IBM Research and the Michael J. Fox Foundation want to use a modeling model to better determine the current status of Parkinson’s disease

IBM Research and the Michael J. Fox Foundation want to use a modeling model to better determine the current status of Parkinson’s disease

New York, 11.8.2020

The Michael J. Fox Foundation (MJFF) and IBM Research have developed an AI model that helps physicians determine how advanced a patient’s Parkinson’s disease is. The organizations have been working together on this progress model since last year with the aim of using machine learning to find treatments and one day a cure for the neurodegenerative disease.

It has been known for years that it is difficult to determine the stage of progression in Parkinson’s disease. „In general, progression in Parkinson’s disease is neither simple nor easy to define,“ IBM Research said. The symptoms and symptom progression of this disease, „manifests itself in a wide range of patients. This makes it difficult for clinicians to definitively and quantitatively assess where a person might be at any given time and how far Parkinson’s disease has actually progressed,“ IBM Research said. An additional challenge is that drugs can significantly influence the symptoms without stopping the progression of the disease. In addition, individual patients may respond quite differently to the same medication. According to the blog post, it is precisely these individual differences that are not even taken into account in many Parkinson’s disease progression models.

With this in mind, IBM Research and the MJFF have developed a modeling methodology to help physicians understand how the disease progresses in relation to the occurrence of symptoms, so that they can determine exactly how far a patient’s Parkinson’s disease has progressed.

Using algorithms, the model has been designed to take into account factors that can obscure the external appearance of Parkinson’s disease, including drugs that can relieve symptoms such as tremors, improve motor control, and modify other common symptoms. „Patients‘ responses to medications may not be consistent across the population, which explains the need for personalized predictions,“ says IBM Research.

„For example, the tremor symptoms of one patient may respond very well to medication, while another patient may experience less relief from medication even though his or her disease is equally advanced … identifying these challenges is key to the success of machine learning in healthcare,“ IBM Research said.

IBM says it hopes that the progression model can be used by clinicians to group patients and better predict the course of disease.

As next steps, IBM plans to train the model using the PD patient data collected by the MJFF, in the hope that each stage of Parkinson’s disease can be defined.