Guest Presenter: Dr. Andrew Pines (Psychiatry resident at Brigham and Women’s Hospital )

Talk Title: Lesions that induce psychosis localize to a common brain circuit involving the hippocampus

Abstract:

Background: Brain lesions resulting in psychosis provide a unique source of causal information. Mapping the connectivity of these lesions could reveal circuit-based therapeutic neuromodulation targets.
Objective: To assess whether lesions causing psychosis map to a common brain circuit.
Methods: We identified published cases of brain lesions associated with psychosis and screened them for causality. Functional connectivity of each lesion was estimated using a human connectome database (n=1000). To determine connections common to lesions causing psychosis, we performed a sensitivity analysis by examining regions covered by both a one-sample T-test and the overlap of the functional correlates of each lesion. We performed a specificity analysis using a Bonferroni-corrected two-sample t-test between the functional maps of lesions causing psychosis and two separate control groups: 490 patients from an observational stroke database, and 258 published cases with lesion-induced, nonpsychotic symptoms.
Results: 148 cases from the literature were determined to be causal of psychosis. One- sample T-test (threshold T>10) and functional overlap tests (70% overlap T>7) identified common functional connections in the posterior hippocampus, ventral tegmental area, anterior/medial dorsal nuclei of the thalamus. The Bonferroni-corrected two-sample t-test revealed the posterior hippocampus as the strongest functional connection distinguishing lesions causing psychosis from lesions that did not cause psychosis. To determine if the effect was driven by lesions intersecting the hippocampus or a distributed brain network, we repeated the sensitivity and specificity analyses after excluding lesions intersecting the hippocampus (n=26, determined by voxels included in the Brainnetome Atlas hippocampus) and found similar functional connections across the whole brain (r=0.98) with the same peak in the posterior hippocampus.
Conclusion: Lesions causing psychosis map to a common brain circuit. The bilateral posterior hippocampi are the most specific areas involved in this circuit. This circuit could provide a new target for neuromodulation.