Identification of mental stress granularity from long-term ECG recordings using novel complexity analysis

Area of Science:

• Cardiology
• Psychiatry
• Complex Systems Analysis

Background:

• Mental health conditions alter physiological signals, but links between long-term physiological data like ECG and psychopathology are underexplored.
• Complex time series exhibit inverse power law waiting time distributions for crucial events, characterized by temporal complexity index μ.
• Diffusion time series from these events have a scaling index δ related to μ.

Purpose of the Study:

• To investigate the relationship between long-term electrocardiogram (ECG) time series and mental stress levels.
• To quantify the complexity of ECG signals during varying stress levels using multiscale modified diffusion entropy analysis (MSMDEA).
• To explore the potential of ECG analysis as an adjunct tool in precision psychiatry.

Main Methods:

• Collected ECG data from 90 participants over four weeks using a chest strap sensor (Polar H10).
• Participants self-reported daily stress levels (-3 to +3).
• Preprocessed 5-minute ECG segments and applied MSMDEA (time scale factors 1-20) to calculate complexity indices μ and δ.

Main Results:

• Significant differences in scaling index δ were observed between high and low-stress groups (p=0.05).
• Analysis of fine-grained stress scores revealed significant differences in δ for over half of the group pairs (p <0.05).
• Lower stress correlated with ECG signals approaching critical complexity (δ=1, μ=2), indicating healthier physiological and psychological states.

Conclusions:

• ECG signal complexity, specifically the scaling index δ, can differentiate stress levels.
• A critical complexity (δ=1 or μ=2) in ECGs is associated with better mental health.
• ECG crucial event analysis shows promise as a biomarker for mental health assessment in precision psychiatry.