Algorithm for recognizing Parkinson's disease using picture of spiral is created in Russia

Parkinson's disease is a severe neurodegenerative disease that occurs in humans gradually, with age, at the age of 30-50, and it affects the cells of the central nervous system. Among its symptoms are slowness of movements, tremor, apathy, change in voice timbre, etc. Early diagnosis of the disease is important because modern treatment methods are aimed exclusively at slowing down the symptoms. Today, according to open source data, there are about 10 million people suffering from Parkinson's disease.

One of the simplest tests for the diagnosis of this disease is for the patient to draw a spiral by dots on a piece of paper. Trembling of the hands, for example, can indicate cognitive disorders of the brain. However, the doctor evaluates the patient's picture "by eye," based on his previous experience, so it becomes urgent to create a mechanism to achieve a more objective result.

"We have created an algorithm that enables to detect the signs that may indicate the presence of cognitive disorders according to the spiral drawn by the patient. The image is analyzed with the help of a neural network. The development can then be used as the basis of a program to support decision-making by clinicians dealing with the diagnosis of Parkinson's disease. With the help of such systems, patients will be able to monitor their condition independently," stated the Saint Petersburg Electrotechnical University of Russia.

Neuronet has undergone study in Open HandPD. This database contains spirals drawn by two groups of respondents; the healthy group: 18 people aged 19-79, and the group of people with Parkinson's disease: 74 people aged 38-78. The participants of the experiment were tested four times, and the respective data contain 368 spirals.

The results of testing the studied model: accuracy: 99.3 percent, sensitivity: 99.2 percent, specificity: 100 percent.

Sensitivity is an important characteristic for medical problems; that is, the probability of finding out the presence of the disease with the help of this algorithm is quite high.

Follow NEWS.am Medicine on Facebook and Twitter