Assessing Arousal Through Multimodal Biosignals: A Preliminary Approach.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Psychology

Background:

• Rising Autism Spectrum Disorder (ASD) prevalence necessitates accessible rehabilitation tools.
• Current computer-based interventions for ASD often lack emotional state awareness, limiting effectiveness.
• Integrating physiological data with neuroimaging offers potential for personalized interventions.

Purpose of the Study:

• To develop models for assessing user emotional states using physiological signals.
• To investigate the feasibility of classifying arousal states from biosignals within a functional Magnetic Resonance Imaging (fMRI) environment.
• To lay the groundwork for future multimodal emotion recognition systems.

Main Methods:

• Acquired physiological signals concurrently with fMRI scans.
• Extracted 35 features from biosignal segments.
• Utilized extracted features for automatic classification of arousal states (High vs. Low Arousal).

Main Results:

• Some extracted physiological features showed statistically significant differences between arousal states.
• Automatic classification of arousal states using physiological signals alone yielded suboptimal results.
• Recording physiological signals within the MRI environment presented significant challenges.

Conclusions:

• Classifying arousal states solely from physiological signals in an fMRI setting is challenging.
• Further research is needed to identify more discriminative features from biosignals.
• Developing robust emotion-aware rehabilitation for ASD requires overcoming technical and analytical hurdles.