We investigated whether the intersession variability of serial fMRI studies
correlates between two activation modalities, i.e. a standardized visual a
nd a standardized motor task. Six volunteers were scanned in at least weekl
y intervals. The number of pixels activated as well as the activation ampli
tude varied widely. The maximal difference of the number of pixels activate
d was 1150%, of the activation amplitude 250%. In three volunteers, the var
iability was highly correlated between the two tasks. Three other volunteer
s showed one outlier each. We conclude that the intersession variability is
due to global factors affecting the whole brain, but that due to unpredict
able outliers, using a standardized task to normalize the data of interest
is of limited value. NeuroReport 11:3843-3847 (C) 2000 Lippincott Williams
& Wilkins.