Identifying diagnostic neuroimaging biomarkers for adolescent major depressive disorder

The increasing incidence of adolescent depression represents a serious public health concern. Despite clear diagnostic criteria, the wide range of symptoms and their overlap with other psychiatric disorders make it difficult to provide effective personalized treatment in adolescents. The integration of resting-state functional magnetic resonance imaging (rs-fMRI) and machine learning has shown promise in identifying diagnostic biomarkers and shedding light on personalized treatments in adult depression. However, equivalent studies in adolescent depression are lacking. Therefore, the present study aimed to identify diagnostic rs-fMRI biomarkers for adolescent depression.

Phenotypic and rs-fMRI data of 127 adolescents (64 adolescents with depression; 63 healthy controls) were acquired from the Boston Adolescent Neuroimaging of Depression and Anxiety dataset. Partial correlation was used to compute the functional connectome of the whole brain. Repeated nested cross validation with Boruta feature selection and support vector machine was employed to build a classification model to discriminate adolescents with depression from healthy controls.

The classification model identified 46 fine-scale connectivity features of the functional connectome as co-biomarkers in adolescent depression. The connectivity between the right medial/superior temporal gyrus and left pars triangularis/rostral middle frontal gyrus, as well as between the right medial orbitofrontal/ rostral anterior cingulate cortex and right precuneus/isthmus cingulate gyrus were identified as the most important features in adolescent depression.

The identification of a novel neuroimaging composite-biomarker panel here sheds light on depression diagnosis in adolescence. The retention of anatomical resolution within these composite biomarkers may facilitate the development of individualized neuromodulation treatment strategies.