Compering AI methods-based and none AI methods-based in diagnosis Parkinson

The current treatments for Parkinson disease have no permanent-modified medicine, therapy, orother approaches, the second most common neurodegenerative disease after Alzheimer’s diseasewhich patient experience motor and non-motor symptoms and have a huge impact on quality oflife. We discuss the pros and cons of the promising artificial intelligence-based and non-artificialintelligence-based methods to not only evaluate the fastest assay to diagnose but also the besttreatment approach to increase the patient’s lifespan.Background and aims: Increasing life expectancy and slowing the progression of the disease todiagnose it in its early stages would prevent rising treatment expenses and place a heavy financialburden on society and families. Various tests and experiments need to be done in order to diagnosethe disease using non-artificial intelligence-based methods, including: inner retinal thinning, theTower of London test, a medical history and a neurological examination, blood and laboratorytests, and brain scans. And artificial intelligence-based approaches include voice database, speechaudio and hand writing database, Genetic biomarkers and other features, which have been analyzedwith different algorithms; Artificial Neural Network, Support Vector Machine, K_NearestNeighbour. As medications or surgery can often provide improvement in the motor symptoms, itis vital to diagnose Parkinson disease in its early stages.Method: From Google Scholar, IEEE Xplore, Springer, PubMed, PubMed Central, Scopus, andother literary sources, a total of ۹۰ articles have been chosen. Here we propose the advantagesand disadvantages of selected methods to diagnose de novo Parkinson disease and other stages ofParkinson. The best probabilistic model obtained by exploring the search engines and articles hasbeen summarized in this article.Results: Overall, ۹۰ articles have been selected for abstract screening and a total of ۲۴ for fulltextreview.Unfortunately, an accurate test that can be used to determine the early stages of Parkinson diseaseis not available, and multiple experiments, including MRI and computed tomography, should beconducted simultaneously. Yet, the presence of a patient’s medical history and artificial intelligencecan speed up and assist the diagnosis process.Conclusion: Using the PRISMA method, a comparison between an artificial intelligence-basedapproach and a non-artificial intelligence-based approach, or combining the two strategies, revealedthe best probabilistic method for Parkinson’s disease diagnosis and treatment.In conclusion, it appears that advancements in the collection of patient medical history data andthe application of various methods to analyze it may make the diagnosis of Parkinson’s diseasewith artificial intelligence a practical method in the near future.

کلیدواژه ها:

artificial intelligence ، motor symptoms ، none-motor symptoms ، Parkinson disease (PD) ، deep learning