Physiological Monitoring of Sound-Based Relaxation Using Binaural Audio and Vibroacoustic Stimulation

Area of Science:

• Neuroscience and Biomedical Engineering
• Human-Computer Interaction
• Psychophysiology

Background:

• Non-invasive methods like immersive audio and vibroacoustic stimulation are explored for emotional and physiological state modulation.
• The SonikB3D platform integrates 3D audio, vibroacoustic stimulation, and physiological monitoring.
• Advancements include a refined protocol and a data-driven framework for automatic relaxation assessment.

Purpose of the Study:

• To evaluate the effectiveness of combined 3D audio and vibroacoustic stimulation on subjective and physiological relaxation.
• To develop and validate a machine learning model for automatic relaxation state classification using multimodal biosignals.

Main Methods:

• A pilot study with 20 participants using the SonikB3D platform with synchronized electroencephalography (EEG), photoplethysmography (PPG), and electrodermal activity (EDA).
• Stimulation involved binaural 3D audio (beats plus music) and vibroacoustic input via a specialized massage table.
• Multimodal physiological features were extracted for a two-stage machine learning pipeline (Random Forest classifier with session-level aggregation).

Main Results:

• Participants reported a statistically significant increase in relaxation, with an average 8% improvement in normalized scores.
• The machine learning model achieved 80% accuracy and an 0.857 F1-score in classifying relaxation states.
• Preliminary evidence suggests measurable changes in relaxation indicators and feasibility of automatic state inference.

Conclusions:

• Combined 3D audio and vibroacoustic stimulation show potential for inducing relaxation.
• Multimodal biosignal analysis and machine learning offer a viable approach for objective relaxation assessment.
• This data-driven methodology advances the study of human responses to multisensory stimuli.