Intellectual Property India Publishes Patent Application for Adhd Disease Classification System Using Artificial Intelligence And Machine Learning With Eeg Signal Analysis

MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202621058345 A) filed by Sayali Harish Zade; Dr. Shrikant Sonekar; Dr. Pravin Kshirsagar; Prof. Avinash Ikhar; Dr. Pradnya Morey; Dr. Sanjay Haridas; Dr. Harna Bodele; Ms. Maitraee Kengale; Ms. Muskan Naidu; Mr. Sujal Mandpe; and Mr. Swayam Bhute on May 07, 2026, for Adhd Disease Classification System Using Artificial Intelligence And Machine Learning With Eeg Signal Analysis.

Inventors include Sayali Harish Zade; Dr. Shrikant Sonekar; Dr. Pravin Kshirsagar; Prof. Avinash Ikhar; Dr. Pradnya Morey; Dr. Sanjay Haridas; Dr. Harna Bodele; Ms. Maitraee Kengale; Ms. Muskan Naidu; Mr. Sujal Mandpe; and Mr. Swayam Bhute.

The application for the patent was published on July 31, 2026, under issue no. 31/2026.

Abstract: The present invention provides a complete ADHD Disease Classification System using Artificial Intelligence and Machine Learning techniques applied to Electroencephalography (EEG) brain signals. The system processes 19-channel EEG data (Fp1, Fp2, F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T3, T4, T5, T6, Fz, Cz, Pz) through a multi-stage machine learning pipeline comprising StandardScaler normalization, SMOTE class balancing, and XGBoost gradient boosting classification. Trained on 2,166,383 synthetic EEG samples with an 80-20 train-test split, the model achieves 77.84% classification accuracy. The system is deployed through a Flask REST API backend and an interactive clinical web interface enabling clinicians to input EEG values, receive real-time ADHD/Non-ADHD predictions with confidence scores, visualize per-channel feature importance, and generate PDF diagnostic reports. The XGBoost ensemble method outperforms alternative models including CNN, Gradient Boosting, and Random Forest. Feature importance analysis confirms that frontal lobe channels (Fp1, Fp2, F3, F4) are the most diagnostically significant, consistent with established ADHD neuroscience. The system serves as an objective clinical decision support tool to supplement existing subjective ADHD diagnostic methods.

Disclaimer: Curated by HT Syndication.