Heart rate-based window segmentation improves accuracy of classifying posttraumatic stress disorder using heart rate...

Area of Science:

• Physiological monitoring
• Autonomic nervous system function
• Medical diagnostics

Background:

• Heart rate variability (HRV) reflects autonomic nervous system function.
• HRV and heart rate (HR) are known to differ in individuals with posttraumatic stress disorder (PTSD).
• Objective physiological measures are needed for PTSD assessment.

Purpose of the Study:

• To develop a classifier for PTSD using HR and HRV measures.
• To improve classifier performance through a novel HR-based window segmentation technique.
• To assess the potential for objective PTSD illness severity tracking.

Main Methods:

• Collected 24-hour single-channel ECG data from 23 PTSD subjects and 25 controls.
• Derived RR intervals to calculate HR and HRV metrics.
• Developed a logistic regression classifier using quiescent HR segments for feature extraction.

Main Results:

• The classifier achieved a median area under the receiver operating curve (AUC) of 0.86 using quiescent segments.
• This performance significantly outperformed classifiers using 24-hour data (AUC 0.72) or random segments (AUC 0.67).
• The four most predictive features from quiescent segments were identified.

Conclusions:

• The developed segmentation approach enhances PTSD classification accuracy from HR and HRV.
• This method holds potential for objective monitoring of PTSD illness severity.
• Future research should prospectively validate classifier changes with PTSD progression or treatment.