ENEA and Tor Vergata Hospital in Rome have developed an innovative video analysis system for early diagnosis of Parkinson’s disease and personalisation of drug therapy. Published in the international journal Applied Sciences, the research focused on alterations in motor skills, such as changes in balance, posture and gait, which are hallmarks of the disease. The analysis of these abilities provides relevant early clues for the diagnosis and study of disease progression, facilitating the initiation and evaluation of therapy.
“For years the scientific community has been trying to find new measures to quantify the motor and non-motor skills of patients under observation in an objective, standardised and consistent manner. The video analysis system we developed adopts modern deep learning techniques, i.e. artificial intelligence that can detect a person’s posture from images taken by the camera”, explained ENEA project leader Andrea Zanela, a researcher in the Robotics and Artificial Intelligence Laboratory.
The new system determines the 3D position of the main joints of the person being filmed and then identifies the characteristics of their body movements by calculating the kinematic parameters of the observed points and the segments joining them.
“The system is not only able to qualitatively recognise and highlight the same gait disorders assessed by the scores of the scales normally used, but it offers a quantitative analysis of the results, detecting a wider and more finely graded range of motor disorders, giving the physician all the richness of an instrumental measurement”, Zanela continued.
This innovative aspect means that the true status of a patient and the progress of the disease can be determined with a high degree of accuracy without affecting their normal activities, both in the early phase of Parkinson’s when symptoms are mild and in the management of therapy.
“The system is therefore not only effective but also has a low impact on the person, whose capabilities – even residual – are assessed by the system in a manner that is as accurate as it is respectful and ergonomic. Therefore, the system developed can be employed via modern telemedicine on an ongoing basis and in places where the people routinely live and work”, Zanela concluded.
Parkinson’s disease is the second-most frequent neurodegenerative disorder after Alzheimer’s disease. In the most industrialised countries it affects about 12 out of every 100,000 people each year. In Europe 1.2 million people currently suffer from this disease, and 6.3 million worldwide.
Initial symptoms make it particularly difficult to arrive at an early, certain diagnosis and may evolve at variable rates over the years, making it hard to monitor its natural course or the response to therapy. Moreover, treatment options are also limited as there are no therapies available to prevent and/or halt the neurodegenerative process.
For more information
Andrea Zanela, ENEA – Robotics and Artificial Intelligence Laboratory, email@example.com