PHILADELPHIA -- Magnetic resonance imaging has the potential for detecting very early brain changes related to memory impairment, Andrew J. Saykin, Psy.D., reported at the Ninth International Conference on Alzheimer's Disease and Related Disorders, presented by the Alzheimer's Association.

New data from the longitudinal Dartmouth Memory and Aging Study suggest that both structural and functional MRI are sensitive enough to detect changes in people who report significant cognitive complaints but who appear normal on neuropsychological testing. The findings have profound implications for early detection and potential intervention, said Dr. Saykin, professor of psychiatry and radiology at Dartmouth Medical School, Lebanon, N.H.

In current practice, structural MRI is used primarily to survey the brain for gross abnormalities such as lesions, atrophy, or stroke, while clinical use of functional MRI is restricted mostly to presurgical brain mapping. The Dartmouth group's research is aimed at broadening the use of the technology. "One has to look at brain structure, brain function, and also cognition, and the way these patterns fit together," Dr. Saykin told this newspaper.

Detailed neuropsychological and geropsychiatric assessments, genetic testing, and structural and functional MRI imaging were performed in 30 subjects with amnestic mild cognitive impairment (MCI), in 30 subjects who reported significant cognitive complaints (CC) but did not meet Mayo criteria for MCI on neuropsychological testing, and in 30 healthy controls.

Structural MRI included volumetric measurements of the hippocampus, entorhinal cortex, and amygdala, as well as voxel-based morphometry, Functional MRI measured brain activity associated with semantic, episodic, and working memory, he said in a poster presentation.

Cross-sectional baseline analysis showed a fairly consistent pattern of abnormality across imaging modalities in the MCI and CC groups, with the CC group showing a strikingly similar pattern of structural and functional imaging to that seen in MCI.

For example, baseline hippocampal volumes were lower in both the MCI and CC groups compared with the healthy controls, while both medial temporal and cortical regions showed reduced gray matter in the MCI and CC groups compared with the controls. (See photos.)

One of the most important uses of this type of volumetric information will be in following patients over time to look for quantitative changes, noted Dr. Saykin, also director of Dartmouth's neuropsychology program and its brain imaging laboratory.

Similarly, functional MRI during a working memory task showed reduced frontal and parietal activation in the MCI and CC subjects, compared with the healthy controls, while those same subjects had greater activation in the retrosplenial regions. Areas of increased activation in the impaired subjects are believed to be compensatory--that is, additional tissue is being recruited in order to maintain function, he explained.

None of these subjects were depressed. "When people come in for testing and do quite well but feel that they're not doing well, I usually look for indications of depression or other explanations. Our current data suggest that if people have memory complaints, there may be changes in their brain structure and activity even if they perform within normal limits on formal testing," Dr. Saykin remarked.

In a subgroup of nine of the MCI subjects, increases in frontal lobe activation correlated with improved test performance after treatment with donepezil. These findings were recently published (Brain 2004;127:1574-83). The ability of MRI to detect changes induced by cholinesterase inhibitors or other pharmacologic agents suggests that routine imaging at certain intervals during treatment may help guide treatment decisions, he said.

Other aspects of the Dartmouth study, presented in separate posters by Dr. Saykin's colleagues in the school's psychiatry department, also supported the potential use of MRI for detecting early brain changes.

Paul J. Wang led a morphometric study in which he measured the area of the corpus callosum in 22 patients with early Alzheimer's disease, 28 with amnestic MCI, 28 with subjective cognitive complaints but no deficits on psychometric testing, and 50 healthy age-matched controls.

Significant differences in area were found between the groups in the callosal subregion that included the isthmus and splenium. Adjusted for intracranial volume, that area was 203.4 [mm.sup.2] in the healthy controls, 189.05 [mm.sup.2] in the CC group, 188.82 in the MCI subjects, and 180.95 [mm.sup.2] among those with early AD.

These findings suggest that changes in the corpus callosum may occur early in dementia, Mr. Wang said.

In a third study led by Laura A. Rabin, Ph.D., functional MRI scans were conducted in 25 subjects each with MCI and CC, and 25 healthy controls during a test of "novelty detection."

Both the MCI and CC groups showed increased brain activation relative to the healthy controls in medial temporal and frontal regions of the brain. The expanded activation correlated negatively with verbal learning scores, Dr. Rabin reported.

Finally, a fourth study pointed to the potential utility of combining genotyping with functional neuroimaging to enhance early detection and tracking of memory disorders.

Heather A. Wishart, Ph.D., led the evaluation of 40 cognitively intact adults, 22 of whom had the apolipoprotein E (apoE) c3/3 genotype (mean age 68.5 years) and 13 with the g3/4 genotype (mean age 63.4). There were no differences between the two groups in neuropsychological test performance, mood, or other cognitive assessments. But the group with the Alzheimer's disease-associated apoE c4 allele showed increased brain activation compared with those without that allele, most prominently in the frontal regions bilaterally and in the left paracentral lobule.

Even after controlling for age and sex, there were no regions in which those without the apoE e4 allele had greater activity than those with it, Dr. Wishart noted.

According to Dr. Saykin, the MRI equipment used in all of these studies is already available in most major medical centers, except for some additional software needed for the functional imaging. Still lacking, though, "is the specialized expertise to design appropriate memory stimuli, deliver them to the patient under computer control, and to analyze the functional scan data," he said.