TY - JOUR
T1 - Distinguishing Patients With a Coordination Disorder From Healthy Controls Using Local Features of Movement Trajectories During the Finger-to-Nose Test
AU - Soancatl Aguilar, Venustiano
AU - Martinez Manzanera, Octavio
AU - Sival, Deborah A.
AU - Maurits, Natasha M.
AU - Roerdink, Jos B. T. M.
PY - 2019/6
Y1 - 2019/6
N2 - Assessment of coordination disorders is valuable for monitoring progression of patients, distinguishing healthy and pathological conditions, and ultimately aiding in clinical decision making, thereby offering the possibility to improve medical care or rehabilitation. A common method to assess movement disorders is by using clinical rating scales. However, rating scales depend on the evaluation and interpretation of an observer, implying that subjective phenotypic assignment precedes the application of the scales. Objective and more accurate methods are under continuous development but gold standards are still scarce. Here, we show how a method we previously developed, originally aimed at assessing dynamic balance by a probabilistic generalized linear model, can be used to assess a broader range of functional movements. In this paper, the method is applied to distinguish patients with coordination disorders from healthy controls. We focused on movements recorded during the finger-to-nose task (FNT), which is commonly used to assess coordination disorders. We also compared clinical FNT scores and model scores. Our method achieved 84% classification accuracy in distinguishing patients and healthy participants, using only two features. Future work could entail testing the reliability of the method by using additional features and other clinical tests such as finger chasing, quiet standing, and/or usage of tracking devices such as depth cameras or force plates.
AB - Assessment of coordination disorders is valuable for monitoring progression of patients, distinguishing healthy and pathological conditions, and ultimately aiding in clinical decision making, thereby offering the possibility to improve medical care or rehabilitation. A common method to assess movement disorders is by using clinical rating scales. However, rating scales depend on the evaluation and interpretation of an observer, implying that subjective phenotypic assignment precedes the application of the scales. Objective and more accurate methods are under continuous development but gold standards are still scarce. Here, we show how a method we previously developed, originally aimed at assessing dynamic balance by a probabilistic generalized linear model, can be used to assess a broader range of functional movements. In this paper, the method is applied to distinguish patients with coordination disorders from healthy controls. We focused on movements recorded during the finger-to-nose task (FNT), which is commonly used to assess coordination disorders. We also compared clinical FNT scores and model scores. Our method achieved 84% classification accuracy in distinguishing patients and healthy participants, using only two features. Future work could entail testing the reliability of the method by using additional features and other clinical tests such as finger chasing, quiet standing, and/or usage of tracking devices such as depth cameras or force plates.
KW - Coordination disorders
KW - classification
KW - finger-to-nose test
KW - generalized linear models
KW - instantaneous speed and local curvature
KW - ATAXIA
KW - CLASSIFICATION
KW - RELIABILITY
KW - CHILDREN
U2 - 10.1109/TBME.2018.2878626
DO - 10.1109/TBME.2018.2878626
M3 - Article
C2 - 30371352
SN - 1558-2531
VL - 66
SP - 1714
EP - 1722
JO - IEEE Trans. Biomedical Engineering
JF - IEEE Trans. Biomedical Engineering
IS - 6
ER -