Neuroimaging Biomarkers of Suicide Risk: A Narrative Review of Cross-Disorder Evidence and Clinical Implications

ანოტაცია

Suicide is recognized as a deliberate act of self-destruction. A major share of deaths worldwide continues to be caused by suicide, despite a decline in its global rates since 1990. Suicide is commonly associated with psychiatric and chronic medical disorders. Suicide risk is poorly predicted by current clinical assessments. Recently, neuroimaging methods have been increasingly used to identify biological markers. Underlying neural abnormalities linked with suicidal behavior are thought to be reflected by these markers. The present review was performed to summarize cross-disorder evidence on neuroimaging biomarkers of suicide risk. Their possible clinical applications were also discussed. A structured literature search was performed in PubMed, Scopus, Web of Science, and APA PsycInfo for studies published from 2015 to 2025. Studies involving in-vivo neuroimaging techniques such as MRI, fMRI, DTI, PET, and MRS were included. Clear measures of suicidal ideation, attempts, or suicide death were required. Relevant data were extracted and organized thematically across structural, functional, and molecular categories. Findings were synthesized narratively. Overlapping and disorder-specific neural signatures were identified. Consistent evidence was observed for reduced gray and white matter volumes in the ventromedial and dorsolateral prefrontal cortex, anterior cingulate cortex, insula, hippocampus, amygdala, and thalamus. Altered connectivity between prefrontal control regions and limbic emotion centers was frequently detected. Disruptions in default mode, salience, and executive control networks were also found. Irregularities in serotonergic and dopaminergic activity were shown by PET findings. These irregularities were related to impulsivity and mood dysregulation. These abnormalities were found across several psychiatric disorders. Suicide vulnerability is indicated to potentially arise from shared fronto-limbic dysfunction, rather than disorder-specific pathology. Research is constrained by small samples, methodological differences, and cross-sectional designs. Larger, longitudinal, and multimodal studies are needed in future work. Support from AI and machine learning is required. Prediction accuracy will be improved, and individualized prevention strategies will be enabled. Suicide prevention is potentially transformed into a more precise, brain-based clinical practice by the early detection of neural biomarkers.