Asynchronous Prediction of Human Gait Intention in a Pseudo Online Paradigm Using Wavelet Transform

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Human voluntary gait intention is crucial for developing brain-controlled assistive locomotion technologies.
• Gait intention is indicated by slow DC potentials and brainwave power shifts detectable 1.5-2 seconds prior to movement onset.
• Reliably predicting gait intention is key for effective neurorehabilitation.

Purpose of the Study:

• To determine if voluntary gait 'starting' and 'stopping' intentions can be reliably detected before movement.
• To develop and validate a computational algorithm for asynchronous gait intention prediction.
• To assess the accuracy, sensitivity, and specificity of the prediction algorithm.

Main Methods:

• A computational algorithm using support vector machines was designed for gait intention prediction.
• Advanced wavelet transform algorithms were employed for signal processing.
• Offline and pseudo-online testing environments were utilized with six healthy subjects performing self-paced gait cycles.

Main Results:

• Offline testing achieved high accuracy (88.23% for start, 87.04% for stop) and specificity (90.24% for start, 89.59% for stop).
• Pseudo-online testing yielded a True Positive Rate of 85.5% for 'start' and 81.2% for 'stop' with low False Positives per Minute.
• Average detection latencies were approximately -1000 ms for both start and stop intentions.

Conclusions:

• Human voluntary gait intention, for both starting and stopping, can be reliably predicted before movement onset.
• The developed algorithm shows promising results in True Positive Rate, False Positives per Minute, and detection latency.
• This prediction capability is significant for advancing real-life assistive technologies in locomotion rehabilitation.