Effective Feature Representation and Attention-based Dense LSTM for Human Brain Activity Recognition using EEG Signals during Yoga and Meditation
Bandari Ramaraju, Ravichander Janapati, Sreedhar Kollem
2025 2nd International Conference on Artificial Intelligence for Innovations in Healthcare Industries December 4, 2025 DOI: 10.1109/icaiihi67124.2025.11403078 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Computational modeling study Peer reviewed |
|---|---|
| Topics | Meditation |
| Key points | The authors propose an attention-based A-DLSTM framework that classifies emotions from EEG signals recorded during meditation and yoga, using features extracted by Restricted Boltzmann Machines plus spectral and wave-based descriptors. They report that this model outperforms conventional approaches on multiple performance metrics, which they present as evidence of its effectiveness for EEG-based brain activity recognition. |
Abstract
Human psychological activities and biological processes are closely linked to brain function, which governs both mental and physical states. Evaluating human activity is vital for well-being, as environmental and physical factors directly affect health. Electroencephalography (EEG) is a non-invasive technique for recording brain signals; however, these signals are often corrupted by non-cerebral artifacts, reducing reliability. Consequently, EEG-based brain activity recognition remains challenging. To address these limitations, this study proposes an attention-based human brain activity recognition framework for emotion classification during meditation and yoga. EEG signals are collected from publicly available datasets, and discriminative features are extracted using Restricted Boltzmann Machines along with spectral and wave-based descriptors. The extracted features are then processed by an Attention-based Dense Long Short-Term Memory (A-DLSTM) network. Experimental evaluation using multiple performance metrics demonstrates that the proposed model outperforms conventional approaches, confirming its effectiveness in accurate human brain activity recognition.