Project Introduction

Nonoperative treatment is widely accepted as the standard care for children with clubfoot, and various technologies, including accelerometers, have allowed researchers to assess functional outcomes and movement mechanics in-depth. However, it is still unclear whether nonoperatively treated children with clubfoot exhibit sided differences in dynamic agility tasks and whether ankle strength affects the quality of their movement in such tasks at skeletal maturity.

Software Overview

A raw .csv data stream of x,y, and z accelerometry vales was compiled into a resultant vector magnitude using the sum of squares. The subsequent data values were spliced into smaller data segments-representing one trial-with timestamps from a LabView created .txt file. Within each trial, a patient moves in 6 main phases: left 3 times and right 3 times. Signal analysis quantities are calculated for each of the three segments. These 6 segments are automatically identified using a proprietary automatic approach. A low-pass Butterworth Filter is applied to the raw data to identify the envelope of the total trial. There are two main frequencies in the accelerometry profile as seen below. The envelope and identification of minimal accelerometry values allowed for splitting each of the 6 phases. Three main quantities are identified: Peak G Force, Stride frequency, and a Vector Magnitude. Peak acceleration is identified using a max command, vector magnitude is identified using a trapezoidal integral function, and the stride frequency is calculated by finding the greatest amplitude frequency after applying the Fast Fourier Transform.

IMU Analysis

Developing software to analyze IMU data to describe agility of patients with clubfoot

  • Category: Biomechanics Research
  • Client: Scottish Rite for Children
  • Project date: 08/2020 - 05/2022

Contributions & Outcomes

I developed a MATLAB analysis software from provenance to validation that read in IMU accelerometry data from Edgren, Single Leg Stance, and L-Test agility tasks performed by patients with clubfoot. The software involved signal processing techniques including: filtering, step detection, envelop creation, and Fourier Transforms. The resulting metrics included step frequency, peak G (acceleration), vector magnitude, split times, and lag times. The software empowered research to identify children with unilateral clubfoot who remained nonoperative until skeletal maturity showed minimal sided differences in time/accelerometer measures during a dynamic agility test. However, ankle strength in the unaffected foot was correlated with better performance and movement quality, highlighting the importance of integrating accelerometry measures for assessing functional deficits and their relationship to ankle strength in clubfoot treatment.

Technical Skills

  • MATLAB
  • Time Series Signal Processing
  • Research
  • IMU
  • Biomechanics