BRIEF RESEARCH REPORT article
Front. Hum. Neurosci.
Sec. Brain Imaging and Stimulation
Volume 19 - 2025 | doi: 10.3389/fnhum.2025.1551168
This article is part of the Research TopicMachine Learning Algorithms for Brain Imaging: New Frontiers in Neurodiagnostics and TreatmentView all 13 articles
An Adaptive Transformer-Based Framework for Advanced Precision Brain Activity Mapping and Intelligent Neurotherapeutic Decision Support Systems
Provisionally accepted
- 1Shree L. R. Tiwari College of Engineering, Mumbai University, Thane, Maharashtra, India
- 2Mukesh Patel School of Technology Management and Engineering, SVKM's Narsee Monjee Institute of Management Studies, Mumbai, India
- 3New Jersey Public Health Association, New Brunswick, New Jersey, United States
- 4IEEE Dallas County, Dallas, North Carolina, United States
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
The identification and treatment of neurological disorders rely significantly on brain imaging and neurotherapeutic decision-making, with EEG signals being a primary tool despite their complexity, noise, and non-stationary nature. This study presents an Adaptive Transformer-based approach that enhances attention mechanisms to effectively extract temporal and spatial relationships in EEG data. The methodology involves preprocessing EEG signals to remove noise and segment them into time-series chunks, followed by channel-wise embeddings and temporal encoding for effective data representation. The proposed transformer architecture incorporates spatial attention for inter-channel interactions, multihead self-attention for capturing temporal dependencies, and an adaptive attention mask for domain-specific adjustments. The model was evaluated on publicly available datasets, including TUH EEG Corpus and CHB-MIT, using accuracy, precision, recall, and F1-score as performance metrics. Results showed that the proposed method outperformed conventional models like CNNs and LSTMs, achieving 98.24% accuracy, with improved precision and recall, allowing for the identification of subtle EEG patterns. Additionally, attention maps highlighted key spatial and temporal regions, enhancing interpretability in clinical applications. The Adaptive Transformer emerges as a powerful tool for EEG-based neurotherapeutic applications, contributing to improved medical decision support and deeper insights into brain function. Future research could explore subject-specific optimizations and real-time system integration to enhance clinical applicability further.
Keywords: EEG signal analysis, transformer architecture, Neurotherapeutic Decision Support, temporal-spatial modeling, Precision Brain Imaging, Adaptive attention mechanism, Machine Learning in Neurology
Received: 24 Dec 2024; Accepted: 05 Sep 2025.
Copyright: © 2025 Nemade, Kulkarni, Shah, Patel, Patel and Goradiya. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Vikram Kulkarni, Mukesh Patel School of Technology Management and Engineering, SVKM's Narsee Monjee Institute of Management Studies, Mumbai, India
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.