Multi-Domain EEG Representation Learning with Orthogonal Mapping and Attention-based Fusion for Cognitive Load Classification

2025-11-15

AI Summary

This study addresses the low classification accuracy and poor robustness in cognitive load state recognition from EEG signals. To this end, we propose a time-frequency dual-domain fusion framework for multimodal representation learning. Specifically, a convolutional encoder is designed to extract discriminative temporal dynamics; multi-spectral topographic maps—constructed from power spectral density—are employed to model spatial-spectral distributions in the frequency domain; and a multi-domain attention mechanism, jointly optimized with an orthogonal projection constraint, is introduced to explicitly capture cross-domain correlations while enhancing inter-class separability and intra-class compactness. Extensive experiments on the CL-Drive and CLARE benchmark datasets demonstrate that our method significantly outperforms single-domain baselines. Moreover, it maintains stable performance under various noise perturbations, validating its superior discriminative capability and strong noise robustness.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

We propose a new representation learning solution for the classification of cognitive load based on Electroencephalogram (EEG). Our method integrates both time and frequency domains by first passing the raw EEG signals through the convolutional encoder to obtain the time domain representations. Next, we measure the Power Spectral Density (PSD) for all five EEG frequency bands and generate the channel power values as 2D images referred to as multi-spectral topography maps. These multi-spectral topography maps are then fed to a separate encoder to obtain the representations in frequency domain. Our solution employs a multi-domain attention module that maps these domain-specific embeddings onto a shared embedding space to emphasize more on important inter-domain relationships to enhance the representations for cognitive load classification. Additionally, we incorporate an orthogonal projection constraint during the training of our method to effectively increase the inter-class distances while improving intra-class clustering. This enhancement allows efficient discrimination between different cognitive states and aids in better grouping of similar states within the feature space. We validate the effectiveness of our model through extensive experiments on two public EEG datasets, CL-Drive and CLARE for cognitive load classification. Our results demonstrate the superiority of our multi-domain approach over the traditional single-domain techniques. Moreover, we conduct ablation and sensitivity analyses to assess the impact of various components of our method. Finally, robustness experiments on different amounts of added noise demonstrate the stability of our method compared to other state-of-the-art solutions.

Problem

• Classifying cognitive load from EEG signals using multi-domain representation learning
• Integrating time and frequency domain features through orthogonal mapping and attention fusion
• Enhancing discrimination between cognitive states while improving intra-class clustering

Innovation

• Convolutional encoder extracts time domain EEG features
• Multi-spectral topography maps encode frequency domain information
• Orthogonal projection constraint enhances inter-class feature discrimination