The metabolic relationship between one cerebral hemisphere and the opposite cerebellum was studied using positron emission tomography (PET) and a double-study method with [18-F] fluorodeoxyglucose (FDG). Subjects were 23 normal volunteers aged 46 to 78 years (mean ± S.D. : 61 ± 10).
All subjects were studied by a procedure incorporating two sequential scans carried out during different behavioral states. For the initial scan, the subjects were at rest, with eyes closed and lightly blindfolded. A bolus of 2.53.5 mCi of FDG was administered intravenously; 30 minutes later, a 20-minute PET scan was performed. Immediately thereafter, an activation paradigm (verbal memory task, described later) was commenced. Five minutes later, a second 2.53.5 mCi bolus of FDG was administered. The activation tesk was continued for the following 30 minutes. The subjects were then repositioned in the scanner in a manner identical to that in the first scan and a second 20-minute PET scan was performed.
In the verbal memory task, subjects were asked to read passages similar to those in the Wechsler Adult Intelligence Scale and immediately thereafter to recall what had been read. Before the scanning session, subjects were familiarized with the task and, during the actual session, were urged to attempt to recall the passages as thoroughly as possible.
Cerebral metabolic rate for glucose (CMRglc) was calculated in a total of 12 pairs of bilateral regions between 10 and 100 mm above the IOM line. Regional CMRglc was calculated by obtaining a mean value for all boxes in the prefrontal, premotor, orbitofrontal, motor, sensory, superior parietal, inferior parietal, superior temporal, medial temporal, occipital, deep grey (basal ganglia + thalamus) and cerebellar regions in each hemisphere.
During the 35-mimute period alloted to the task, the subjects read an average of 18.6 ± 2.6 passages. The number of recalled passages averaged 50.6 ± 14.5%. The mean CMRglc value at rest was 6.0 ± 1.3 mg/100 g/min and no significant right-left asymmetry in either cerebral or cerebellar CMRglc was noted. The CMRglc during activation was 7.1 ± 1.0 mg/100 g/min, showing an 18.3% increase. During activation, regional side-to-side asymmetries of CMRglc were produced. The premotor, orbitofrontal and motor regions in the left hemisphere showed significantly higher CMRglc increases than the corresponding regions in the right hemisphere, whereas the right cerebellum showed a significantly higher CMRglc increase than the left cerebellum.
The regional CMRglc increase data were analyzed in terms of regional (Left-Right) values of CMRglc increase. Correlations of these asymmetrical indices between various regions in the hemisphere and the cerebellum were examined. In Pearson correlation analysis, the motor region showed the highest correlation coefficient (r=-0.6, p=0.003) with cerebellum, followed by the premotor (r=-0.48, p=0.02), the prefrontal (r=-0.41, p=0.06) and the sensory (r=-0.41, p=0.06) regions. In multiple regression analysis, the only variable correlating with the cerebellum was the motor region (t=-2.6, p=0.02).
The clinical approach used here has provided one of the first quantitative demonstrations of highly specific metabolic coupling between the cerebral region and the cerebellum. Our finding suggests that the metabolic activation in the motor region during a verbal memory task is best coupled with the metabolic activation in the cerebellum.
View full abstract