Cheryl L. Grady

Spatial Pattern Analysis of Functional Brain Images Using Partial Least Squares

This paper introduces a new tool for functional neuroimage analysis: partial least squares (PLS). It is unique as a multivariate method in its choice of emphasis for analysis, that being the covariance between brain images and exogenous blocks representing either the experiment design or some behavioral measure. What emerges are spatial patterns of brain activity that represent the optimal association between the images and either of the blocks. This process differs substantially from other multivariate methods in that rather than attempting to predict the individual values of the image pixels, PLS attempts to explain the relation between image pixels and task or behavior. Data from a face encoding and recognition PET rCBF study are used to illustrate two types of PLS analysis: an activation analysis of task with images and a brain-behavior analysis. The commonalities across the two analyses are suggestive of a general face memory network differentially engaged during encoding and recognition. PLS thus serves as an important extension by extracting new information from imaging data that is not accessible through other currently used univariate and multivariate image analysis tools.

The effect of white matter hyperintensity volume on brain structure, cognitive performance, and cerebral metabolism of glucose in 51 healthy adults

Objective To assess the association of MRI white matter hyperintensities (WMHI) with cognitive performance, cerebral structure, and cerebral metabolism in 51 healthy individuals aged 19 to 91 years without cerebrovascular risk factors. Background Abnormal white matter signals have been associated with brain atrophy, reduced cerebral blood flow, focal neurologic signs, gait disorder, and poorer neuropsychological test performance. Most studies of WMHI, however, include subjects with hypertension or other identifiable causes of cerebrovascular disease that may have an independent effect on brain structure and function. To assess brain changes associated with WMHI independent of cerebrovascular risk factors, we determined WMHI volume, brain volume, cerebral metabolism, and cognitive performance for a group of subjects free of medical illness. Regional cerebral metabolism and cognitive domains were also assessed to evaluate the possible role of frontal lobe dysfunction in subjects with WMHI. Design Cross-sectional study of 51 very healthy subjects aged 19 to 91 years. Methods WMHI, brain, and CSF volumes were determined by MRI segmentation. Neuropsychological tests were employed to assess multiple cognitive domains. Brain metabolism was determined from 18-fluoro-2-deoxy-d-glucose PET. Multivariate relations were tested with stepwise linear regression. Models included the potential confounders of age and education where appropriate. Results The distribution of WMHI volume was bimodal, with five subjects having WMHI volumes beyond three SDs from the normally distributed population. A WMHI volume of greater than 0.5% of intracranial volume was considered abnormal. Within the multivariate models, WMHI volumes were significantly predictive of increased ventricular volume, reduced brain volume, and reduced cognitive scores. Subjects with greater than 0.5% WMHI volume also had significantly lower frontal lobe metabolism, significantly higher systolic blood pressure, significantly larger ventricular volume, and significantly lower scores on frontal lobe-mediated neuropsychological tests than age-matched controls. Conclusion WMHI volume is associated with structural and functional brain changes even within a group of very healthy individuals. WMHI is associated with poorer frontal lobe cognitive function and, when severe, is accompanied by significantly reduced frontal lobe metabolism. Subjects with large WMHI volumes have significantly higher systolic blood pressure, brain atrophy, reduced cerebral metabolism, and lower scores on tests of frontal lobe function than age-matched controls. Large amounts of WMHI are, therefore, pathologic and may be related to elevated systolic blood pressure even when it is within the normal age-related range.

Longitudinal study of the early neuropsychological and cerebral metabolic changes in dementia of the Alzheimer type

To examine the progression of neuropsychologic and metabolic changes in the early stages of dementia of the Alzheimer type (DAT), we studied 11 midly demented patients longitudinally. Three aspects of neuropsychological function were measured: memory, attention to complex sets and abstract reasoning, and lateralized functions, i.e., language and visuoconstruction. Regional cerebral metabolic rates for glucose were measured in frontal, parietal, and temporal association cortices. Our results show that, in general, memory deficits are the first neuropsychological impairments to occur in DAT, followed by problems with attention to complex cognitive sets and abstract reasoning, which are followed in turn by deficits in language and visuospatial abilities. In addition, neocortical metabolic abnormalities usually precede impairment of neocortically mediated attention and abstract reasoning by 8 to 16 months, and precede impairment of neocortically mediated language and visuospatial function by 12 to 37 months. These findings suggest that the first nonmnestic neuropsychological consequence of neocortical physiological dysfunction in DAT is a loss of attentional capacity. Since neocortical metabolic changes generally precede the appearance of neocortically mediated neuropsychological dysfunction, physiologic dysfunction may exist for some time before cognition is affected.

Relations between Neuropsychological and Cerebral Metabolic Asymmetries in Early Alzheimer's Disease

Regional CMRglc (rCMRglc) values were determined with positron emission tomography (PET) in 10 patients with mild to moderate clinically diagnosed Alzheimers disease (AD) and in 26 healthy controls. rCMRglc in frontal, parietal, and temporal association cortices were significantly more laterally asymmetrical in AD patients than in controls (p < 0.05). Furthermore, lateral asymmetry of rCMRglc in AD patients but not in the control subjects correlated significantly with asymmetry of language and visuospatial functions such that lower left than right rCMRglc was associated with relatively greater impairment of language and vice versa. The results demonstrate that discrepancies between language and visuospatial deficits in patients with early AD are related to asymmetrical reductions in cerebral cortical glucose metabolism.

The Metabolic Topography of Normal Aging

Normal aging is associated with the degeneration of specific neural systems. We used [18F]fluorode-oxyglucose (FDG)/positron emission tomography (PET) and a statistical model of regional covariation to explore the metabolic topography of this process. We calculated global and regional metabolic rates for glucose (GMR and rCMRglc) in two groups of normal subjects studied independently on different tomographs: Group 1—130 normal subjects (62 men and 68 women; range 21–90 years); Group 2—20 normal subjects (10 men and 10 women; range 24–78 years). In each of the two groups, the Scaled Subprofile Model (SSM) was applied to rCMRglc data to identify specific age-related profiles. The validity of these profiles as aging markers was assessed by correlating the associated subject scores with chronological age in both normal populations. SSM analysis disclosed two significant topographic profiles associated with aging. The first topographic profile, extracted in an analysis of group 1 normals, was characterized by relative frontal hypometabolism associated with covariate metabolic increases in the parietooccipital association areas, basal ganglia, mid-brain, and cerebellum. Subject scores for this profile correlated significantly with age in both normal groups (R2 = 0.48 and 0.33, p < 0.0001 for groups 1 and 2, respectively). Because of clinical similarities between normal motoric aging and parkinsonism, we explored the possibility of shared elements in the metabolic topography of both processes. We performed a combined group SSM analysis of the 20 group 2 normals and 22 age-matched Parkinsons disease patients, and identified another aging-related topographic profile. This profile was characterized by relative basal ganglia hypermetabolism associated with covariate decreases in frontal premotor cortex. Subject scores for this profile also correlated significantly with age in both normal groups (group 1: R2 = 0.30, p < 0.00001; group 2: R2 = 0.59, p < 0.01). Healthy aging is associated with reproducible topographic covariation profiles associated with specific neural systems. FDG/PET may provide a useful metabolic marker of the normal agingprocess.

Effect of task difficulty on cerebral blood flow during perceptual matching of faces

To aid our understanding of age‐related changes in brain activation during visuoperceptual processing, we designed an experiment to test the effect of task difficulty on regional cerebral blood flow (rCBF) as measured by positron emission tomography (PET). We report here the results from 10 young subjects engaged in match‐to‐sample tasks of progressively degraded faces. The tasks consisted of a control task, a face matching task with no stimulus degradation, and five levels of degradation: 20%, 40%, 50%, 60%, and 70%. Both performance accuracy and reaction times deteriorated significantly with increasing face degradation. There was a significant increase of rCBF in bilateral fusiform gyri during all face‐matching conditions compared to the control task, and bilateral prefrontal activation during the 70% degradation condition. Linear regression analyses revealed a significant increase of rCBF in the right prefrontal cortex, and linear decreases of rCBF in the striate and fusiform cortex as face degradation increased. Performance on the 70% task was correlated positively with rCBF in right prefrontal and bilateral fusiform gyri, and negatively with left prefrontal and striate rCBF. These results show that the right prefrontal, striate, and ventral extrastriate cortex are the principal brain regions that modulate their activity as this visual discrimination task becomes more difficult. The right prefrontal increase probably represents an increasing demand on working memory or attention, whereas decreased rCBF in the striate cortex may be due to changes in the characteristics of the stimuli, or to suppression of low‐level processing by one of a number of mechanisms. This experiment has implications both for the design of neuroimaging experiments, and for interpreting differences in rCBF activation between groups.

Dissociation of object and spatial vision in human extrastriate cortex: Age-related changes in activation of regional cerebral blood flow measured with [15 o]water and positron emission tomography

We previously reported selective activation of regional cerebral blood flow (rCBF) in occipitotemporal cortex during a face matching task (object vision) and activation in superior parietal cortex during a dot-location matching task (spatial vision) in young subjects, The purpose of the present study was to determine the effects of aging on these extrastriate visual processing systems. Eleven young (mean age 27 4 years) and nine old (mean age 72 7 years) male subjects were studied. Positron emission tomographic scans were performed using a Scanditronix PC10247B tomograph and H215O to measure rCBF. To locate brain areas that were activated by the visual tasks, pixel-by-pixel difference images were computed between images from a control task and images from the face and dot-location matching tasks. Both young and old subjects showed rCBF activation during face matching primarily in occipitotemporal cortex, and activation of superior parietal cortex during dot-location matching. Statistical comparisons of these activations showed that the old subjects had more activation of occipitotemporal cortex during the spatial task and more activation of superior parietal cortex during the object task than did the young subjects. These results show less functional separation of the dorsal and ventral visual pathways in older subjects, and may reflect an age-related reduction in the processing efficiency of these visual cortical areas.

Stability of metabolic and neuropsychological asymmetries in dementia of the Alzheimer type

We examined right/left asymmetries in cerebral glucose utilization and neuropsychological performance in 16 patients with dementia of the Alzheimer type (DAT) who were studied twice over a mean interval of 15 months. Neuropsychological asymmetry was expressed as the difference between performance on a visuoconstructive task and a language task. Test/retest comparisons showed significant declines on the neuropsychological measures, but no significant changes in the neuropsychological and metabolic asymmetry indices. At both evaluations, significant correlations were found between neuropsychological asymmetry and metabolic asymmetry in frontal and parietal cortex. Individual patterns of metabolic and neuropsychological asymmetry in these patients with DAT seem to be stable.

Neuropsychological and cerebral metabolic function in early vs late onset dementia of the Alzheimer type

Differences in age at onset of dementia of the Alzheimer type (DAT) have been associated with differences in cognitive impairments and course of the disease. To investigate this, we examined cerebral metabolism and cognitive performance in early and late onset DAT patients, who had equivalent duration and severity of illness. Regional cerebral metabolic rates for glucose were measured in the resting state by positron emission tomography using [18F]2-fluoro-2-deoxy-D-glucose. A cross-sectional analysis showed no significant differences between the two groups in performance on neuropsychological tests, but the early onset patients showed significantly more parietal metabolic dysfunction than did the late onset patients. Longitudinal analysis showed no significant differences between early and late onset patients in rate of cognitive decline over a mean interval of 19 months. Thus, our results do not support the hypothesis of different subgroups in DAT based on age at onset, nor suggest a faster rate of cognitive decline in younger patients.

Mapping the functional neuroanatomy of the intact human brain with brain work imaging

The recent development of noninvasive methods for measuring local rates of energy metabolism or blood flow in the brain has made it possible to investigate functional neuroanatomy in healthy human subjects. The best of these methods, high resolution measurement of regional cerebral blood flow (rCBF) with positron emission tomography (PET), provides a precision of anatomical localization that far exceeds that attainable with human brain lesion studies. Moreover, the study of healthy subjects avoids possible confounding effects of brain lesions, such as compensatory reorganization of brain function. PET-rCBF studies have already identified several cortical areas involved in higher-order visual processing, indicating that functional neuroimaging may yield a map of human visual cortex analogous to maps that have been developed by vision research in nonhuman primates. PET-rCBF studies of imagery and language demonstrate the potential of functional neuroimaging to map regions of human cortex that perform functions that cannot be studied so easily in nonhuman primates or perform functions that humans do not share with other species.