An fMRI Pilot Study of Cognitive Flexibility in Trichotillomania

Abstract

Trichotillomania is a relatively common psychiatric condition, although its neurobiological basis is unknown. Abnormalities of flexible responding have been implicated in the pathophysiology of obsessive-compulsive disorder and thus may be relevant in trichotillomania. The purpose of this study was to probe reversal learning and attentional set-shifting in trichotillomania. Twelve adults with trichotillomania and 13 matched healthy control subjects undertook a functional MRI task of cognitive flexibility. Group-level activation maps for extradimensional and reversal switches were independently parcellated into discrete regions of interest using a custom watershed algorithm. Activation magnitudes were extracted from each region of interest and study subject and compared at the group level. Reversal events evoked the expected patterns of insula and parietal regions and activity in the frontal dorsal cortex extending anterior to the frontal poles, whereas extradimensional shifts evoked the expected frontal dorsolateral and parietal pattern of activity. Trichotillomania was associated with significantly increased right middle frontal and reduced right occipital cortex activation during reversal and set-shifting. Elevated frontal activation coupled with reduced activation in more posterior brain regions was identified. These pilot data suggest potentially important neural dysfunction associated with trichotillomania.

Keywords: Cognition; Flexibility; Imaging; Trichotillomania.

Figure 1. Analyses of regional activity observed at the group level.

a) Extradimensional switches showed activation at the group level across frontal dorso-lateral & parietal regions (1-Sample t-test, FDR cluster-corrected p<0.05). b) Reversal Learning events showed activation at the group level across dorsal fronto-parietal activity that extended to the frontal poles & anterior insula (1-Sample t-test, FDR cluster-corrected p<0.05). Group level activation maps were parcellated into discrete regions of interest and compared using 2-sample t-tests across the control and TTM groups (uncorrected p<0.05). ROI’s showing significant differences in activation magnitude are shaded in black for the extradimensional switches (bi – Table 4 & Reversal learning events (bii – Table 5). Anatomical labels come from the AAL atlas and were assigned by the mode voxels of an ROI.