Electrodermal activity processing: a convex optimization approach.

Area of Science:

• Psychophysiology
• Computational Neuroscience
• Signal Processing

Background:

• Electrodermal activity (EDA) and skin conductance (SC) are key measures of autonomic nervous system responses.
• Previous models for SC analysis have limitations in separating tonic and phasic components.
• Understanding SC components is crucial for interpreting affective and cognitive states.

Purpose of the Study:

• To introduce a novel convex optimization model for decomposing SC into tonic and phasic components.
• To validate the model's ability to differentiate physiological responses to affective stimuli.
• To assess the distinct contributions of tonic and phasic SC to arousal detection.

Main Methods:

• Developed a convex optimization model to decompose SC signals.
• Incorporated physiological knowledge via constraints and regularizers.
• Conducted an experiment with 9 healthy subjects viewing negative-valence high-arousal and neutral images from the IAPS database.
• Extracted features from model-estimated tonic and phasic signals for statistical analysis.

Main Results:

• The model's extracted phasic driver significantly differentiated between arousal and neutral image conditions.
• No significant differences were found in the estimated tonic components between conditions.
• Results align with literature, confirming the phasic component's link to sympathetic activity.

Conclusions:

• The novel convex optimization model effectively isolates the phasic component of SC.
• The phasic component is a sensitive indicator of sympathetic responses to affective stimuli.
• Preliminary findings support the model's potential for advanced psychophysiological research.