Wearable data with AI effectively integrates symptoms of Parkinson’s disease

Substantia nigra in the human brain, illustration
Black core. Illustration showing healthy black matter in a human brain. The substantia nigra plays an important role in reward, addiction, and movement. Degeneration of this structure is characteristic of Parkinson’s disease. [KATERYNA KON/SCIENCE PHOTO LIBRARY/Getty Images]

New research from brain research and advocacy non-profit Cohen Veterans Bioscience (CVB) shows that wearable data combined with an artificial intelligence (AI) algorithm can effectively identify people with disabilities. Parkinson’s disease And those who do not have.

Results, Posted last week in the magazine sensors powered by Michael J Fox Foundation (MJFF), data analysis from Parkinson’s Disease Progress Signs Initiative The PPMI group collected data continuously using the Verily Study Watch in the subject’s natural environment. The researchers then used this data to create and train an AI algorithm to see if it could detect PD from a person’s daily activity.

In a proof-of-concept study, the research team used inertial sensor data from Verily Study watches that subjects wore for up to 23 hours per day over several months to distinguish between seven people with cerebral palsy and four people without.

“Because movement-related Parkinson’s symptoms such as bradykinesia and gait abnormalities typically appear when a person with diabetes walks, we initially used Human Activity Recognition (HAR) techniques to identify walking-like activity in the unrestricted, unclassified data,” the researchers wrote next. We used these ‘walking-like’ events to train one-dimensional convolutional neural networks (1D-CNN) to determine the presence of PD.”

The new AI tool delivered good results with researchers who were about to characterize people with and without PD using their wearable data with 90% accuracy in individual walking actions and 100% accuracy when analyzing data collected for an entire day.

“This study demonstrates the feasibility of leveraging untethered, unlabeled wearable sensor data for accurate detection of Parkinson’s disease using powerful deep learning methods,” said Lee Lancashire, the study’s principal investigator and chief information officer at CVB. “Through this combination of wearables and artificial intelligence, we are one step closer to monitoring individual health care-related activity, such as motor function outside the clinic, and unlocking the potential for early detection and diagnosis of diseases such as Parkinson’s.”

There is a need to develop innovative methods for identifying and diagnosing Parkinson’s disease, as it is one of the fastest growing neurological disorders and there are currently no objective biomarkers of the disease. Due to the gradual decline in motor and non-motor symptoms (eg, cognition and mood), the diagnosis of Parkinson’s today is complex and usually relies on subjective questionnaires to assess symptom severity, which can result in symptoms being undetected or misclassified.

Sensor technology has demonstrated its potential to aid in diagnosis and as a means of developing digital biomarkers of disease. The results of a proof-of-concept study of the use of sensors as a potential diagnosis of PD pave the way for the use of sensors as a tool to monitor symptoms and disease progression.

“Although additional studies are needed, we are excited that sensor data obtained through normal patient activity can be used to enable clinicians to monitor and categorize Parkinson’s symptoms through the ease of obtaining objective measures that can be used to improve clinical decision-making and guide therapeutic interventions.” , noted Mark Fraser, co-author of the paper and CSO of MJFF.

Leave a Comment