Evaluating the performance of EEG based on ANN to predict the effectiveness of tDCS combined evaluative conditioning on obsession symptoms reduction in contamination OCD patients: Secondary analysis of data from a randomized controlled trial.

Authors

Fateme Asadollahzadeh Shamkhal, Ali Moghimi, Hamid Reza Kobravi, Javad Salehi Fadardi, Faezeh Raeis Al Mohaddesin

Published in

PloS one. Volume 21. Issue 8. Pages e0354614. Epub Aug 20, 2026.

Abstract

Over 40% of Obsessive-Compulsive Disorder (OCD) patients do not respond to common treatments. This study was a secondary analysis of data from a randomized controlled trial, for predictability of effectiveness of Transcranial Direct Current Stimulation (tDCS) with Contamination-Based OCD (C-OCD) using artificial neural networks (ANN) and electroencephalography (EEG) signals. Out of 54 C-OCD patients, 48 were randomized into 4 groups. Using a 2 × 2 factorial design, each group received a combined intervention (real or sham) based on tDCS and Disgust Reduction Evaluative Conditioning (DREC) in 10 sessions. Evaluations included the Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) and EEG recordings at rest (eyes open). Among the various features extracted from EEG (Fuzzy Synchronization Likelihood (FSL), Power Spectrum, and Recurrence Quantification Analysis (RQA), the Relief algorithm identified appropriate features based on intervention effectiveness for each frequency band and group allocation. The obsessive symptoms reduction was predicted using appropriate features, the fully connected feedforward network and Radial Basis Functions (RBF) networks. The ANN inputs were the appropriate features extracted from the EEG signal before the interventions. To predict the effectiveness level, the Y-BOCS change score (pre- to post-intervention) was given to the models as the desired output. Both networks had the best results for the first three proposed Relief features. The fully connected feedforward network optimized by Gray Wolf Optimizer (GWO) achieved the best predictive performance with an average RMSE 0.57 ± 0.4. Based on the calculated error and comparison with Y-BOCS change intervals, EEG data from C-OCD patients combined with ANN effectively predicted the extent to which each type of intervention can change patients' Y-BOCS score. So, the therapist can decide whether to perform or select each type of intervention for the patient before starting treatment. Our findings confirm the feasibility of using pre-intervention EEG signals combined with ANN to predict individual responses in C-OCD patients.

PMID:
42623372
Bibliographic data and abstract were imported from PubMed on 21 Aug 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.