TY - JOUR
T1 - Artificial intelligence-based classification of motor unit action potentials in real-world needle EMG recordings
AU - Hubers, Deborah
AU - Potters, Wouter
AU - Paalvast, Olivier
AU - de Jonge, Sterre
AU - Doelkahar, Brian
AU - Tannemaat, Martijn
AU - Wieske, Luuk
AU - Verhamme, Camiel
N1 - Funding Information: We would like to thank prof. dr. ir. M.J.A.M. van Putten for his input during the design and course of this study. Several authors of this publication are members of the Netherlands Neuromuscular Centre (NL-NMD) and the European Reference Network for rare neuromuscular diseases (ERN-EURO-NMD). Deborah Hubers: Design or conceptualization of the study; major role in the acquisition of data; analysis or interpretation of the data; drafting or revising the manuscript for intellectual content. Wouter Potters: Design or conceptualization of the study; major role in the acquisition of data; analysis or interpretation of the data; drafting or revising the manuscript for intellectual content. Olivier Paalvast: major role in the acquisition of data; analysis or interpretation of the data; drafting or revising the manuscript for intellectual content. Sterre de Jonge: major role in the acquisition of data; analysis or interpretation of the data; drafting or revising the manuscript for intellectual content. Brian Doelkahar: analysis or interpretation of the data; drafting or revising the manuscript for intellectual content. Martijn Tannemaat: major role in the acquisition of data; analysis or interpretation of the data; drafting or revising the manuscript for intellectual content. Luuk Wieske: Design or conceptualization of the study; major role in the acquisition of data; analysis or interpretation of the data; drafting or revising the manuscript for intellectual content. Camiel Verhamme: Design or conceptualization of the study; major role in the acquisition of data; analysis or interpretation of the data; drafting or revising the manuscript for intellectual content. Funding Information: D. Hubers, W. Potters, O. Paalvast, S. de Jonge, B. Doelkahar, L. Wieske, C. Verhamme report no disclosures relevant to the manuscript. M. Tannemaat reports trial support from Argenx and Alexion, consultancies for Argenx and UCB Pharma and research funding from Argenx and NMD Pharma, with all reimbursements received by Leiden University Medical Center. Publisher Copyright: © 2023 International Federation of Clinical Neurophysiology
PY - 2023/12/1
Y1 - 2023/12/1
N2 - Objective: To develop an artificial neural network (ANN) for classification of motor unit action potential (MUAP) duration in real-word, unselected and uncleaned needle electromyography (n-EMG) recordings. Methods: Two nested ANN models were trained, the first discerning muscle rest, contraction and artifacts in n-EMG recordings from 2674 individual muscles from 326 patients obtained as part of daily care. The second ANN model subsequently used segments labeled as contraction for prediction of prolonged, normal and shortened MUAPs. Model performance was assessed in one internal and two external validation datasets of 184, 30 and 50 muscles, respectively. Results: The first model discerned rest, contraction and artifacts with an accuracy of 96%. The second model predicted prolonged, normal and shortened MUAPs with an accuracy of 67%, 83% and 68% in the different validation sets. Conclusions: We developed a two-step ANN that classifies rest, muscle contraction and artifacts from real-world n-EMG recordings with very high accuracy. MUAP duration classification had moderate accuracy. Significance: This is the first study to show that an ANN can classify MUAPs in real-world n-EMG recordings highlighting the potential for AI assisted MUAP classification as a clinical tool.
AB - Objective: To develop an artificial neural network (ANN) for classification of motor unit action potential (MUAP) duration in real-word, unselected and uncleaned needle electromyography (n-EMG) recordings. Methods: Two nested ANN models were trained, the first discerning muscle rest, contraction and artifacts in n-EMG recordings from 2674 individual muscles from 326 patients obtained as part of daily care. The second ANN model subsequently used segments labeled as contraction for prediction of prolonged, normal and shortened MUAPs. Model performance was assessed in one internal and two external validation datasets of 184, 30 and 50 muscles, respectively. Results: The first model discerned rest, contraction and artifacts with an accuracy of 96%. The second model predicted prolonged, normal and shortened MUAPs with an accuracy of 67%, 83% and 68% in the different validation sets. Conclusions: We developed a two-step ANN that classifies rest, muscle contraction and artifacts from real-world n-EMG recordings with very high accuracy. MUAP duration classification had moderate accuracy. Significance: This is the first study to show that an ANN can classify MUAPs in real-world n-EMG recordings highlighting the potential for AI assisted MUAP classification as a clinical tool.
KW - Artificial intelligence
KW - Diagnostic accuracy
KW - Motor unit action potentials
KW - Needle EMG
UR - http://www.scopus.com/inward/record.url?scp=85176948123&partnerID=8YFLogxK
U2 - https://doi.org/10.1016/j.clinph.2023.10.008
DO - https://doi.org/10.1016/j.clinph.2023.10.008
M3 - Article
C2 - 37976609
SN - 1388-2457
VL - 156
SP - 220
EP - 227
JO - Clinical neurophysiology
JF - Clinical neurophysiology
ER -