PEPbench

PEPbench benchmarks algorithms that extract the pre-ejection period (PEP) from electrocardiography (ECG) and impedance cardiography (ICG) signals, providing a standardized framework and annotated datasets for reproducible evaluation in psychophysiology.

Key Features:

• Algorithm Library: PEPbench implements multiple Q-peak and B-point detection algorithms from literature that can be combined to create comprehensive PEP extraction pipelines.
• Standardized Evaluation Framework: PEPbench provides a standardized framework for evaluating PEP extraction algorithms, enabling systematic comparison across different algorithm combinations.
• Public Datasets with Reference Annotations: PEPbench includes two datasets with manually annotated Q-peaks and B-points, which are the first publicly available datasets with reference PEP annotations.
• Systematic Benchmarking: PEPbench enables comparison of 108 different algorithm combinations to assess performance differences and error rates.

Scientific Applications:

• Cardiac Sympathetic Nervous System Assessment: PEPbench supports evaluation of algorithms measuring PEP, which serves as a marker of cardiac sympathetic nervous system (SNS) activity in psychophysiological research.
• Algorithm Performance Evaluation: PEPbench enables researchers to assess the accuracy and reliability of automated PEP extraction algorithms, particularly for beat-to-beat level analysis where error rates are relatively high.
• Reproducible Psychophysiology Research: PEPbench facilitates reproducible research by providing open-source algorithms and annotated datasets for standardized comparison of PEP extraction methods.

Methodology:

PEPbench combines different Q-peak detection algorithms and B-point detection algorithms from literature to create PEP extraction pipelines, then evaluates these combinations against manually annotated Q-peaks and B-points in two datasets using a standardized comparison framework.