Title : Beyond traditional screening: AI-driven early detection of cognitive disorders and dementia
Abstract:
In Hungary, the process of diagnosing dementia is slow and time-consuming. Currently, there are no well-developed systems and strategies at the local level for screening, and there is no artificial intelligence algorithm that could determine early signs of dementia based on the patient`s digital behavioral patterns.
Two PILOTs were conducted using the PreDem platform. Over the PILOTs` duration (2021.09.01.-2024.01.11), 42,711 test data were analyzed. The 259 participants completed SDMT-type tests, Stroop tests, and memory/word games.
We established a unified system for task evaluation, facilitating cross-test result comparisons. Participants were grouped into three categories: 1st - presumed dementia patients, 2nd - diagnosed MS patients, and 3rd - presumably normal population. Comparing the first two groups to the normal population revealed significant differences, vividly illustrated by density functions. Our results show that the first symptoms of dementia can appear as early as the 20s-30s, and can be clearly detected in the 40s and 50s.
The platform used in the study is a promising new tool for the early detection and prevention of dementia. During the PILOT studies, the platform reliably identified dementia patients, significantly surpassing the accuracy of traditional diagnostic methods. Therefore, it can be said that this method is suitable for detecting the first, otherwise unnoticed signs of dementia through risk analysis based on artificial intelligence processing. Early detection is crucial for effective management of dementia. Early diagnosis allows patients to begin necessary treatments, which can slow the progression of the disease, preserve cognitive functions, and improve quality of life.
