Experts from St. Petersburg Electrotechnical University "LETI" have presented an innovative algorithm based on a neural network that can detect signs of cognitive impairment, including Parkinson's disease.
One of the most accessible methods for diagnosing the disease is drawing a spiral using dots on a piece of paper. Deviations in the drawing, such as hand tremors, may indicate cognitive impairments. Currently, the evaluation of such drawings is done manually by doctors, which can be subjective and depend on their experience, highlighting the need for the introduction of more objective tools.
According to the author of the work, Ekaterina Syrtsova, the university has created an algorithm that analyzes photographs of spirals drawn by patients and helps to make a diagnosis. The neural network ensures high accuracy of image analysis.
The test results of the algorithm showed an accuracy of 99.3%. This development has the potential to become the basis for systems that support doctors in the rapid diagnosis of neurodegenerative diseases.
In the future, if there is interest from medical institutions, the algorithm can be integrated into decision support systems for doctors in the diagnosis of Parkinson's disease. The algorithm may also be useful for patients to independently monitor their condition.
Read materials on the topic:
Belgorod scientists received equipment for studying Alzheimer's and Parkinson's diseases
A new center for neuroprosthetics and cybernetic medicine will be created in Moscow