Extraction of stride events from gait accelerometry during treadmill walking

Area of Science:

• Biomechanics
• Neurology
• Signal Processing

Background:

• Evaluating stride events is crucial for understanding gait changes in aging and neurological diseases.
• Accurate identification of heel contact and toe-off events for gait cycle analysis is challenging.
• Current methods for extracting stride events can be cumbersome.

Purpose of the Study:

• To propose and validate a method for extracting stride cycle events from tri-axial accelerometry signals.
• To assess the accuracy of the proposed method by comparing it with motion capture data.
• To evaluate the method's ability to differentiate between subject groups.

Main Methods:

• A novel method was developed to extract stride cycle events from tri-axial accelerometry data.
• Data were collected from healthy controls, individuals with Parkinson's disease, and individuals with peripheral neuropathy walking on a treadmill.
• Gait accelerometry signals were captured using a tri-axial accelerometer at the L3 lumbar segment, with motion capture data serving as the comparison.

Main Results:

• The proposed methodology accurately extracted heel and toe contact events from both feet.
• Mean gait cycle intervals derived from accelerometry matched those from motion capture.
• Cycle-to-cycle variability measures were within 1.5%, and subject group differences were identifiable using both methods.

Conclusions:

• A simple tri-axial accelerometer and signal processing algorithm can effectively capture stride events.
• This algorithm facilitates stride event assessment during treadmill walking.
• This represents a significant step towards real-world stride event assessment using tri-axial accelerometers.