AI Methods Compared for Interpreting EEG Models in Depression Detection

A new study published on arXiv explores various post-hoc explainable AI (XAI) methods to interpret black-box EEG models used for detecting Major Depressive Disorder (MDD). Researchers applied techniques like DeepSHAP, Integrated Gradients, GradCAM, Occlusion, and Permutation Feature Importance to an InceptionTime architecture. The analysis revealed partially overlapping attribution patterns, with a recurring emphasis on specific EEG regions, particularly in the right hemisphere. While some methods showed agreement, others produced distinct results, highlighting how different XAI approaches can influence interpretations of EEG-based deep learning models for psychiatric applications. AI

IMPACT This research explores how AI interpretability methods can be applied to medical diagnostics, potentially improving trust and understanding in AI-driven healthcare solutions.

RANK_REASON The cluster contains an academic paper detailing a comparative study of AI methods. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• EEG
• GradCAM
• InceptionTime
• Integrated Gradients
• Major Depressive Disorder
• Nikolina Frid PhD
• Permutation Feature Importance

• paper
• safety

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →