Network Changes in Insula and Amygdala Connectivity Accompany Implicit Suicidal Associations

Limited knowledge exists regarding the neurobiology of suicidal thoughts, given that there are currently no direct probes of the suicidal state. This pilot study used magnetoencephalography (MEG) to evaluate correlates of the implicit association between the self and death compared to the self and life as objective markers of suicide risk. Healthy volunteers (HVs; n=21) completed a modified version of the Suicide Implicit Association Task (S-IAT) during MEG scanning. Gamma power-which is considered a proxy measure of excitation-inhibition balance-was directly compared in the self-death/self-life contrast. As a proof-of-concept, the ability of dynamic causal modeling to categorize HVs versus four individuals with recent suicide crisis (SC) was evaluated. In HVs, enhanced gamma power in both amygdala and anterior insula were found for the self-death compared with self-life contrast. In addition, connectivity estimates between early visual cortex, anterior insula, and amygdala correctly categorized SC participants with 77% to 82% sensitivity and 80% to 85% specificity. These findings, which implicate network-level changes in salience network and amygdala connectivity in mediating suicidal associations, require further replication in larger samples. Direct probing of suicidal thoughts with the S-IAT may provide foundational markers of neural circuits associated with suicide risk.