Classification of Extremity Movements by Visual Observation of Signal Transforms

Published: 19 April 2022| Version 2 | DOI: 10.17632/fjxt6cjptn.2
Contributors:
,
,
,
,
,
,

Description

A low-cost quantitative continuous measurement of movements in the extremities of people with Parkinson's disease (McKay, 2019) provides the means to express the dysfunction of movements commonly seen in people with Parkinson's disease (PD) in the form of the signals of instrumentation capturing the three-dimensional position in space of the extremities during movements. The goal of the current protocol is to provide the means to obtain objective assessments of the signals and transforms of the output of our low-cost quantitative continuous measurement of movements in the extremities of people with PD (McKay, 2019; Harrigan, 2020). To attain this end, we sought to develop a method for 35 experts to blindly rate the signals and transforms of our quantitative continuous measurement of movements in the extremities of cohorts of people with PD and control and comparison groups (McKay, 2019; Harrigan, 2020; Ziegelman, 2020). Thus, we conducted an investigation to apply our accelerometry-based method for the acquisition of motion data for the twelve tasks (McKay, 2019) on 20 patients with PD, one patient with multiple system atrophy (MSA), a condition with some traits characteristic of PD, and 8 healthy age- and sex-matched healthy individuals with typical development (TD). The original output from the instrumentation was stored on the laptop used for the study. Subsequent analysis has been restricted to the five repetitive tasks (3.4 Finger tapping, 3.5 Hand movements, 3.6 Pronation-supination movements of hands, 3.7 Toe tapping, 3.8 Leg agility) (Ziegelman 2021). The signals and the fast Fourier transforms (FFTs) and continuous wavelet transforms (CWTs) of the signals of the five repetitive items have been published (Harrigan 2020, 2022). However, the published materials express data in varying formats that cannot be correlated for blind ratings by experts without access to the original data. The published data are in separate datasets that cannot be combined by humans using visual observation. For this reason, we sought to develop a protocol to express the signals and the transforms of the five repetitive movements of each of the cohorts in a format suitable for blind rating by experts unfamiliar with the original data. In order to verify that each rater was qualified to participate as a rater, potential raters participated in weekly online research team meetings including instruction by a biomedical engineer and a mechanical engineer in the expression of output signals as FFTs and CWTs. Raters were asked to perform ratings independently without consultation with others. Because the research team of expert raters included people throughout the world, raters were allowed to view the images as long as they wanted. They were allowed to change ratings. They could take as much time as they wanted to complete and submit each set of ratings.

Files

Steps to reproduce

In order to develop a procedure to rate by visual observation the signals and their transforms of movements with a procedure analogous to the rating of the movements themselves directly by visual observation, we constructed images of the individual movements of each procedure in each location for each person. Throughout all image representations, separate panels were constructed to express the signal and its FFT or the CWT without the signal for assessments by trained raters. The images were presented randomly without identification of laterality (left or right) or status (PD, MSA, or TD) for all rating sessions. For each rating session the output representation (signal and its FFT or CWT without its signal), the location (finger, wrist, toe, or ankle), and the task (3.4 Finger tapping, 3.5 Hand movements, 3.6 Pronation-supination movements of hands, 3.7 Toe tapping, 3.8 Leg agility) were always identified to the raters. Specific protocols were constructed for the random presentations of test and retest sessions of the individual cohorts (single assessment of MSA, single assessment of PD, two assessments of PD or TD randomly presented without identification of status (PD or TD). The images of the person with MSA and the persons with PD with single assessments were presented randomly to raters as panels with six images of the x, y, and z positions of either the finger (upper row) and wrist (lower row) or the toe (upper row) and ankle (lower row) for separate ratings. The images of the persons with PD or TD with two assessments were presented without identification of status (PD or TD) as two images representing averaged representations of the x, y, and z axes for the finger and wrist or the toe and the ankle in two sequential groups (Alpha and Beta Group Study Start Links) for raters to complete before proceeding to the next. Raters were required to complete each rating session in sequence (MSA, PD with single session, PD and TD without identification of status (PD or TD) in two sequential groups. Raters were provided images of six outputs for each extremity of a test results for the assessment of a 54-year-old man with MSA with random presentations of items (1960 test, 2352 test, 2822 test, and 3386 test). After raters completed the images for the person with MSA, they were presented test present images of six outputs for each extremity of ten participants with TD who had completed single test sessions. After completing and submitting all these images, raters were presented the output in two images for each extremity of two groups of test and retest sessions of 20 participants with PD and 8 participants with TD. They were required to complete the first group of test and retest images (Alpha Group Study Start Link) before proceeding to the next one (Beta Group Study Start Link). Each group of images was counterbalanced to minimize learning effects. The overall plan is represented in the Form-Link Master Sheet.

Institutions

University of Illinois at Urbana-Champaign Neuroscience Program, Johns Hopkins University

Categories

Signal Processing, Movement Disorder, Fast Fourier Transform, Parkinson's Disease, Image Visualization, Accuracy Analysis, Application of Sensors, Continuous Wavelet Transform

License