Brian Keane¹, Yonatan Abrham¹, Kirsten Peterson², Michael Cole², Paul Geha¹, Skylar DeWitt¹; ¹University of Rochester, ²Rutgers University, Newark

Psychosis patients functionally exhibit thalamo-cortical hyperconnectivity and cortico-cortical hypoconnectivity in somatomotor and secondary visual (visual2) networks. These dysconnectivity patterns have been combined to form a “somato-visual” biomarker that is highly robust, reliable, and generalizable across samples (Keane et al., 2024, Mol. Psych.). Can traces of this biomarker be found in white matter? To address this question, we leveraged diffusion MRI data from the Human Connectome Early Psychosis project, which included 33 healthy controls (22 males; mean age=25) and 86 early-stage psychosis patients (50 males; mean age=24). Diffusion data were preprocessed using QSIPrep. Tractography was performed using DSI Studio. Structural connectomes (361x361) were derived from the thalamus plus 360 cortical parcels. Thalamo-cortical connectivity was computed by averaging streamline counts between the thalamus and visual2/somatomotor networks. Cortico-cortical connectivity was computed by averaging streamline counts between cortical parcels of the same two networks. To compute the structural variant of the biomarker, we normalized and subtracted the two averaged values (thalamo-cortical – cortico-cortical). To provide a comparison, we re-derived the functional variant of the biomarker from the same subjects in the same atlas space using already-published connectomes (ibid). Patients exhibited expectedly higher functional biomarker values (p=5e-07, Hedges’ g=1.0) but non-significantly lower structural biomarker values (p=.10, g=-.3; BF01=14.1, one-tailed). The functional and structural biomarker variants were uncorrelated across subjects (r=-.05, p=.66). The foregoing null results could not be ascribed to noisy data since the connectomes passed numerous quality checks, e.g., showing larger structural/functional correlations within- than between-subjects (p=9e-07) and larger thalamo-cortical structural connectivity with primary versus secondary visual networks (p=6e-06). To conclude, somato-visual functional dysconnectivity in early psychosis may emerge synaptically (e.g. due to pathology to the NMDA receptor) without corresponding large-scale changes in white matter. A next goal is to understand the behavioral consequences of this dysconnectivity (e.g., oculomotor control, visually-guided reaching, perceptual organization).

Acknowledgements: This work was supported by a Del Monte Institute pilot grant to BPK