Unveiling the neuroanatomical basis of antidepressant TMS using models of polysynaptic connectome communication

Transcranial magnetic stimulation (TMS) of the left dorsolateral prefrontal cortex (DLPFC) is a frontier intervention for refractory depression. The antidepressant effects of TMS are thought to involve communication between the DLPFC and the subgenual cortex (SGC), a deep cortical structure that is not accessible to non-invasive stimulation. Interestingly, DLPFC and SGC are not connected by a direct white matter fibre. TMS propagation between these regions is therefore polysynaptic, i.e., mediated by intermediate regions interlinked via white matter connectivity. This limits the use of current tractography-based methods—which only consider direct, “one-hop”, fibres—to understand the anatomical basis of DLPFC-SGC communication. Here, we use advances in network neuroscience to model the polysynaptic propagation of TMS from the DLPFC to the SGC via the connectome. We apply models of network communication to quantify the structural capacity for signal transmission between stimulation sites in the DLPFC and the SGC. The models identify polysynaptic routes comprising sequences of two or more white matter tracts, and can therefore delineate multi-hop pathways between gray matter structures that are not connected by a direct fiber. We show that DLPFC-SGC communication via the connectome predicts symptom remission in two independent cohorts of depression patients. We delineate two putative pathways underpinning this process. A primary subcortical pathway mediated by the caudate and nucleus accumbens, and a secondary cortical pathway comprising superior frontal and anterior cingulate structures. Our work sheds light into the mechanisms of antidepressant TMS and paves the way for stimulation protocols based on polysynaptic fiber pathways.