A Silent Speech Decoding System from EEG and EMG with Heterogenous Electrode Configurations

TL;DR

This paper presents neural networks capable of decoding silent speech from EEG and EMG data with heterogeneous electrode configurations, achieving high accuracy and demonstrating potential for aiding speech-impaired individuals.

Contribution

Introduces neural network models that handle heterogeneous EEG/EMG electrode setups and demonstrates their effectiveness in silent speech decoding across subjects and languages.

Findings

95.3% word classification accuracy in healthy participants

54.5% accuracy in speech-impaired patient

Significant improvement over single-subject models

Abstract

Silent speech decoding, which performs unvocalized human speech recognition from electroencephalography/electromyography (EEG/EMG), increases accessibility for speech-impaired humans. However, data collection is difficult and performed using varying experimental setups, making it nontrivial to collect a large, homogeneous dataset. In this study we introduce neural networks that can handle EEG/EMG with heterogeneous electrode placements and show strong performance in silent speech decoding via multi-task training on large-scale EEG/EMG datasets. We achieve improved word classification accuracy in both healthy participants (95.3%), and a speech-impaired patient (54.5%), substantially outperforming models trained on single-subject data (70.1% and 13.2%). Moreover, our models also show gains in cross-language calibration performance. This increase in accuracy suggests the feasibility of developing practical silent speech decoding systems, particularly for speech-impaired patients.