Brain Connectomics in Early Adolescence Predict Suicidal Ideation Severity in Later Adolescence

Suicidal ideation (SI) typically emerges during adolescence but is challenging to predict. Given the consequences of SI, it is important to identify neurobiological and psychological predictors of SI in adolescents in order to improve suicide prevention strategies. We used graph theoretical methods to examine whether, and which, local properties of functional brain topology in early adolescence can predict severity of subsequent SI. In 109 participants (61 female) ages 9-13 years, we assessed clinical and participant characteristics and obtained resting-state fMRI data (baseline). We examined inter-regional functional connectedness across 250 brain regions by computing measures of nodal interconnectedness: local efficiency, eigenvector centrality, nodal degree, within-module z-score, and participation coefficient. A LASSO regression identified a linear combination of the most important predictors of SI severity measured approximately 4 years later. The LASSO analysis identified a combination of 8 predictors of future SI severity (R2=.23) including internalizing symptom severity as the largest predictor (β=.13). The remaining predictors were brain-based including nodal degree of the inferior frontal gyrus (β=-.08), precentral gyrus (β=-.07), fusiform gyrus (β=.05), and inferior temporal gyrus (β=.03), within-module degree of the substantia nigra (β=.06) and inferior parietal lobe (β=-.03), and eigenvector centrality of the subgenual cingulate gyrus (β=.03). Our findings suggest that along with early symptoms of internalizing behaviors, network properties of the brain in early adolescence, a period when SI may not be clinically evident, are markers of vulnerability for SI severity later in adolescence. Research is needed to validate the clinical utility of these markers as predictors of suicidal thoughts.