Decoding Human Somatosensory Sensitivity Through Resting EEG and Behavioral Analysis: A Multimodal Fusion Approach

Area of Science:

• Neuroscience
• Pain Research
• Biomedical Engineering

Background:

• Objective assessment of somatosensory sensitivity is crucial for precision medicine and pain management.
• Current methods lack quantitative, objective indicators for somatosensory sensitivity.
• Distinguishing between different somatosensory profiles is essential for tailored clinical interventions.

Purpose of the Study:

• To propose and validate a fusion approach for decoding human somatosensory sensitivity.
• To combine quantitative sensory testing (QST) and neurophysiological data for classification.
• To reveal distinct brain activation patterns associated with different somatosensory sensitivity types.

Main Methods:

• Sixty healthy participants underwent QST (cold, heat, mechanical, pressure) and resting-state electroencephalography (EEG).
• QST scores were clustered into four subgroups: generally hypersensitive (HS), generally non-sensitive (NS), predominantly thermally sensitive (TS), and predominantly mechanically sensitive (MS).
• EEG features (PSD, connectivity) were selected using a PCMRMR protocol and classified using SVM, kNN, RF, and GC models.

Main Results:

• Frequency-based connectivity (FBC) emerged as a superior EEG-derived brain signature for classifying somatosensory sensitivity types.
• A classification accuracy of 87% was achieved in distinguishing between HS, NS, TS, and MS groups.
• Fused multimodal data successfully decoded distinct brain network patterns for each somatosensory type.

Conclusions:

• A quantitative, multi-parameter approach to somatosensory sensitivity assessment is feasible with high accuracy.
• The developed fusion method offers significant potential for objective pain perception evaluation in clinical practice.
• Objective classification of somatosensory sensitivity can guide personalized pain management strategies.