Exploring the Relationship between Brain Hemisphere States and Frequency Bands through Deep Learning Optimization Techniques

2025-09-17

AI Summary

This study investigates the impact of EEG frequency-band specificity and optimizer selection on binary classification of left- versus right-hemisphere brain states. We systematically evaluate four optimizers—Adagrad, RMSprop, SGD, and FTRL—across Alpha to Gamma bands using three deep learning architectures: a deep fully connected network, a shallow MLP, and a CNN. Classification performance is quantified via accuracy, and SHAP-based interpretability analysis is employed to elucidate feature contributions. Results show Adagrad achieves peak performance in the Beta band, while RMSprop excels in the Gamma band; CNN attains the second-highest accuracy due to its capacity for spatial feature modeling, and the deep network demonstrates competitive capability in complex pattern recognition. Notably, Adagrad and RMSprop exhibit significantly greater training stability than SGD and FTRL. This work provides the first empirical evidence of synergistic frequency-band–optimizer interactions in EEG decoding, establishing a neurophysiologically grounded foundation for adaptive deep learning paradigms in brain–computer interface applications.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Optimization
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

This study investigates classifier performance across EEG frequency bands using various optimizers and evaluates efficient class prediction for the left and right hemispheres. Three neural network architectures - a deep dense network, a shallow three-layer network, and a convolutional neural network (CNN) - are implemented and compared using the TensorFlow and PyTorch frameworks. Results indicate that the Adagrad and RMSprop optimizers consistently perform well across different frequency bands, with Adadelta exhibiting robust performance in cross-model evaluations. Specifically, Adagrad excels in the beta band, while RMSprop achieves superior performance in the gamma band. Conversely, SGD and FTRL exhibit inconsistent performance. Among the models, the CNN demonstrates the second highest accuracy, particularly in capturing spatial features of EEG data. The deep dense network shows competitive performance in learning complex patterns, whereas the shallow three-layer network, sometimes being less accurate, provides computational efficiency. SHAP (Shapley Additive Explanations) plots are employed to identify efficient class prediction, revealing nuanced contributions of EEG frequency bands to model accuracy. Overall, the study highlights the importance of optimizer selection, model architecture, and EEG frequency band analysis in enhancing classifier performance and understanding feature importance in neuroimaging-based classification tasks.

Problem

• Investigating classifier performance across EEG frequency bands
• Evaluating optimizer efficiency for brain hemisphere classification
• Comparing neural network architectures for EEG spatial feature extraction

Innovation

• Deep learning optimizers for EEG classification
• CNN and dense networks compared for neuroimaging
• SHAP analysis reveals frequency band contributions