When someone has Parkinson’s disease, the progression of the disease needs to be frequently monitored. This usually involves numerous visits to the hospital to be examined by medical experts. Having to follow this intensive regime can put a strain on both patients and hospital resources.
Now researchers at the University of Oxford, U.K., have found a possible solution to this stressful situation: simply analyze the speech patterns of Parkinson’s patients to determine the severity of the symptoms of the disease.
A research team made up of British and U.S. scientists analyzed almost 6,000 speech recordings from 42 people with Parkinson’s and developed algorithms using the data to estimate how bad the symptoms were.
Parkinson’s is a disease affecting the nervous system. No one has so far found out exactly what causes the onset of the disease.
In primary Parkinson’s, the brain loses the ability to manufacture three chemicals. These three chemicals are important in maintaining the health of the nervous system. Common symptoms of Parkinson’s include tremors, a slow and heavy gait, and muscular stiffness.
For the U.K. study, researchers dispensed with the standard method of assessing symptom severity using the Unified Parkinson’s Disease Rating Scale (UPDRS). UPDRS requires a doctor or specialist to make a clinical evaluation of the patient’s ability to cope with a range of tasks. Instead, the researchers used simple, self-administered speech tests that didn’t require the patient’s physical presence in a clinic.
By applying a wide range of known speech signal processing algorithms to a large database (approximately 6,000 recordings from 42 Parkinson’s patients, recruited to a six-month, multi-center trial), the research team was able to locate algorithms that seem to reveal symptoms in Parkinson’s disease more accurately than conventional approaches.
The researchers compared estimates made using the new technique with assessments made by doctors. They found that, based on around 140 speech samples from each of the 42 patients, the estimates made using the algorithms differed from doctors’ UPDRS ratings by around two points.
The researchers say that the study provided good evidence that speech impairment and the average overall severity of other Parkinson’s disease symptoms are very closely linked. This means that symptom severity should be able to be accurately measured just by analyzing speech.
Remote monitoring, or “telemonitoring” could be particularly important for people with Parkinson’s who may find it difficult and stressful to make frequent hospital visits.