The research arm of the Big Blue and The Michael J. Fox Foundation (MJFF) have built an AI model that can group typical symptom patterns of Parkinson's disease (PD) and accurately pinpoint the progression of these symptoms in a patient, despite whether or not they are taking medications to mask those symptoms.
The discovery, published in The Lancet Digital Health, was one of the key goals the two organisations had aimed to achieve at its outset. IBM Research and MJFF have been working together since July 2018 to examine how machine learning could be applied to help clinicians further understand the underlying biology of PD, particularly as it progresses so differently from individual to individual.
As part of developing the AI model, the researchers used de-identified datasets from the Parkinson's Progression Markers Initiative (PPMI).
"The dataset served as the input to the machine learning approach, enabling the discovery of complex symptom and progression patterns," IBM Research said in the research paper.
"While many previous studies have focused on characterising Parkinson's disease using only baseline information, our method relies on up to seven years of patient data. Also, our model makes limited a priori assumptions about the progression pathways, compared to previous studies."
IBM Research said, as a result, the researchers uncovered that a patient's state could vary in several factors, including the ability to perform activities of daily living; issues around slowness of movement, tremor and postural instability; and non-motor symptoms such as depression, anxiety, cognitive impairment and sleep disorders. At the same time, the model could predict when a patient would progress to a severe state of PD.
"We have found that the results support the hypothesis of diverse progression pathways, as indicated by the many disease trajectories we've observed. However, the AI model is still able to make accurate predictions. Having learned the model using one dataset, it was able to successfully predict an advanced state of Parkinson's disease associated with outcomes such as dementia and the inability to walk unassisted," IBM Research said.
From this finding, the researchers are hopeful the model can be used to assist clinicians on advising patient management, as well as identify those who may benefit from a clinical trial.
As next steps, the plan is to further refine the model by incorporating genomic and neuroimaging measurements, so it can provide even more granular characterisation of the disease.