Reliability of Visual Observation to Classify Continuous Wavelet Transforms of Signals of Extremity Movements
Abdelwahab Elshourbagy1, Liran Ziegleman2, Timothy Harrigan3, Manuel Hernandez4, James Brasic5
1Misr University for Science and Technology, 2University of Illinois, 3Johns Hopkins University, 4Carle Illinois College of Medicine, University of Illinois at Urbana-Champaign, 5Bellevue Hospital Center
Objective:

To construct a reliable instrument to virtually classify extremity movements through structured examinations

Background:

Structured motor assessments are hindered by the limits of visual observation by human examiners, infectious diseases, international conflicts, and other catastrophes. We sought to develop a tool to analyze the signals of technology during a structured motor assessment that can be remotely assessed to deliver the optimal care during catastrophes and in underserved regions. 

Design/Methods:

A low-cost quantitative continuous measurement of movements in the extremities of people with Parkinson’s disease (McKay, et al., MethodsX 2019;6:169-189) was administered to ten participants with Parkinson’s disease (PD) and eight age- and sex-matched healthy controls (HC) in person by a trained examiner and recorded by a technologist. Images of continuous wavelet transforms (CWTs) of the output signals of five repetitive tasks (finger tapping, hand movements, leg agility, pronation-supination movements of hands, toe tapping) were presented online to 31 trained raters without knowledge of the source (age, sex, diagnosis, laterality) to be classified according to: [A. Interruptions or freezing; B. Slowing; C. Amplitude reductions]. The images were presented randomly in different random orders for the initial rating session. We compared the original in-person clinical scores with the scores of the CWTs. The sum of the scores of each participant for the test of the five repetitive procedures were summed for each rater.

Results:

Utilizing SPSS, Cronbach’s alpha was 0.918 for PD and 0.895 for HC and Intraclass correlation coefficients were PD [0.255 (F test significance, 0.000) for single measures and 0.918 (F test significance, 0.000) for average measures] and HC [0.067 (F test significance, 0.000) for single measures and 0.704 (F test significance, 0.000) for average measures].

Conclusions:

Trained raters can reliably classify continuous wavelet transforms images of accelerometer signal of extremity movements of participants with Parkinson’s disease and age- and sex-matched healthy adults.

10.1212/WNL.0000000000210711
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.