Late-life depression is a severe mental health condition affecting older adults. It is associated with an increased risk of suicide, which is higher in this population compared to any other age group. The prevalence of depression in older adults is rising, highlighting the need for effective screening of suicide risk in this population.
In this study, researchers used machine learning to predict suicide risk in late-life depression patients by analyzing whole-brain resting-state functional connectivity and white matter structural connectivity data.
A recent study published in Nature has shown that late-life depression affects almost 20% of the aging population, with over one-third of patients unable to achieve complete remission after treatment. Suicidal ideation, plans, and behavior are serious health issues among older adults, with a higher likelihood of death than any other age group.
While previous research has identified risk factors and questionnaires to measure general suicidal tendencies, there is considerable heterogeneity within individuals with suicidality, making it challenging for patient management. To address this, the study employed connectome-based predictive modeling using whole-brain resting-state functional and white matter structural connectivity data to predict suicide risk in late-life depression patients.
The study found that brain patterns could predict suicide risk, with the explained variance up to 30.34%, improving classification-prediction accuracy compared to using questionnaire scores alone. The findings suggest that multimodal brain connectivity could capture individual differences in suicide risk among late-life depression patients, potentially helping clinicians identify patients who need detailed assessments and interventions.
It was found that brain connectivity features could be used to predict the severity of suicide risk in a heterogeneous population with Late-Life Depression (LLD). The study used multimodal neuroimaging data and a machine learning approach to demonstrate the effectiveness of connectome profiles for solving a classification-prediction problem.
The research team found that the network strength of FC (functional connectivity) and SC (structural connectivity) profiles showed discriminant between-group differences and improved classification-prediction accuracy. The CPM (connectome-based predictive model) models were further generalized to classify groups with different levels of suicide risk in external datasets.
The findings suggest that brain connectivity derived from a data-driven procedure might provide valuable information about an LLD patient’s past and current suicide risk. CPM models using FC and SC can capture the variability of suicide risk among LLD patients. However, the predictive models are currently tricky to implement due to several factors.
First, diagnosing and managing patients with suicide risk may not allow the acquisition and analysis of imaging data. Second, clinicians need to receive training in using predictive methods. Third, the predictive models were not tested for predicting future suicide risk, and future studies need to apply a longitudinal design or develop further models to predict future suicide risk.
Several other factors have to be taken into account when interpreting the findings. Suicide risk increases among older adults with medical illnesses that cause disability. Medical and psychiatric medications might also affect suicide risk and brain connectivity in older adults. Sex-related differences in suicide risk and brain connectivity have been widely explored in the literature, and future studies should consider recruiting samples with a more balanced sex ratio.
The study has potential limitations, including a relatively small sample size and the exclusion of patients with comorbid psychiatric disorders or significant medical illnesses. Additionally, one of the external validations used the sample collected by the research team, which is not a perfectly stringent out-of-sample validation.
Nevertheless, the study highlights the potential for brain connectivity features to be used in predicting the severity of suicide risk in LLD patients, which could lead to improved clinical assessment and management.
