Purpose: We sought to determine the cause of cerebellar dysfunction in
epilepsy and whether this dysfunction was directly related to seizure
s. Methods: Cerebellar metabolism was evaluated in 48 patients with a
well-defined region of seizure onset and with corresponding hypometabo
lism. Regions of interest (ROI) were drawn according to a standardized
template. If the ROI/nonepileptogenic cortex count rate ratio was out
side the 95% confidence interval (CI) of controls, the ROI was defined
as abnormal. The ratios from cerebellar hemispheres (defined as ipsi-
or contralateral to the seizure onset region), were compared among co
ntrols (n = 8); patients who had seizure onsets and corresponding hypo
metabolism mesially in a temporal lobe (patient group 1, n = 19); pati
ents whose seizures had onset mesially in a temporal lobe but spread r
apidly to the ipsilateral frontal lobe and who had hypometabolism both
in the affected temporal lobe and frontal lobe (patient group 2, n =
23); and patients who had seizure onsets and corresponding hypometabol
ism in the frontal lobe (patient group 3, n = 6). Results: Significant
hypometabolism was noted in the contralateral cerebellum of patients
in groups 2 and 3 [p = 0.007 and p = 0.008, respectively; two-way anal
ysis of variance (ANOVA)]. In contrast, patients in group 1 tended to
have lower values in the ipsilateral cerebellum (p = 0.057). Conclusio
ns: The observed cerebellar changes are consistent with animal data sh
owing that cerebellar connections to frontal lobes are numerous and cr
ossed, whereas the connections to mesial temporal lobes are less abund
ant, bilateral, with an ipsilateral predominance. The difference betwe
en the two groups of patients with mesial temporal seizures suggests t
hat cerebellar dysfunction in partial epilepsy, at least to a certain
extent, is related to mechanisms involved in seizure generation and sp
read.