A Novel Framework for Quantifying Accuracy and Precision of Event Detection Algorithms in FES-Cycling

Sensors, 2021 · DOI: 10.3390/s21134571 · Published: July 3, 2021

Simple Explanation

Functional electrical stimulation (FES) is a rehabilitation technique that uses electrical pulses to activate muscles in paralyzed limbs, helping to restore movement or function. In FES-cycling, this stimulation is often based on the angle of the pedals. However, changes in seating position or other factors can disrupt the timing of these stimulations. Adaptive algorithms using real-time interpretation of body positions could solve this, but their accuracy needs to be tested. This study evaluates three algorithms (Hilbert, BSgonio, and Gait Cycle Index (GCI) Observer) using data from passive cycling of participants with spinal cord injury to see how accurately they can detect key events in the cycling motion.

Study Duration

Not specified

Participants

Six participants with complete SCI

Evidence Level

Not specified

Key Findings

1
The Hilbert and BSgonio algorithms met the set accuracy criteria, showing good agreement with the baseline reference. BSgonio proved an accurate and precise way of detecting events reliably in real time.
2
The GCI Observer algorithm did not meet the accuracy criteria and may not be sufficient for practical use in FES-cycling. The GCI Observer, on the other hand, failed our evaluation criteria.
3
The study introduces a method for evaluating the accuracy and precision of real-time control algorithms in FES-cycling, which can be adapted for other applications needing algorithmic accuracy.

Research Summary

This study introduces a framework for evaluating the accuracy and precision of event detection algorithms used in FES-cycling. Three algorithms (Hilbert, BSgonio, and GCI Observer) were tested using inertial data from participants with spinal cord injury. The algorithms were evaluated based on their ability to detect specific events within a pedaling cycle, compared to a baseline reference. Limits of agreement and Lin’s concordance correlation coefficient were used to assess accuracy and precision. The BSgonio and Hilbert algorithms demonstrated sufficient accuracy and precision for real-time event detection, while the GCI Observer algorithm did not meet the established criteria. The BSgonio algorithm proved an accurate and precise way of detecting events reliably in real time.

Practical Implications

Improved FES-Cycling Control

The BSgonio algorithm can be used for more accurate and reliable control of FES-cycling, potentially leading to more effective rehabilitation outcomes.

Framework for Algorithm Evaluation

The presented framework provides a standardized method for evaluating and comparing different event detection algorithms in FES-cycling and other applications.

Real-time Adaptability

Adaptive and reliable algorithms facilitate FES systems in practical applications.

Study Limitations

1
Small number of SCI subjects
2
Inertial measurement data acquired from passive leg movements
3
Results to be interpreted with care due to experimental conditions differing from FES-induced cycling

Your Feedback

Was this summary helpful?

Back to Biomedical