Back to search

University of North Texas

Deep Learning Models for Gait Analysis and Treatment Planning for Children With Cerebral Palsy

Abstract

dc:description

Cerebral palsy is the most common disorder that affects the mobility of children across the world. Children with cerebral palsy display a wide range of symptoms varying from mild toe walking to limited mobility. Multiple procedures performed during one surgical session are also a common form of treatment to reduce the number of surgeries needed and to facilitate a faster recovery. Treatment planning for children with cerebral palsy is a complex and multi-disciplined task with many possible options. To assist clinicians with finding the best treatment plan for a child with cerebral palsy, deep learning machine learning models were developed to quantify gait quality to recommend procedures that can best improve gait quality. 1) A deep learning autoencoder was used to compress the person's instrumented gait data into a single value, then derived a gait quality using a typically developing population. The autoencoder was able to capture approximately 90\% of the explained variance of the original data. Using a typically developing population, we are able to calculate a distance from that population as the gait quality. 2) A propensity model was trained to predict the likelihood of procedures given the gait data of an individual. The propensity model achieved a label ranking average precision of 76\%, meaning that the procedures that were performed were more likely to be ranked higher. The results can serve as guidance for clinicians when considering procedures. Surgery outcome prediction models were also trained to predict changes post-surgery. While the surgery outcome prediction models' performances were better than the standard deviation of the dataset, they can still provide reference when comparing the outcomes for different procedures. This dissertation serves to advance the clinical care and treatment planning for children with cerebral palsy.

Degree

thesis:*
Grantor dc:publisher
University of North Texas
Year dc:date
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Wang, Shou-Jen Steve 1982-
Contributors dc:contributor
  • Harris, Gerald
  • Albert, Mark V.
  • Xiao, Ting
  • Bozdag, Serdar

Subjects

dc:subject × 3

Rights

dc:rights
Statement dc:rights
  • Public
  • Wang, Shou-Jen Steve 1982-
  • Copyright
  • Copyright is held by the author, unless otherwise noted. All rights Reserved.
Language dc:language
English

Identifiers

dc:identifier.*
Identifier
local-cont-no: submission_4944
ark: ark:/67531/metadc2652452

Chain of custody

source
Harvested from
University of North Texas
Base URL
digital.library.unt.edu/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Wang, Shou-Jen Steve 1982-. Deep Learning Models for Gait Analysis and Treatment Planning for Children With Cerebral Palsy. University of North Texas, 2026. https://digital.library.unt.edu/ark:/67531/metadc2652452/