The influence of different flavor peptides on brain perception via scalp electroencephalogram and development of a taste model

Traditional taste evaluation methods often rely on subjective assessments, introducing biases. To address this, we propose using electroencephalography (EEG) to explore the link between brain activity and taste perception. Our EEG analysis showed significant activity differences in specific brain regions, particularly at electrodes Pz, FT7, F7, and TP7, highlighting their role in taste signal processing. Consistent activity at Pz across various tastes supports the development of a mathematical model and sensory evaluation system. We used wavelet packet transform for EEG signal preprocessing, followed by feature extraction and classification with the Common Spatial Pattern (CSP) and Support Vector Machine (SVM) algorithms. Testing five taste categories-sour, sweet, bitter, salty, and umami-resulted in an overall prediction accuracy of 0.7613, with the highest accuracy of 0.8235 for "sweet" taste. Despite challenges in predicting "sour" and "salty" tastes, our study demonstrates the potential of combining wavelet packet transform, CSP, and SVM for EEG-based taste classification.