Technology to classify movements by analysis of quantitative continuous outputs of sensors

Parkinson's disease is one of the most common neurodegenerative disorders, caused by the progressive deterioration of dopaminergic cells in the substantia nigra pars compacta. It is the second leading cause of death in the United States, following Alzheimer's disease. The diagnosis of Parkinson's disease primarily relies on physical and neurological examinations, supported by laboratory data and structured interviews. Motor assessments are typically performed through the visual observation of trained raters, who evaluate patients from a distance as they perform motor tasks (Goetz CG, et al. Mov Disord 2008).

However, quantifying the severity of motor symptoms by the naked human eye presents significant challenges and limitations. The utilization of advanced technologies is required to address these challenges and provide more precise and objective assessments.

To reduce the uncertainty in the motor assessment of people with Parkinson’s disease by visual observation from several feet away, we developed a low-cost, quantitative, continuous measurement of movements in the extremities of individuals with Parkinson’s disease (McKay GN, et al. MethodsX 2019).

A low-cost, quantitative, continuous measurement of movements in the extremities of people with Parkinson's disease was conducted by trained raters on 20 individuals with Parkinson’s disease and 8 age- and sex-matched healthy participants with typical development. Representations of the output signals and their transforms (Harrigan TP, et al., Data Brief 2022) were evaluated by 35 trained raters. Signals and fast Fourier and continuous wavelet transforms were presented to trained raters, without clinical assessments, to be scored for halts, interruptions, amplitude decrements, and slowing. This scoring utilized a scheme (Hernandez ME, et al., MethodsX 2022) similar to the schemes used for rating clinical assessments based on visual observation (Goetz CG, et al. Mov Disord 2008; McKay GN, et al. MethodsX 2019).

An online procedure allowed 35 trained raters to complete structured ratings, including halts, amplitude decrements, and slowing of the signals and transforms (Hernandez ME, et al., MethodsX 2022). The scores for movements with no, minimal, or mild impairments were more challenging to classify compared to those indicating moderate or worse impairments. These results were analyzed using a parent regression exponential model (Y=-0.00291e1.13124+0.44694) with an alpha level of 0.05 (Brasic JR, et al., Mov Disord in press).

This poster was presented at Neurology Exchange Virtual Conference, September 20-22, 2022, www.neurology-exchange.com