- November 23, 2021
New research was recently published by the group focusing on a new artificial intelligence model that groups typical symptom patterns of Parkinson’s disease. The model can predict the progression of the disease by finding the timing and severity of known symptoms. The model predicts timing and severity by learning from longitudinal patient data.
Details of the new AI model were published in “Lancet Digital Health,” with researchers noting the model can predict the timing and severity of the disease by leveraging longitudinal patient data, which is a description of the patient’s clinical status gathered over time. Researchers say their goal is to use AI to help patient management and clinical trial design.
Parkinson’s is a fairly common disease impacting as many as 6 million people globally. Despite how well-known the condition is and how widespread around the globe, people fighting the disease experience various motor and non-motor symptoms. The goal of the new AI is to use machine learning to learn from large amounts of patient data and provide clinicians and researchers a better tool to predict the progression of symptoms in individual patients.
Researchers note the patient data used by the AI has been de-identified, and it’s one of the largest Parkinson’s datasets in the world. Having access to such a massive data set is critical for success in machine learning models. Past studies focused on characterizing Parkinson’s disease using baseline information. However, the new method relies on up to seven years of patient data. Despite the diverse progression pathways of the disease, the AI model can make accurate predictions.
REFERENCE: Slash Gear; 30 JUL 2021; Shane McGlaun