Multi-scale EEG feature decoding with Swin Transformers for subject independent motor imagery BCIs

High inter-subject variability and the non-stationary nature of EEG signals pose significant challenges for subject-independent Brain-Computer Interfaces (BCIs) leading to poor model generalization. Differences in neural activity patterns, electrode placements, and external noise further degrade performance making it difficult to develop BCIs that remain reliable across users without extensive recalibration. This study presents a Compact Convolutional Swin Transformer (CCST) to address this issue by using hierarchical window based self-attention combined with convolutional feature extraction to efficiently capture both local electrode interactions and global temporal dependencies. This multi-scale feature representation enhances generalization across subjects, a critical factor for real world BCI deployment. We evaluated CCST on the BCI Competition IV (2a, 2b) and PhysioNet MI datasets using Leave-One-Subject-Out (LOSO) cross-validation achieving state-of-the-art classification accuracies of 68.27%, 76.61%, and 71.70% respectively. Our statistical analysis using the Wilcoxon signed-rank test with Bonferroni correction confirms significant performance improvements over benchmark models. Additionally, CCST achieves a reduction in parameters and a decrease in FLOPs compared to full self-attention models making it more efficient for real-time BCI applications. These results establish CCST as a scalable and efficient framework for adaptive subject-independent BCIs with promising applications in neurorehabilitation, assistive technology, and cognitive training.

• Publication year

  2026

• Venue

  Scientific Reports

• Publication date

  2026-01-20

• Fields of study

  Medicine, Computer Science, Engineering

• Identifiers
  DOI 10.1038/s41598-025-32207-3PMID 41559106PMCID 12820037
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar, PubMed

• No claims are published for this paper.

• No concepts are published for this paper.

• TMSA-Net:A novel attention mechanism for improved motor imagery EEG signal processing

  2025cited by this paper
• TCANet: a temporal convolutional attention network for motor imagery EEG decoding

  2025cited by this paper
• A Novel 3D Approach with a CNN and Swin Transformer for Decoding EEG-Based Motor Imagery Classification

  2025cited by this paper
• Multi-scale convolutional transformer network for motor imagery brain-computer interface

  2025cited by this paper
• Transformers in EEG Analysis: A Review of Architectures and Applications in Motor Imagery, Seizure, and Emotion Classification

  2025cited by this paper
• A composite improved attention convolutional network for motor imagery EEG classification

  2025cited by this paper
• Enhanced electroencephalogram signal classification: A hybrid convolutional neural network with attention-based feature selection.

  2025cited by this paper
• Motor Imagery EEG signals classification using a Transformer-GCN approach

  2025cited by this paper
• A causal perspective on brainwave modeling for brain–computer interfaces

  2024cited by this paper
• Compact convolutional transformer for subject-independent motor imagery EEG-based BCIs

  2024cited by this paper
• CTNet: a convolutional transformer network for EEG-based motor imagery classification

  2024cited by this paper
• Brain–computer interfaces: the innovative key to unlocking neurological conditions

  2024cited by this paper
• Learning High-Frequency Functions Made Easy with Sinusoidal Positional Encoding

  2024cited by this paper
• Large Brain Model for Learning Generic Representations with Tremendous EEG Data in BCI

  2024cited by this paper
• Data Constraints and Performance Optimization for Transformer-Based Models in EEG-Based Brain-Computer Interfaces: A Survey

  2024cited by this paper
• Toward Domain-Free Transformer for Generalized EEG Pre-Training

  2024cited by this paper
• Temporal-spatial transformer based motor imagery classification for BCI using independent component analysis

  2024cited by this paper
• Deep Learning in EEG-Based BCIs: A Comprehensive Review of Transformer Models, Advantages, Challenges, and Applications

  2023cited by this paper
• Exploring the Potential of Attention Mechanism-Based Deep Learning for Robust Subject-Independent Motor-Imagery Based BCIs

  2023cited by this paper
• BrainWave-Scattering Net: a lightweight network for EEG-based motor imagery recognition

  2023cited by this paper
• Decoding the Continuous Motion Imagery Trajectories of Upper Limb Skeleton Points for EEG-Based Brain–Computer Interface

  2023cited by this paper
• EEG-CDILNet: a lightweight and accurate CNN network using circular dilated convolution for motor imagery classification

  2023cited by this paper
• A shallow mirror transformer for subject-independent motor imagery BCI

  2023cited by this paper
• Global Adaptive Transformer for Cross-Subject Enhanced EEG Classification

  2023cited by this paper
• Transformer-Based Network with Optimization for Cross-Subject Motor Imagery Identification

  2023cited by this paper
• Electroencephalogram based brain-computer interface: Applications, challenges, and opportunities

  2023cited by this paper
• Integrating Dropout Regularization Technique at Different Layers to Improve the Performance of Neural Networks

  2023cited by this paper
• Hierarchical Transformer for Brain Computer Interface

  2023cited by this paper
• Physics-Informed Attention Temporal Convolutional Network for EEG-Based Motor Imagery Classification

  2023cited by this paper
• Discrepancy between inter- and intra-subject variability in EEG-based motor imagery brain-computer interface: Evidence from multiple perspectives

  2023cited by this paper
• EEG classification algorithm of motor imagery based on CNN-Transformer fusion network

  2022cited by this paper
• A Survey on Dropout Methods and Experimental Verification in Recommendation

  2022cited by this paper
• EEG temporal–spatial transformer for person identification

  2022cited by this paper
• A novel hybrid CNN-Transformer model for EEG Motor Imagery classification

  2022cited by this paper
• MI-EEG classification using Shannon complex wavelet and convolutional neural networks

  2022cited by this paper
• A review of critical challenges in MI-BCI: From conventional to deep learning methods.

  2022cited by this paper
• Coupling Convolution, Transformer and Graph Embedding for Motor Imagery Brain-Computer Interfaces

  2022cited by this paper
• A Time Series is Worth 64 Words: Long-term Forecasting with Transformers

  2022cited by this paper
• EEG Conformer: Convolutional Transformer for EEG Decoding and Visualization

  2022cited by this paper
• Editorial: Inter- and Intra-subject Variability in Brain Imaging and Decoding

  2021cited by this paper
• Gated Transformer for Decoding Human Brain EEG Signals

  2021cited by this paper
• Time-frequency analysis methods and their application in developmental EEG data

  2021cited by this paper
• H-Transformer-1D: Fast One-Dimensional Hierarchical Attention for Sequences

  2021cited by this paper
• On Position Embeddings in BERT

  2021cited by this paper
• EEG Classification with Transformer-Based Models

  2021cited by this paper
• A Practical Survey on Faster and Lighter Transformers

  2021cited by this paper
• HS-CNN: a CNN with hybrid convolution scale for EEG motor imagery classification

  2020cited by this paper
• A Brute-Force CNN Model Selection for Accurate Classification of Sensorimotor Rhythms in BCIs

  2020cited by this paper
• EEG-Inception: A Novel Deep Convolutional Neural Network for Assistive ERP-Based Brain-Computer Interfaces

  2020cited by this paper
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

  2020cited by this paper
• Improve Transformer Models with Better Relative Position Embeddings

  2020cited by this paper
• Thinker invariance: enabling deep neural networks for BCI across more people

  2020cited by this paper
• Electrocorticographic dissociation of alpha and beta rhythmic activity in the human sensorimotor system

  2019cited by this paper
• A Convolutional Recurrent Attention Model for Subject-Independent EEG Signal Analysis

  2019cited by this paper
• Design and Optimization of a BCI-Driven Telepresence Robot Through Programming by Demonstration

  2019cited by this paper
• Deep Learning for EEG motor imagery classification based on multi-layer CNNs feature fusion

  2019cited by this paper
• An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling

  2018cited by this paper
• Supplementary for: Deep learning with convolutional neural networks for EEG decoding and visualization

  2017cited by this paper
• DeepSleepNet: A Model for Automatic Sleep Stage Scoring Based on Raw Single-Channel EEG

  2017cited by this paper
• Attention is All you Need

  2017cited by this paper
• Convolutional Sequence to Sequence Learning

  2017cited by this paper
• EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces

  2016cited by this paper
• Window Size Impact in Human Activity Recognition

  2014cited by this paper
• Dropout: a simple way to prevent neural networks from overfitting

  2014cited by this paper
• Dominant frequencies of resting human brain activity as measured by the electrocorticogram

  2013cited by this paper
• Review of the BCI Competition IV

  2012cited by this paper
• Brain-computer interfaces in medicine.

  2012cited by this paper
• An Analysis of Variance Test for Normality (Complete Samples)

  1965cited by this paper
• Multiple Comparisons among Means

  1961cited by this paper

• No citing papers are available for this paper.