Scholarlink

<p>Emotion recognition in quadriplegic (QP) patients using Electroencephalogram (EEG) signals is highly challenging due to continuous medication, which causes overlapping of EEG bands in Pharmaco-EEG (pEEG) signals and introduces severe artefacts such as involuntary spasms and respiratory disturbances. Conventional emotion recognition techniques designed for drug-free EEG are ineffective under these conditions. To address these challenges, this paper proposes a novel Rehabilitation BCI with Music Therapy (RBMT) framework for reliable emotion detection and therapeutic intervention. A SiO₂ nano-coated graphene EEG electrode is developed to reduce band overlap and enhance signal quality. The acquired pEEG signals are denoised using Discrete and Stationary Wavelet Transforms. Spatial–temporal features are extracted using an SBOA-optimized LSTM integrated with a mutual cross-attention mechanism and SoftMax classifier to identify anxiety and depression based on valence–arousal analysis. The predicted emotional state triggers adaptive music therapy via a BCI audio interface. Experimental results demonstrate that the proposed RBMT framework achieves 98% emotion recognition accuracy, outperforming existing approaches. The proposed RBMT method achieved 98% emotion recognition accuracy, 20 dB SNR, 0.15 training loss, 2.8&#xa0;s response time, 25% attention weight, and 9.3 comfort score.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Pharmaco-EEG–Driven Rehabilitation BCI Framework Using Nano-Coated Graphene Electrodes and SBOA-LSTM for Emotion-Adaptive Music Therapy in Quadriplegic Patients

  • 期刊论文

  • 发表日期：

    2026年4月4日

查看全文
  • B. Ramesh,
  • Ushus. S. Kumar

摘要

Emotion recognition in quadriplegic (QP) patients using Electroencephalogram (EEG) signals is highly challenging due to continuous medication, which causes overlapping of EEG bands in Pharmaco-EEG (pEEG) signals and introduces severe artefacts such as involuntary spasms and respiratory disturbances. Conventional emotion recognition techniques designed for drug-free EEG are ineffective under these conditions. To address these challenges, this paper proposes a novel Rehabilitation BCI with Music Therapy (RBMT) framework for reliable emotion detection and therapeutic intervention. A SiO₂ nano-coated graphene EEG electrode is developed to reduce band overlap and enhance signal quality. The acquired pEEG signals are denoised using Discrete and Stationary Wavelet Transforms. Spatial–temporal features are extracted using an SBOA-optimized LSTM integrated with a mutual cross-attention mechanism and SoftMax classifier to identify anxiety and depression based on valence–arousal analysis. The predicted emotional state triggers adaptive music therapy via a BCI audio interface. Experimental results demonstrate that the proposed RBMT framework achieves 98% emotion recognition accuracy, outperforming existing approaches. The proposed RBMT method achieved 98% emotion recognition accuracy, 20 dB SNR, 0.15 training loss, 2.8 s response time, 25% attention weight, and 9.3 comfort score.