Machine Learning to Quantify Physical Activity in Children with Cerebral Palsy: Comparison of Group, Group-Personalized, and Fully-Personalized Activity Classification Models
Pattern recognition methodologies, such as those utilizing machine learning (ML) approaches, have the potential to improve the accuracy and versatility of accelerometer-based assessments of physical activity (PA). Children with cerebral palsy (CP) exhibit significant heterogeneity in relation to imp...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-07-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/20/14/3976 |
| _version_ | 1827712891617280000 |
|---|---|
| author | Matthew N. Ahmadi Margaret E. O’Neil Emmah Baque Roslyn N. Boyd Stewart G. Trost |
| author_facet | Matthew N. Ahmadi Margaret E. O’Neil Emmah Baque Roslyn N. Boyd Stewart G. Trost |
| author_sort | Matthew N. Ahmadi |
| collection | DOAJ |
| description | Pattern recognition methodologies, such as those utilizing machine learning (ML) approaches, have the potential to improve the accuracy and versatility of accelerometer-based assessments of physical activity (PA). Children with cerebral palsy (CP) exhibit significant heterogeneity in relation to impairment and activity limitations; however, studies conducted to date have implemented “one-size fits all” group (G) models. Group-personalized (GP) models specific to the Gross Motor Function Classification (GMFCS) level and fully-personalized (FP) models trained on individual data may provide more accurate assessments of PA; however, these approaches have not been investigated in children with CP. In this study, 38 children classified at GMFCS I to III completed laboratory trials and a simulated free-living protocol while wearing an ActiGraph GT3X+ on the wrist, hip, and ankle. Activities were classified as sedentary, standing utilitarian movements, or walking. In the cross-validation, FP random forest classifiers (99.0–99.3%) exhibited a significantly higher accuracy than G (80.9–94.7%) and GP classifiers (78.7–94.1%), with the largest differential observed in children at GMFCS III. When evaluated under free-living conditions, all model types exhibited significant declines in accuracy, with FP models outperforming G and GP models in GMFCS levels I and II, but not III. Future studies should evaluate the comparative accuracy of personalized models trained on free-living accelerometer data. |
| first_indexed | 2024-03-10T18:24:41Z |
| format | Article |
| id | doaj.art-5f5cb3f3c82e457b80a619af59c3adf6 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T18:24:41Z |
| publishDate | 2020-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-5f5cb3f3c82e457b80a619af59c3adf62023-11-20T07:05:15ZengMDPI AGSensors1424-82202020-07-012014397610.3390/s20143976Machine Learning to Quantify Physical Activity in Children with Cerebral Palsy: Comparison of Group, Group-Personalized, and Fully-Personalized Activity Classification ModelsMatthew N. Ahmadi0Margaret E. O’Neil1Emmah Baque2Roslyn N. Boyd3Stewart G. Trost4Institute of Health and Biomedical Innovation at Queensland Centre for Children’s Health Research, Queensland University of Technology, South Brisbane 4101, AustraliaDepartment of Rehabilitation and Regenerative Medicine, Columbia University Irving Medical Center, New York, NY 10032, USAInstitute of Health and Biomedical Innovation at Queensland Centre for Children’s Health Research, Queensland University of Technology, South Brisbane 4101, AustraliaQueensland Cerebral Palsy and Rehabilitation Research Centre, UQ Child Health Research Centre, Faculty of Medicine, The University of Queensland, South Brisbane 4101, AustraliaInstitute of Health and Biomedical Innovation at Queensland Centre for Children’s Health Research, Queensland University of Technology, South Brisbane 4101, AustraliaPattern recognition methodologies, such as those utilizing machine learning (ML) approaches, have the potential to improve the accuracy and versatility of accelerometer-based assessments of physical activity (PA). Children with cerebral palsy (CP) exhibit significant heterogeneity in relation to impairment and activity limitations; however, studies conducted to date have implemented “one-size fits all” group (G) models. Group-personalized (GP) models specific to the Gross Motor Function Classification (GMFCS) level and fully-personalized (FP) models trained on individual data may provide more accurate assessments of PA; however, these approaches have not been investigated in children with CP. In this study, 38 children classified at GMFCS I to III completed laboratory trials and a simulated free-living protocol while wearing an ActiGraph GT3X+ on the wrist, hip, and ankle. Activities were classified as sedentary, standing utilitarian movements, or walking. In the cross-validation, FP random forest classifiers (99.0–99.3%) exhibited a significantly higher accuracy than G (80.9–94.7%) and GP classifiers (78.7–94.1%), with the largest differential observed in children at GMFCS III. When evaluated under free-living conditions, all model types exhibited significant declines in accuracy, with FP models outperforming G and GP models in GMFCS levels I and II, but not III. Future studies should evaluate the comparative accuracy of personalized models trained on free-living accelerometer data.https://www.mdpi.com/1424-8220/20/14/3976accelerometerswearable sensorsexercisemeasurementGMFCS level |
| spellingShingle | Matthew N. Ahmadi Margaret E. O’Neil Emmah Baque Roslyn N. Boyd Stewart G. Trost Machine Learning to Quantify Physical Activity in Children with Cerebral Palsy: Comparison of Group, Group-Personalized, and Fully-Personalized Activity Classification Models Sensors accelerometers wearable sensors exercise measurement GMFCS level |
| title | Machine Learning to Quantify Physical Activity in Children with Cerebral Palsy: Comparison of Group, Group-Personalized, and Fully-Personalized Activity Classification Models |
| title_full | Machine Learning to Quantify Physical Activity in Children with Cerebral Palsy: Comparison of Group, Group-Personalized, and Fully-Personalized Activity Classification Models |
| title_fullStr | Machine Learning to Quantify Physical Activity in Children with Cerebral Palsy: Comparison of Group, Group-Personalized, and Fully-Personalized Activity Classification Models |
| title_full_unstemmed | Machine Learning to Quantify Physical Activity in Children with Cerebral Palsy: Comparison of Group, Group-Personalized, and Fully-Personalized Activity Classification Models |
| title_short | Machine Learning to Quantify Physical Activity in Children with Cerebral Palsy: Comparison of Group, Group-Personalized, and Fully-Personalized Activity Classification Models |
| title_sort | machine learning to quantify physical activity in children with cerebral palsy comparison of group group personalized and fully personalized activity classification models |
| topic | accelerometers wearable sensors exercise measurement GMFCS level |
| url | https://www.mdpi.com/1424-8220/20/14/3976 |
| work_keys_str_mv | AT matthewnahmadi machinelearningtoquantifyphysicalactivityinchildrenwithcerebralpalsycomparisonofgroupgrouppersonalizedandfullypersonalizedactivityclassificationmodels AT margareteoneil machinelearningtoquantifyphysicalactivityinchildrenwithcerebralpalsycomparisonofgroupgrouppersonalizedandfullypersonalizedactivityclassificationmodels AT emmahbaque machinelearningtoquantifyphysicalactivityinchildrenwithcerebralpalsycomparisonofgroupgrouppersonalizedandfullypersonalizedactivityclassificationmodels AT roslynnboyd machinelearningtoquantifyphysicalactivityinchildrenwithcerebralpalsycomparisonofgroupgrouppersonalizedandfullypersonalizedactivityclassificationmodels AT stewartgtrost machinelearningtoquantifyphysicalactivityinchildrenwithcerebralpalsycomparisonofgroupgrouppersonalizedandfullypersonalizedactivityclassificationmodels |