Utility of Brain Imaging Methods in Research on Aging
Christine R. Hartel
National Research Council
Randy L. Buckner
Washington University in St. Louis
The powerful techniques of neuroimaging have enabled the field of cognitive neuroscience to flourish, and today these techniques can be used in experiments designed to illuminate various aspects of social cognition. At the same time, the methods of social and cognitive psychology are importantly influencing experimental design in cognitive neuroscience; they have revealed that there are special measurement challenges in using neuroimaging in experiments with older adults.
The most commonly used methods of neuroimaging in cognitive studies are positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) (see reviews by Buxton, 2002; Carson, Daube-Witherspoon, and Herscovitch, 1997). These two techniques measure the way the cellular functioning of the brain supports mental activities such as memory, cognition, and emotion. PET does this by measuring the changes in local blood flow that occur whenever there are changes in neural activity. Magnetic resonance imaging (MRI) measures a parallel phenomenon: the very small changes in the concentration of oxygen in blood at the sites of brain activity. The intensity of an MRI signal is proportional to the amount of oxygen carried by hemoglobin in the blood. Functional MRI (fMRI) is the use of MRI to detect what are known as blood-oxygen-level-dependent (BOLD) signals to map brain function during various activities (Ogawa et al., 1992; Kwong et al., 1992).
PET and MRI scanners, and in particular the latter, are available in virtually all hospitals today across the United States and almost all are now
capable of making functional imaging measurements. There are many research centers that have scanners exclusively dedicated for research endeavors in psychology and in neuroscience. Therefore, neuroimaging resources are readily available for use as indirect measures of brain activities—including affective processes, executive function, or memory—that are not necessarily under the subject’s voluntary control.
PSYCHOLOGICAL APPLICATIONS OF NEUROIMAGING
Neuroimaging methods have a temporal resolution of a few seconds. This makes it possible to ask a question such as: Are there brain areas in which activity predicts whether or not items will be remembered or forgotten? In fact, Wagner and colleagues (1998) have found that the magnitude of activity in the left prefrontal cortex and in the medial temporal cortex predicts whether or not memories will form. Such a study is important not only in relation to understanding memory but also because it illustrates how these methods can be used in experiments in social psychology. Because the methods are indirect, they might be ideal for use in experimental work in which direct questions may not (or cannot) elicit an appropriate response, although questions remain about neuroimaging methodology and measurement in the elderly population.
It is important to remember that neuroimaging techniques are based on indirect measures of blood flow response properties. There are many physical and physiological properties that change with aging and these must be assessed methodologically. For example, the vascular system changes with age and many older individuals are hypertensive. There are all sorts of small, low-level effects on the arteries during aging, and all of these changes may affect the signals that result in the neuroimage. The coupling of neural activity to the neuroimaging signal—the basic assumption of MRI use—may have different properties in young and elderly subjects (D’Esposito, Zarahn, Aguirre, and Rypma, 1999).
As an analogy, the brain imaging methods available today produce images that are much like satellite pictures of earth: they take a relatively long time to complete and provide only a survey of the activities going on. The resolution is poor relative to the brain’s structural components. It would be useful to measure brain activity in a way that would show the activity of individual neurons and to observe it dynamically and across networks. But what is available is an indirect measure that provides a good cursory look at the broad regions of the brain that show activity. Even so, this measure appears to be sufficient for some of the questions being asked, especially in efforts to observe how broad pathways such as the affective systems or executive systems respond and change with age.
USES OF NEUROIMAGING IN RESEARCH ON AGING
The quality of contemporary fMRI images is also good enough that one can examine frontal cortex changes during aging. It is possible to actually reconstruct the thickness of the cortex and observe changes as people age, seeing when and where both white and gray cortical matter change. Among the many changes that occur, the ventricles enlarge and the gray matter thins. O’Sullivan and colleagues (2001) have found evidence that suggests that reduction of integrity in the white matter tracts of the brain may be an important mechanism of age-related cognitive decline. Their findings further suggest that disruption of the brain’s executive functioning is a form of cerebral “disconnection” and accounts for cognitive changes in normal aging. Mikels and Reuter-Lorenz (2004) found that the corpus callosum, which interconnects the left and right cerebral hemispheres of the brain, plays an important role in resource allocation and selective attention for certain types of memory tasks, suggesting that changes in it are another cause of age-related decline. In earlier studies, Reuter-Lorenz and Stanczak (2000) had established that older adults use both hemispheres to their advantage to solve problems at lower levels of task complexity compared to young adults, but the older adults also experience impaired sensorimotor transfer between hemispheres through the corpus callosum. These studies in the functional anatomy of the aging brain use behavioral output as the end measure of brain activity.
Methodological studies are also necessary to determine whether neuroimaging in the elderly gives reliable results that can be compared to those obtained from a younger population. To illustrate with a relatively simple fMRI experiment what one type of methodological study looks like: Buckner, Snyder, Sanders, Raichle, and Morris (2000) conducted a visual flicker experiment with healthy younger and older adults and with older demented adults. Subjects were presented with the image of a large, flickering checkerboard and were asked to press a key as quickly as they could as soon as they saw it. In young adults, this resulted in a clear response in the visual cortex at the back of the brain and an equally clear motor response in the front. In both groups of older adults, the amplitude of the hemodynamic response was reduced significantly in visual cortex. The variance in regional hemodynamic response magnitude has been seen in other studies (D’Esposito et al., 1999), and differences in measured variance have been marked. Such differences mean that the main effects between groups of subjects may be difficult to interpret, since direct comparisons between groups assume the variance is similar.
Buckner and his colleagues (2000) found similar response amplitudes in the motor cortex between young and older adults, and also between nondemented and demented older adults, but because of the assumptions
noted above, the results must be interpreted cautiously. Buckner et al. (2000) suggest that it is more conservative to look at group-by-region interactions, which would account for the differences in absolute response amplitude, or to look at relative change across conditions within a group rather than changes between them. These could be analyses of group-by-condition interactions or between-group parametric manipulations. The latter is particularly powerful when more than two levels of a condition are considered in each experimental group. Thus, even though various properties of the brain change with aging, and some baseline effects have to be taken into account, neuroimaging methods can be useful in these populations, but there are still measurement issues that must be attended to.
Neuroimaging techniques have also revealed that there are paradoxical differences in the ways older adults use their brains for certain kinds of memory functions. That is, in experimental memory tasks, young adults tend to use the left side of the brain for encoding and the right side for recall, while older adults recruit resources from both brain hemispheres to accomplish the task. The latter is called nonselective recruitment. Cabeza (2002) and Reuter-Lorenz (2002) have reviewed the PET and fMRI experiments that led to that generalization, and they suggest that the change observed in hemispheric asymmetry in older adults during verbal recall is not a task-specific occurrence but rather a general phenomenon of aging. Cabeza (2002), Reuter-Lorenz (2002), Buckner (2003), and others believe that this is a compensatory phenomenon. But Logan, Sanders, Snyder, Morris, and Buckner (2002) also raise the possibility that “nonselective recruitment reflects a breakdown in the appropriate selection of regions associated with controlled task performance,” citing the work of O’Sullivan et al. (2001) on cortical “disconnection” described above. Logan, Sanders, Snyder, Morris, and Buckner (2002) suggest that age-related changes in white matter integrity may cause nonselective recruitment that in turn leads to the processing difficulties so common in an elderly population.
Furthermore, in memory tasks, older adults, unlike younger ones, do not recruit the frontal areas of the brain that are the sites of the control processes needed for memory; this is called underrecruitment. The phenomena of nonselective recruitment and underrecruitment are paradoxical and seem to be due not to context or other effects but simply to differences in how the systems are recruited by younger and older adults. Researchers are really just beginning to understand how these work, both neuroanatomically and functionally.
However, under certain conditions, older adults can recruit the same executive areas of the brain and at the same level of intensity as younger adults. In fact, under a different set of conditions, the brains of younger adults can be made to look much like those of older adults in their recruitment strategies. This is a very important measurement issue, and is, in a
sense, like that described by Schwarz (this volume) with respect to surveys: results depend on the context in which you ask the question, and differences between younger and older adults can appear or disappear depending on the context. Such effects apply in many different experimental situations and must be taken into account in designing properly controlled experiments.
For example, Logan and colleagues (2002) have looked at memory formation in younger and older adults under several conditions. Under one condition, subjects were asked to memorize (intentional encoding) words (verbal stimuli) and faces (nonverbal stimuli) and told that they would be tested on these later while in the MRI scanner. This type of memory encoding is actually rather difficult because one must initiate strategies for memorization by oneself. In this situation, there are large differences between older and younger adults in the recruitment of the frontal regions (especially the left hemisphere) that are part of the brain’s executive system to encode memories. One might conclude that during aging the brain is actually losing some of its physical properties because of atrophy in these areas and that executive resources simply disappear with aging. But this turns out not to be true.
Under a different experimental condition (for words only), environmental support for memory encoding was provided by giving subjects an effective strategy for memorization by using meaning-based (semantic) elaboration. Under this condition, the memory of older adults was improved but not fully to that of younger adults, while the left frontal activity performance was nearly the same in both groups. However, the lack of selectivity in recruiting various brain areas was not reversed by semantic elaboration techniques in the elderly, suggesting a true and permanent decline in functioning with age. When both groups learned words “accidentally” (shallow incidental encoding), left frontal activity and memory in younger adults decreased to levels seen in older adults. In these experiments, care was taken to ensure that baseline differences in hemodynamic response properties that might be observed in between-population comparisons were controlled for. Baseline differences could result from internal physiological properties of individuals or from external properties of the analysis, like misregistration. Thus it seems clear that context effects are among the most salient, even at the neural level, in exploring the differences between younger and older adults.
Efforts to determine causal mechanisms for these phenomena are under way in many laboratories, but our understanding is still very limited. It is important to note that nonimaging methods can also provide important data for understanding causal mechanisms. For example, older people are not unaware of their memory deficits, and sometimes if asked to perform in a memory experiment they demur, asking to do some other task. By chang-
ing the nature of the experiment slightly, this problem can be overcome. For example, one can take advantage of the fact that people accidentally memorize things (incidental encoding) during an experiment that does not appear to be a memory test: participants may be shown a series of words, asked to complete an unrelated task, and then given another task that uses the original series of words without reference to the fact that it is the same list of words seen earlier. Neuroimages can be obtained during each phase of the experiment, taking advantage of the fact that participants will have unintentionally remembered some of the words and forgotten others.
The convergence of data from experiments using techniques like these that were developed in other sciences, and the use of the data, methods, and techniques of other sciences generally, will both improve scientists’ understanding and also constrain what can be learned from the current level of neuroimage resolution. While the resolution of PET and fMRI images is not of the quality desired to carry out all the types of experiments needed to explore the brain phenomena of aging, the quality of images we have is still remarkable. And the promise of these techniques—and their successors—is enormous for all fields of research involving brain function and the aging population.
Buckner, R.L. (2003). Functional-anatomic correlates of control processes in memory. Journal of Neuroscience, 23, 3999-4004.
Buckner, R.L., and Logan, J.M. (2002). Frontal contributions to episodic memory encoding in the young and elderly. In A.E. Parker, E.L. Wilding, and T. Bussey (Eds.), The cognitive neuroscience of memory encoding and retrieval. Philadelphia: Psychology Press.
Buckner, R.L., Snyder, A.Z., Sanders, A.L., Raichle, M.E., and Morris, J.C. (2000). Functional brain imaging of young, nondemented, and demented older adults. Journal of Cognitive Neuroscience, 12(Suppl. 2), 24-34.
Buxton, R.B. (2002). Introduction to functional magnetic resonance imaging: Principles and techniques . New York: Cambridge University Press.
Cabeza, R. (2002). Hemispheric asymmetry reduction in older adults: The HAROLD model. Psychology and Aging, 17, 85-100.
Carson, R.E., Daube-Witherspoon, M.E., and Herscovitch, P. (1997). Quantitative functional brain imaging with positron emission tomography. San Diego: Academic Press.
D’Esposito, M., Zarahn, E., Aguirre, G.K. and Rypma, B. (1999). The effects of normal aging on the coupling of neural activity to the BOLD hemodynamic response. NeuroImage, 10, 6-14.
Head, D., Snyder, A.Z., Girton, L.E., Morris, J.C., and Buckner, R.L. (2004). Frontal-hippocampal double dissociation between normal aging and Alzheimer’s disease. Cerebral Cortex, 14, 410-423.
Kwong, K.K., Belliveau, J.W., Chesler, D., Goldberg, I.E., Weiskoff, R.J., Poncelet, B.P., Kennedy, D.N., Hoppel, B.E., Cohen, J.S., Turner, R., Cheng, H., Brady, T.J., and Rosen, B.R. (1992). Dynamic magnetic resonance imaging of human brain activity during primary sensory simulation. Proceedings of the National Academy of Sciences, 89, 5675-5679.
Logan, J.M., Sanders, A.L., Snyder, A.Z., Morris, J.C., and Buckner, R.L. (2002). Under-recruitment and nonselective recruitment: Dissociable neural mechanisms associated with aging. Neuron, 33(5), 827-840.
Mikels, J.A., and Reuter-Lorenz, P.A. (2004). Neural gate keeping: The role of interhemispheric interactions in resource allocation and selective filtering. Neuropsychology, 18, 328-339.
Ogawa, S., Tank, D.W., Menon, R., Ellermann, J.M., Kim, S.-G., Merkle, H., and Ugurbil, K. (1992). Intrinsic signal changes accompanying sensory stimulation: Functional brain mapping with magnetic resonance. Proceedings of the National Academy of Sciences, 89, 5951-5955.
O’Sullivan, M., Jones, D.K., Summers, P.E., Morris, R.G., Williams, S.C.R., and Markus, H.S. (2001). Evidence for cortical “disconnection” as a mechanism of age-related cognitive decline. Neurology, 57, 632-638.
Reuter-Lorenz, P. (2002). New visions of the aging mind and brain. Trends in Cognitive Science, 6, 394-400.
Reuter-Lorenz, P.A., and Stanczak, L. (2000). Differential effects of aging on the functions of the corpus callosum. Developmental Neuropsychology, 18, 113-137.
Wagner, A.D., Schacter, D.L., Rotte, M., Koutstaal, W., Maril, A., Dale, A.M., Rosen, B.R., and Buckner, R.L. (1998). Building memories: Remembering and forgetting of verbal experiences as predicted by brain activity. Science, 281, 1188-1191.