An article in Gait magazine The researchers found that four gait characteristics were maximally vital for the diagnosis of Parkinson's disease: step length, speed, width and consistency (or width variability). To assess the severity of the disease, significant maximum points were variability in step width and double time (during which both feet are in contact with the ground).
“Our study broke new ground compared to the clinical literature through a larger database than the same old one for diagnostic purposes. We chose gait parameters as a key criterion because gait disorders appear to be early in Parkinson's disease and worsen over time, and also because they are not correlated with physiological parameters such as age, height and weight,” Fabio Augusto Barbieri, co-author of the article, told Agência FAPESP. Barbieri is a professor in the Department of Physical Education of the Faculty of Sciences (CF) of UNESP.
The pattern examined included 63 participants of Ativa Parkinson, a systematized multidisciplinary physical activity program for Parkinson's patients performed at FC-UNESP, and 63 healthy controls. All volunteers were over 50 years old. The knowledge collected and fed into the repository used in device learning processes for seven years.
A baseline assessment produced by analyzing the gait parameters of healthy controls and comparing them with the degrees expected for this age group. This referred to the use of a special motion capture camera to measure each person's steps in length, width, duration, speed, cadence, and singleness and double downtime, as well as step variability and asymmetry.
The researchers used the knowledge to create two other device models, one to diagnose the disease and another to estimate its severity in the patient being tested. Scientists from the School of Engineering at the University of Porto in Portugal collaborated on this component of the study.
They passed knowledge through six algorithms: Naïve Bayes (NB), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Logistic Regression (LR) and Multilayer Perceptron (MLP). NB achieved a diagnostic accuracy of 84. 6%, while NB and FR scored in the severity assessment.
“The typical accuracy of clinical assessments is around 80%. It is possible that the likelihood of misdiagnosis is particularly high when combining clinical assessment with artificial intelligence,” Barbieri said.
Challenges ahead
Parkinson's disease is due, at least in part, to the degeneration of nerve cells in the spaces of the brain that move, due to deficient dopamine production. Dopamine is the neurotransmitter that transmits signals to the extremities. Low levels of dopamine affect movement and generate symptoms such as tremors, slow gait, stiffness and poor balance, as well as alterations in speech and writing.
The diagnosis is ultimately based on the patient's medical history and neurological examination, without urgent tests. Precise data are not available, but it is estimated that 3-4% of the population over 65 years of age has Parkinson's disease.
According to another co-author, Ph. D. candidate Tiago Penedo, whose studies are overseen by Barbieri, the effects of the test will be useful for better diagnostic evaluation in the future, but the burden may be only an inhibitory factor. “We made progress with the tool and helped expand the database, however, we use expensive devices that are difficult to locate in clinics and doctors' offices,” he said.
The device used in the study costs about $100,000. ” It is imaginable to analyze the gait with less expensive techniques, using a stopwatch, a force platform, etc. , but the effects are not precise,” Penedo said.
The techniques used in the study could contribute to a greater understanding of the mechanisms underlying the disease, namely gait patterns, the researchers believe.
An earlier study, reported in a paper published in 2021, with Barbieri as the last author, found a 53% lower pitch duration synergy when overcoming obstacles in Parkinson's patients than in healthy subjects of the same age and weight. The synergy here refers to the dexterity of the musculoskeletal (or musculoskeletal) formula to adapt movement, combining points such as speed and position of the feet, when descending a sidewalk, for example.