Evaluation Traps in EEG Disease Classification: Identity Leakage, Lucky Folds, and Objective Mismatch Replicated Across AD, FTD, MDD, and SCZ

June 10, 2026Open Access

View Full Paper

Ask AI

Bookmark

Share

Key Points

This research aims to investigate evaluation traps in EEG disease classification, focusing on identity leakage and objective mismatches.
Review of current EEG classification practices across various diseases including AD, FTD, MDD, and SCZ.
Analysis of identity leakage and its impact on classification robustness.
Exploration of fold selection and objective mismatch in EEG applications.
Identification of identity leakage leading to inflated performance metrics across all evaluated diseases.
Demonstration of inconsistent results due to lucky folds in EEG classification experiments.
Highlighting significant objective mismatches, questioning the trust in EEG-derived diagnoses.

We only have this paper's title and publication details.

Evaluation traps affect EEG disease classification in conditions like AD, FTD, MDD, and SCZ, suggesting caution is needed.

Preprint, currently under peer review.

No takes yet. Share an insight, caveat, or question.

Adel Sahuc (2026) studied this question.

View Full Paper

Ask AI

Bookmark

Share