Quantitatively Recognizing Stimuli Intensity of Primary Taste Based on Surface Electromyography

Area of Science:

• Neuroscience
• Biomedical Engineering
• Sensory Science

Background:

• Quantitative recognition of taste intensity is crucial for understanding gustatory perception.
• Surface electromyography (sEMG) offers a potential non-invasive method for measuring facial muscle activity related to taste.
• Previous methods for taste intensity recognition have limitations in quantitative accuracy.

Purpose of the Study:

• To develop and validate a novel approach for quantitatively recognizing the intensity of primary taste stimuli using sEMG.
• To assess the feasibility of distinguishing and quantifying different taste intensities (Sour, Sweet, Bitter, Salty, Umami).
• To optimize feature extraction and dimensionality for improved sEMG-based taste recognition models.

Main Methods:

• Acquisition of sEMG signals from facial muscles under various primary taste stimuli (citric acid, sucrose, magnesium chloride, sodium chloride, sodium glutamate) at different concentrations.
• Signal preprocessing including Quadratic Variation Reduction and adaptive notch filtering to remove noise and artifacts.
• Extraction of 330 features per sample, followed by Support Vector Machine (SVM) regression with five-fold cross-validation for taste intensity recognition.
• Feature dimensionality reduction and optimization for each taste type.

Main Results:

• The SVM model achieved R² scores of 0.7277 (Sour), 0.1963 (Sweet), 0.7450 (Bitter), 0.7642 (Salty), and 0.5055 (Umami), demonstrating successful quantitative recognition for Sour, Bitter, and Salty tastes.
• Facial muscle activity patterns were analyzed and compared across different taste types and intensities.
• Feature dimensionality was significantly reduced from 330 to between 120 and 260 features per taste type without compromising performance.
• Model performance was analyzed across multiple subjects, indicating the robustness of the approach.

Conclusions:

• Quantitative recognition of primary taste intensity using sEMG and SVM is feasible, particularly for Sour, Bitter, and Salty tastes.
• The developed method provides a foundation for objective and quantitative assessment of taste perception.
• Further research and applications in taste stimuli recognition can benefit from this sEMG-based approach, with potential for simplified models and improved accuracy.