CardioPy

CardioPy performs heart rate variability (HRV) analysis on single-lead EKG signals, detecting R-peaks and computing time- and frequency-domain HRV metrics for short-term (5-minute) recordings.

Key Features:

• Implementation: Python-based software optimized for short-term (5-minute) single-lead EKG recordings.
• R-peak detection: An adaptive thresholding algorithm for R-peak detection in single-lead EKG signals.
• Artifact handling and parameter adjustment: Manual artifact removal and manual parameter adjustment for data quality control and refinement of detections.
• Analyses: Standard time-domain (e.g., SDNN) and frequency-domain (e.g., low frequency power, low frequency-to-high frequency ratio) HRV analyses including power spectral analysis.
• Validation against Kubios: Power spectra compared to Kubios HRV, with CardioPy spectra explaining an average of 99.7% of the variance in Kubios spectra.
• Performance metrics: Validation on EKG segments from five healthy individuals reporting mean sensitivity 100.0% (SD 0.05%) and positive predictive value 99.8% (SD 0.20%).
• Demonstrated results: Application to brain-state comparisons showed reductions in SDNN, low frequency power, and LF/HF ratio during slow wave sleep versus wakefulness.

Scientific Applications:

• Brain-state HRV comparison: Comparison of HRV metrics between wakefulness and slow wave sleep.
• Clinical research on HRV biomarkers: Use in studies where HRV is applied as a marker of health and disease.
• Algorithm benchmarking: Quantitative comparison and benchmarking of HRV outputs and power spectra against Kubios HRV.

Methodology:

R-peak detection using an adaptive thresholding algorithm, manual artifact removal and parameter adjustment, computation of time-domain metrics (e.g., SDNN) and frequency-domain metrics (low frequency power, LF/HF ratio), power spectral analysis and comparison to Kubios, and validation reported via sensitivity and positive predictive value on five healthy individuals' EKG segments optimized for 5-minute recordings.