Neurocognitive and Social Cognitive Predictors of Adolescent Major Depressive Disorder: A Machine Learning...

Area of Science:

• Computational psychiatry
• Neuroscience
• Machine learning applications in mental health

Background:

• Major Depressive Disorder (MDD) is a prevalent mental health condition in adolescents.
• Accurate differentiation between adolescents with MDD and healthy controls (HC) is crucial for timely intervention.
• Neurocognitive deficits are often observed in individuals with MDD.

Purpose of the Study:

• To evaluate the efficacy of machine learning (ML) algorithms in differentiating adolescents with MDD from HC using neurocognitive data.
• To identify key neurocognitive features that contribute to the accurate classification of MDD in adolescents.
• To explore the potential of ML-based cognitive profiling for early detection of MDD.

Main Methods:

• Neurocognitive functions were assessed in adolescents diagnosed with MDD and HC using a comprehensive battery of tests.
• A tree-based approach was employed for feature selection, followed by the implementation of various ML algorithms.
• Techniques such as Synthetic Minority Over-sampling Technique and stratified 10-fold cross-validation were used to address class imbalance and optimize model performance.
• Shapley Additive Explanations (SHAP) values were utilized for interpreting feature contributions.

Main Results:

• The study included 117 adolescents with MDD and 67 HC.
• Support Vector Classifier (SVC) demonstrated the highest performance, achieving a mean accuracy of 76.0% and an Area Under Curve (AUC) of 79.0%.
• SHAP analysis identified symbol coding, categorical fluency, and Stroop Test parameters as the most influential features in distinguishing between groups.

Conclusions:

• Machine learning techniques show significant potential in accurately differentiating adolescents with MDD from HC based on neurocognitive assessments.
• Cognitive domains related to processing speed and executive functions are critical indicators for MDD in adolescents.
• ML-based cognitive profiling may serve as a valuable tool for supporting the early recognition and diagnosis of MDD in adolescent populations.