Unmet Need
According to the Parkinson’s foundation, 1.2 million people in the United States will be living with Parkinson’s disease by the year 2030. The present rating scale to measure the disease’s severity and progression is performed by trained clinicians and movement disorder specialists, which may not always be perfect due to the heterogeneity of the disease’s presentation. Furthermore, patients with Parkinson’s disease in rural areas are faced with limited, or no, access to specialists and providers—the present, and sole, means of accessing diagnostic and assessment processes. There exists a strong need for technology to be developed to address health disparities in care delivery for patients in rural areas, as well as to improve the standards of diagnosis, assessment, pathology, and response to treatment of patients with Parkinson’s disease, globally.
Technology Overview
Inventors at Johns Hopkins developed a low-cost accelerometer and image classification system to objectively quantify the movements and tremors of patients with Parkinson’s disease, correlating to the present standard diagnostic tool. Using machine learning, data from the low-cost accelerometer is used to deliver highly accurate classifications of Parkinson’s disease in the absence of a physician or trained specialist. This technique has applications in telemedicine for the evaluation of movements in patients. Further, it provides the foundations for identification of electronic signals to facilitate the diagnosis, monitoring, and treatment of PD and other neurodegenerative disorders.
Stage of Development
Experimental data is available demonstrating the accuracy of the approach. Repeated trials and increased sample sizes may strengthen the image classification system.
Publications
Suresh, Akanksha; Hernandez, Manuel; Brasic, James, “Predicting Scores of Repetitive Movement Measurements using Image Classification”, Mendeley Data, V1 2020, https://doi.org/10.1016/j.mex.2022.101739
Hernandez MH, Ziegelman L, Kosuri T, Hakim H, Zhao L, Mills KA, Brašić JR. Classification of extremity movements by visual observation of signals and their transforms. MethodsX 2022; 9: 101739. https://doi.org/10.1016/j.mex.2022.101739
Ziegelman L, Kosuri T, Hakim H, Zhao L, Elshourbagy A, Mills KA, Harrigan TP, Hernandez ME, Brašić JR. Dataset of quality assurance measurements of rhythmic movements. Data Brief 2023; 50:109556. https://doi.org/10.1016/j.dib.2023.109556
Singh M, Prakash P, Kaur R, Sowers R, Brašić JR, Hernandez ME. A deep learning approach for automatic and objective grading of the motor impairment severity in Parkinson’s disease for use in tele-assessments. Sensors (Basel) 2023; 23: 9004. https://doi.org/10.3390/s23219004
Prakash P, Kaur R, Levy J, Sowers R, Brašić J, Hernandez ME. A deep learning approach for grading of motor impairment severity in Parkinson’s disease. Annu Int Conf IEEE Eng Med Biol Soc 2023; 2023: 1-4. https://doi.org/10.1109/EMBC40787.2023.10341122
