A Music Interaction System Based on the Emotional Recognition of Older Adults in Smart Homes: Cross-Sectional Experimental Study

TL;DR

This study explores how music can help older adults regulate emotions in smart homes by analyzing EEG signals and developing a new adaptive music interaction model.

Contribution

A novel Multimodal Fusion-Adaptive Wellness Interaction model was developed to match suitable music with older adults' emotional needs.

Findings

Music with positive valence and high arousal significantly soothes older adults' emotions.

EEG data showed significant activation in specific brain regions in response to music stimuli.

The model provides a theoretical basis for emotion regulation in smart home systems.

Abstract

The advent of the Smart Home 3.0 era imposes higher technical requirements for the construction of the home ecosystem. Music is an effective means for humans to regulate their emotions. The emergence of multimodal technology facilitates the use of music for emotional regulation by older adults in their home environment. Therefore, it is necessary to carry out further research on the emotional issues of the music interaction system in smart homes that are targeted at older adult users.

This study aimed to establish the mapping relationship between music parameters and emotional states based on the Valence-Arousal emotional model and through the analysis of electroencephalogram (EEG) signals. In addition, a novel Multimodal Fusion-Adaptive Wellness Interaction model was constructed to match the most suitable music and meet the emotional needs of older adult users.

A total of 68 older adult participants aged between 60 and 75 years were recruited. Four groups of music stimulus materials were formulated to conduct an experimental study on the relationship between the emotions of older adult participants and music parameters. Finally, the Multimodal Fusion-Adaptive Wellness Interaction model was constructed. With the recreational sofa as the main body, multimodal, emotional intelligent adaptive interaction was realized.

During the experiment, approximately 68,970,528 EEG signal data were collected. After analyzing the EEG bands and extracting the event-related potential data, the following results were obtained: For P200, the music stimulus materials showed significant differences and activation in the left parietal and occipital lobe regions (P7, P3, O1, O2; P<.001). For N400, there were significant differences and activation in the prefrontal lobe region due to the music stimulus materials (O1, CZ, O2; P<.001). For P300, the music stimulus materials led to significant differences and activation in the occipital lobe region and the central line region of the temporal area (F7, FPZ, FZ, F8; P<.001).

Older adult users respond in a way that the A1 stimulus material, namely the music material with positive valence and high arousal (+V + A), can soothe the emotional state of older adult users. This provides a theoretical basis and evidence for the music interaction regulation module in smart home systems.

Figures