An Explainable Spatial-Temporal Graphical Convolutional Network to Score Freezing of Gait in Parkinsonian Patients
Freezing of gait (FOG) is a poorly understood heterogeneous gait disorder seen in patients with parkinsonism which contributes to significant morbidity and social isolation. FOG is currently measured with scales that are typically performed by movement disorders specialists (i.e., MDS-UPDRS), or thr...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-02-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/23/4/1766 |
_version_ | 1797618403347791872 |
---|---|
author | Hyeokhyen Kwon Gari D. Clifford Imari Genias Doug Bernhard Christine D. Esper Stewart A. Factor J. Lucas McKay |
author_facet | Hyeokhyen Kwon Gari D. Clifford Imari Genias Doug Bernhard Christine D. Esper Stewart A. Factor J. Lucas McKay |
author_sort | Hyeokhyen Kwon |
collection | DOAJ |
description | Freezing of gait (FOG) is a poorly understood heterogeneous gait disorder seen in patients with parkinsonism which contributes to significant morbidity and social isolation. FOG is currently measured with scales that are typically performed by movement disorders specialists (i.e., MDS-UPDRS), or through patient completed questionnaires (N-FOG-Q) both of which are inadequate in addressing the heterogeneous nature of the disorder and are unsuitable for use in clinical trials The purpose of this study was to devise a method to measure FOG objectively, hence improving our ability to identify it and accurately evaluate new therapies. A major innovation of our study is that it is the first study of its kind that uses the largest sample size (>30 h, N = 57) in order to apply explainable, multi-task deep learning models for quantifying FOG over the course of the medication cycle and at varying levels of parkinsonism severity. We trained interpretable deep learning models with multi-task learning to simultaneously score FOG (cross-validated F1 score 97.6%), identify medication state (OFF vs. ON levodopa; cross-validated F1 score 96.8%), and measure total PD severity (MDS-UPDRS-III score prediction error ≤ 2.7 points) using kinematic data of a well-characterized sample of N = 57 patients during levodopa challenge tests. The proposed model was able to <i>explain</i> how kinematic movements are associated with each FOG severity level that were highly consistent with the features, in which movement disorders specialists are trained to identify as characteristics of freezing. Overall, we demonstrate that deep learning models’ capability to capture complex movement patterns in kinematic data can automatically and objectively score FOG with high accuracy. These models have the potential to discover novel kinematic biomarkers for FOG that can be used for hypothesis generation and potentially as clinical trial outcome measures. |
first_indexed | 2024-03-11T08:12:25Z |
format | Article |
id | doaj.art-fbc56730aef44d1b903a0db13eefd447 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-11T08:12:25Z |
publishDate | 2023-02-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-fbc56730aef44d1b903a0db13eefd4472023-11-16T23:05:50ZengMDPI AGSensors1424-82202023-02-01234176610.3390/s23041766An Explainable Spatial-Temporal Graphical Convolutional Network to Score Freezing of Gait in Parkinsonian PatientsHyeokhyen Kwon0Gari D. Clifford1Imari Genias2Doug Bernhard3Christine D. Esper4Stewart A. Factor5J. Lucas McKay6Department of Biomedical Informatics, School of Medicine, Emory University, Atlanta, GA 30322, USADepartment of Biomedical Informatics, School of Medicine, Emory University, Atlanta, GA 30322, USADepartment of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, GA 30322, USAJean and Paul Amos Parkinson’s Disease and Movement Disorders Program, Department of Neurology, School of Medicine, Emory University, Atlanta, GA 30322, USAJean and Paul Amos Parkinson’s Disease and Movement Disorders Program, Department of Neurology, School of Medicine, Emory University, Atlanta, GA 30322, USAJean and Paul Amos Parkinson’s Disease and Movement Disorders Program, Department of Neurology, School of Medicine, Emory University, Atlanta, GA 30322, USADepartment of Biomedical Informatics, School of Medicine, Emory University, Atlanta, GA 30322, USAFreezing of gait (FOG) is a poorly understood heterogeneous gait disorder seen in patients with parkinsonism which contributes to significant morbidity and social isolation. FOG is currently measured with scales that are typically performed by movement disorders specialists (i.e., MDS-UPDRS), or through patient completed questionnaires (N-FOG-Q) both of which are inadequate in addressing the heterogeneous nature of the disorder and are unsuitable for use in clinical trials The purpose of this study was to devise a method to measure FOG objectively, hence improving our ability to identify it and accurately evaluate new therapies. A major innovation of our study is that it is the first study of its kind that uses the largest sample size (>30 h, N = 57) in order to apply explainable, multi-task deep learning models for quantifying FOG over the course of the medication cycle and at varying levels of parkinsonism severity. We trained interpretable deep learning models with multi-task learning to simultaneously score FOG (cross-validated F1 score 97.6%), identify medication state (OFF vs. ON levodopa; cross-validated F1 score 96.8%), and measure total PD severity (MDS-UPDRS-III score prediction error ≤ 2.7 points) using kinematic data of a well-characterized sample of N = 57 patients during levodopa challenge tests. The proposed model was able to <i>explain</i> how kinematic movements are associated with each FOG severity level that were highly consistent with the features, in which movement disorders specialists are trained to identify as characteristics of freezing. Overall, we demonstrate that deep learning models’ capability to capture complex movement patterns in kinematic data can automatically and objectively score FOG with high accuracy. These models have the potential to discover novel kinematic biomarkers for FOG that can be used for hypothesis generation and potentially as clinical trial outcome measures.https://www.mdpi.com/1424-8220/23/4/1766deep learningmotion capturemulti-task learningParkinson’s disease |
spellingShingle | Hyeokhyen Kwon Gari D. Clifford Imari Genias Doug Bernhard Christine D. Esper Stewart A. Factor J. Lucas McKay An Explainable Spatial-Temporal Graphical Convolutional Network to Score Freezing of Gait in Parkinsonian Patients Sensors deep learning motion capture multi-task learning Parkinson’s disease |
title | An Explainable Spatial-Temporal Graphical Convolutional Network to Score Freezing of Gait in Parkinsonian Patients |
title_full | An Explainable Spatial-Temporal Graphical Convolutional Network to Score Freezing of Gait in Parkinsonian Patients |
title_fullStr | An Explainable Spatial-Temporal Graphical Convolutional Network to Score Freezing of Gait in Parkinsonian Patients |
title_full_unstemmed | An Explainable Spatial-Temporal Graphical Convolutional Network to Score Freezing of Gait in Parkinsonian Patients |
title_short | An Explainable Spatial-Temporal Graphical Convolutional Network to Score Freezing of Gait in Parkinsonian Patients |
title_sort | explainable spatial temporal graphical convolutional network to score freezing of gait in parkinsonian patients |
topic | deep learning motion capture multi-task learning Parkinson’s disease |
url | https://www.mdpi.com/1424-8220/23/4/1766 |
work_keys_str_mv | AT hyeokhyenkwon anexplainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT garidclifford anexplainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT imarigenias anexplainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT dougbernhard anexplainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT christinedesper anexplainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT stewartafactor anexplainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT jlucasmckay anexplainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT hyeokhyenkwon explainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT garidclifford explainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT imarigenias explainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT dougbernhard explainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT christinedesper explainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT stewartafactor explainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients AT jlucasmckay explainablespatialtemporalgraphicalconvolutionalnetworktoscorefreezingofgaitinparkinsonianpatients |