EEG Signal Classification: From Brain Activity to Text Representation

Read full text Summarise with VIRAT
Authors: R. Jeen Retna Kumar, R. S. Pavithra, A. Mallikarjuna Reddy, A. Phanendra Kumar, Victor Daniel, Varsha Ranjalkar
Cited by: 0
References: 9

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Electroencephalography(EEG) is a non invasive technique for recording the scalp potentials that reflects the synchronous neural activity.This study explores a complete EEG to text framework that spans the data acquisition,preprocessing,feature extraction,supervised classification and for the symbolic natural language mapping.Using a public EEG emotion dataset,a Random Forest classifier achieved 87% accuracy in distinguishing between the three affective states (Positive,Neutral,Negative).Each predicted state is then represented by a short text templates to demonstrate the brain-to-text conversion.The results confirm that the time frequency features enable the meaningful decoding of emotional intent,while highlighting the limitations of current symbolic text generation.The Future directions include integration with a neural language models for context aware EEG-to-text generation.

Fields of Study & Subjects

Emotion and Mood Recognition EEG and Brain-Computer Interfaces Gaze Tracking and Assistive Technology

Publication details

Journal / source
Advances in engineering research/Advances in Engineering Research
Pages
331–341
Publication date
30 March 2026
Language
EN
Document type
Book chapter