Brian Keane from Rochester will present.

Thalamic and cortical sensory dysconnectivity as a biomarker for psychosis

Past work has shown that people with psychosis exhibit thalamo-cortical hyperconnectivity and cortico-cortical hypoconnectivity in sensory networks.  Do these dysconnectivity patterns apply to all sensory networks?  Can such results form the basis of a viable biomarker?  To address these questions, we analyzed publicly available resting-state data from 105 early-stage psychosis patients and 54 healthy controls. We found that both affective and non-affective psychosis patients exhibited cortico-cortical hypoconnectivity and thalamo-cortical hyperconnectivity in somatomotor and secondary visual networks but not in the auditory or primary visual networks. These dysconnectivity patterns could be combined in a simple way to generate a robust “somato-visual” biomarker (p = 2e-10, Hedges’ g = 1.05). This biomarker was present in antipsychotic-naive patients and did not depend on over a dozen confounds (e.g., stress, nicotine, comorbidities, sociodemographics). The biomarker had moderate test-retest reliability (ICC = 0.62), could be recovered in five-minute scans, could generalize to other data sets, and could not be detected in a clinical control group. Preliminary analyses using diffusion MRI show that there were no corresponding changes in the structural connectome, suggesting a possible molecular or synaptic origin to the biomarker.  This work sets the stage for future studies aiming to develop interventions, predict future illness onset, or stratify patients into clinically meaningful subgroups.