Research on Signal-to-Noise Ratio Comparison and Optimization of EEG Signals in Brain-Computer Interface Systems

Abstract

Electroencephalogram (EEG) signals are widely used in brain-computer interface (BCI) systems due to their high temporal resolution and non-invasive nature. However, their low amplitude, susceptibility to noise, and poor spatial resolution often lead to inaccurate classification of brain intentions. To address these limitations, this study explores a multimodal approach that integrates functional near-infrared spectroscopy (fNIRS) with EEG to enhance signal quality and classification accuracy. The complementary advantages of EEG’s millisecond-level temporal resolution and fNIRS’s high spatial resolution enable more robust and interpretable decoding of brain activities. Furthermore, we propose the potential integration of transcranial focused ultrasound (tFUS) technology, which offers precise neuromodulation capabilities, for future multimodal BCI systems. Experimental results demonstrate that the EEG-fNIRS combination significantly improves classification performance compared to single-modal EEG, with higher accuracy and reduced noise interference. This research highlights the promising direction of multimodal signal fusion for developing more reliable and efficient non-invasive BCI systems, with broad applications in medical rehabilitation, assistive technology, and cognitive neuroscience.