Arjun Balachandar1, Musleh Algarni2, Lais Oliveira2, Luca Marsili3, Aristide Merola4, Andrea Sturchio3, Alberto J Espay3, William D Hutchison5,6, Aniruddh Balasubramaniam7, Frank Rudzicz8,9,10, Alfonso Fasano11,12,13. 1. Department of Medicine, University of Toronto, Toronto, Canada. 2. Division of Neurology, Edmond J. Safra Program in Parkinson's Disease and Morton and Gloria Shulman Movement Disorders Clinic, Toronto Western Hospital, UHN, University of Toronto, 399 Bathurst St, 7MacL412, Toronto, ON, M5T 2S8, Canada. 3. James J and Joan A. Gardner Center for Parkinson's Disease and Movement Disorders, University of Cincinnati Medical Center, Cincinnati, OH, USA. 4. Department of Neurology, Ohio State University Wexner Medical Center, Columbus, OH, USA. 5. Department of Physiology, University of Toronto, Toronto, ON, Canada. 6. Krembil Research Institute, Toronto, ON, Canada. 7. Department of Mathematics, Indian Institute of Science, Bangalore, India. 8. Department of Computer Science, University of Toronto, Toronto, ON, Canada. 9. Li Ka Shing Knowledge Institute, St Michael's Hospital, Toronto, ON, Canada. 10. Vector Institute for Artificial Intelligence, Toronto, ON, Canada. 11. Division of Neurology, Edmond J. Safra Program in Parkinson's Disease and Morton and Gloria Shulman Movement Disorders Clinic, Toronto Western Hospital, UHN, University of Toronto, 399 Bathurst St, 7MacL412, Toronto, ON, M5T 2S8, Canada. alfonso.fasano@uhn.ca. 12. Krembil Research Institute, Toronto, ON, Canada. alfonso.fasano@uhn.ca. 13. CenteR for Advancing Neurotechnological Innovation to Application (CRANIA), Toronto, ON, Canada. alfonso.fasano@uhn.ca.
Abstract
BACKGROUND: Patients with essential tremor (ET), Parkinson's disease (PD) and dystonic tremor (DT) can be difficult to classify and often share similar characteristics. OBJECTIVES: To use ubiquitous smartphone accelerometers with and without clinical features to automate tremor classification using supervised machine learning, and to use unsupervised learning to evaluate if natural clusterings of patients correspond to assigned clinical diagnoses. METHODS: A supervised machine learning classifier was trained to classify 78 tremor patients using leave-one-out cross-validation to estimate performance on unseen accelerometer data. An independent cohort of 27 patients were also studied. Next, we focused on a subset of 48 patients with both smartphone-based tremor measurements and detailed clinical assessment metrics and compared two separate machine learning classifiers trained on these data. RESULTS: The classifier yielded a total accuracy of 74.4% and F1-score of 0.74 for a trinary classification with an area under the curve of 0.904, average F1-score of 0.94, specificity of 97% and sensitivity of 84% in classifying PD from ET or DT. The algorithm classified ET from non-ET with 88% accuracy, but only classified DT from non-DT with 29% accuracy. A poorer performance was found in the independent cohort. Classifiers trained on accelerometer and clinical data respectively obtained similar results. CONCLUSIONS: Machine learning classifiers achieved a high accuracy of PD, however moderate accuracy of ET, and poor accuracy of DT classification. This underscores the difficulty of using AI to classify some tremors due to lack of specificity in clinical and neuropathological features, reinforcing that they may represent overlapping syndromes.
BACKGROUND: Patients with essential tremor (ET), Parkinson's disease (PD) and dystonic tremor (DT) can be difficult to classify and often share similar characteristics. OBJECTIVES: To use ubiquitous smartphone accelerometers with and without clinical features to automate tremor classification using supervised machine learning, and to use unsupervised learning to evaluate if natural clusterings of patients correspond to assigned clinical diagnoses. METHODS: A supervised machine learning classifier was trained to classify 78 tremor patients using leave-one-out cross-validation to estimate performance on unseen accelerometer data. An independent cohort of 27 patients were also studied. Next, we focused on a subset of 48 patients with both smartphone-based tremor measurements and detailed clinical assessment metrics and compared two separate machine learning classifiers trained on these data. RESULTS: The classifier yielded a total accuracy of 74.4% and F1-score of 0.74 for a trinary classification with an area under the curve of 0.904, average F1-score of 0.94, specificity of 97% and sensitivity of 84% in classifying PD from ET or DT. The algorithm classified ET from non-ET with 88% accuracy, but only classified DT from non-DT with 29% accuracy. A poorer performance was found in the independent cohort. Classifiers trained on accelerometer and clinical data respectively obtained similar results. CONCLUSIONS: Machine learning classifiers achieved a high accuracy of PD, however moderate accuracy of ET, and poor accuracy of DT classification. This underscores the difficulty of using AI to classify some tremors due to lack of specificity in clinical and neuropathological features, reinforcing that they may represent overlapping syndromes.
Authors: T Schmitz-Hübsch; S Tezenas du Montcel; L Baliko; J Berciano; S Boesch; C Depondt; P Giunti; C Globas; J Infante; J-S Kang; B Kremer; C Mariotti; B Melegh; M Pandolfo; M Rakowicz; P Ribai; R Rola; L Schöls; S Szymanski; B P van de Warrenburg; A Dürr; T Klockgether; Roberto Fancellu Journal: Neurology Date: 2006-06-13 Impact factor: 9.910
Authors: Roberto López-Blanco; Miguel A Velasco; Antonio Méndez-Guerrero; Juan Pablo Romero; María Dolores Del Castillo; J Ignacio Serrano; Eduardo Rocon; Julián Benito-León Journal: J Neurol Sci Date: 2019-04-09 Impact factor: 3.181
Authors: Kailash P Bhatia; Peter Bain; Nin Bajaj; Rodger J Elble; Mark Hallett; Elan D Louis; Jan Raethjen; Maria Stamelou; Claudia M Testa; Guenther Deuschl Journal: Mov Disord Date: 2017-11-30 Impact factor: 10.338
Authors: Christopher G Goetz; Barbara C Tilley; Stephanie R Shaftman; Glenn T Stebbins; Stanley Fahn; Pablo Martinez-Martin; Werner Poewe; Cristina Sampaio; Matthew B Stern; Richard Dodel; Bruno Dubois; Robert Holloway; Joseph Jankovic; Jaime Kulisevsky; Anthony E Lang; Andrew Lees; Sue Leurgans; Peter A LeWitt; David Nyenhuis; C Warren Olanow; Olivier Rascol; Anette Schrag; Jeanne A Teresi; Jacobus J van Hilten; Nancy LaPelle Journal: Mov Disord Date: 2008-11-15 Impact factor: 10.338
Authors: Jean-François Daneault; Benoit Carignan; Carl Éric Codère; Abbas F Sadikot; Christian Duval Journal: Front Hum Neurosci Date: 2013-01-18 Impact factor: 3.169