrossRef][Medline]
  18. P. Reuter-Lorenz, New visions of the aging mind and brain. Trends Cognit. Sci. 6, 394 (2002).[CrossRef][Medline]
  19. R. Cabeza, Hemispheric asymmetry reduction in older adults: the HAROLD model. Psychol. Aging 17, 85-100 (2002).[CrossRef][Medline]
  20. S. C. Li, U. Lindenberger, Cross-Level Unification: A Computational Exploration of the Link Between Deterioration of Neurotransmitter Systems Dedifferentiation of Cognitive Abilities in Old Age (Hogrefe & Huber, Seattle, WA, 1999).
  21. P. B. Baltes, U. Lindenberger, 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 (1997).[CrossRef][Medline]
  22. C. L. Grady, Age-related differences in face processing: a meta-analysis of three functional neuroimaging experiments. Can. J. Exp. Psychol. 56, 208-220 (2002).[Medline]
  23. P. A. Reuter-Lorenz, L. Stanczak, A. Miller, Neural recruitment and cognitive aging: Two hemispheres are better than one, especially as you age. Psychol. Sci. 10, 494-500 (1999).[Abstract/Free Full Text]
  24. C. L. Grady, A. R. McIntosh, B. Horwitz, J. M. Maisog, L. G. Ungerleider, M. J. Mentis, P. Pietrini, M. B. Schapiro, J. V. Haxby, Age-related reductions in human recognition memory due to impaired encoding. Science 269, 218-221 (1995).[Abstract/Free Full Text]
  25. B. Rypma, M. D'Esposito, Isolating the neural mechanisms of age-related changes in human working memory. Nature Neurosci. 3, 509-515 (2000).[CrossRef][Medline]
  26. R. Cabeza, C. L. Grady, L. Nyberg, A. R. McIntosh, E. Tulving, S. Kapur, J. M. Jennings, S. Houle, F. I. Craik, Age-related differences in neural activity during memory encoding and retrieval: a positron emission tomography study. J. Neurosci. 17, 391-400 (1997).[Abstract/Free Full Text]
  27. C. L. Grady, L. J. Bernstein, S. Beig, A. L. Siegenthaler, The effects of encoding task on age-related differences in the functional neuroanatomy of face memory. Psychol. Aging 17, 7-23 (2002).[CrossRef][Medline]
  28. P. A. Reuter-Lorenz, J. Jonides, E. E. Smith, A. Hartley, A. Miller, C. Marshuetz, R. A. Koeppe, Age differences in the frontal lateralization of verbal and spatial working memory revealed by PET. J. Cogn. Neurosci. 12, 174-187 (2000).[CrossRef][Medline]
  29. J. M. Logan, A. L. Sanders, A. Z. Snyder, J. C. Morris, R. L. Buckner, Under-recruitment and nonselective recruitment: dissociable neural mechanisms associated with aging. Neuron 33, 827-840 (2002).[CrossRef][Medline]
  30. J. B. Krause, J. G. Taylor, D. Schmidt, H. Hautzel, F. M. Mottaghy, H. W. Muller-Gartner, Imaging and neural modelling in episodic and working memory processes. Neural Netw. 13, 847-859 (2000).[CrossRef][Medline]
  31. V. Della-Maggiore, A. B. Sekuler, C. L. Grady, P. J. Bennett, R. Sekuler, A. R. McIntosh, Corticolimbic interactions associated with performance on a short-term memory task are modified by age. J. Neurosci. 20, 8410-8416 (2000).[Abstract/Free Full Text]
  32. A. R. McIntosh, A. B. Sekuler, C. Penpeci, M. N. Rajah, C. L. Grady, R. Sekuler, P. J. Bennett, Recruitment of unique neural systems to support visual memory in normal aging. Curr. Biol. 9, 1275-1278 (1999).[CrossRef][Medline]
  33. A. R. McIntosh, C. L. Grady, L. G. Ungerleider, J. V. Haxby, S. I. Rapoport, B. Horwitz, Network analysis of cortical visual pathways mapped with PET. J. Neurosci. 14, 655-666 (1994).[Abstract]
  34. B. Horwitz, A. R. McIntosh, J. V. Haxby, M. Furey, J. A. Salerno, M. B. Schapiro, S. I. Rapoport, C. L. Grady, Network analysis of PET-mapped visual pathways in Alzheimer type dementia. Neuroreport 6, 2287-2292 (1995).[Medline]
  35. M. Kobari, J. S. Meyer, M. Ichijo, Leuko-araiosis, cerebral atrophy, and cerebral perfusion in normal aging. Arch. Neurol. 47, 161-165 (1990).[CrossRef][Medline]
  36. M. D'Esposito, E. Zarahn, G. K. Aguirre, B. Rypma, The effect of normal aging on the coupling of neural activity to the bold hemodynamic response. Neuroimage 10, 6-14 (1999).[CrossRef][Medline]
  37. M. H. Ross, D. A. Yurgelun-Todd, P. F. Renshaw, L. C. Maas, J. H. Mendelson, N. K. Mello, B. M. Cohen, J. M. Levin, Age-related reduction in functional MRI response to photic stimulation. Neurology 48, 173-176 (1997).[Abstract/Free Full Text]
  38. R. L. Buckner, A. Z. Snyder, A. L. Sanders, M. E. Raichle, J. C. Morris, Functional brain imaging of young, nondemented, and demented older adults. J. Cogn. Neurosci. 12 Suppl. 2, 24-34 (2000).
  39. H. Haug, The Aging Human Cerebral Cortex: Morphometry of Areal Differences and Their Functional Meaning (Elsevier, Amsterdam, 1997).
  40. R. Cunnington, R. Iansek, K. A. Johnson, J. L. Bradshaw, Movement-related potentials in Parkinson's disease. Motor imagery and movement preparation. Brain 120 (Pt. 8), 1339-1353 (1997).[Abstract/Free Full Text]
Citation: A. Gazzaley, M. D'Esposito, The Contribution of Functional Brain Imaging to Our Understanding of Cognitive Aging. Sci. SAGE KE 2003, pe2 (29 January 2003)
http://sageke.sciencemag.org/cgi/content/full/sageke;2003/4/pe2




THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES:
Keeping Priorities: The Role of Working Memory and Selective Attention in Cognitive Aging.
J. W. de Fockert (2005)
Sci. Aging Knowl. Environ. 2005, pe34
   Abstract »    Full Text »




Science of Aging Knowledge Environment. ISSN 1539-6150