Early detection of dementia using deep learning:potential as point of care testing

Kenji Karako

Department of Human and Engineered Environmental Studies, Graduate School of Frontier Sciences, The University of Tokyo

The advancement in the detection of Mild Cognitive Impairment （MCI）　is evolving by harnessing the power of deep learning models for early diagnosis.　Recent studies have indicated the potential to facilitate early MCI detection by utilizing information that can be easily acquired, such as speech data, facial video recordings, blood test results, and inertial information during walking.　Further advancements in data collection and model enhancement are expected to lead to the development of convenient Point of Care Testing that can conduct MCI assessments on the spot.　Particularly, data such as speech information, facial video recordings, and information during walking can be collected daily through smartphones and wearable devices, potentially enabling constant data measurement and MCI estimation in everyday life.　In the future, a comprehensive realization of MCI estimation and data acquisition is anticipated, fostering a system where individual cognitive function abnormalities can be detected instantly on-site.

Address correspondence to Dr. Kenji Karako, Department of Human and Engineered Environmental Studies, Graduate School of Frontier Sciences, The University of Tokyo （5-1-5, Kashiwanoha, Kashiwa, Chiba 277-8563, Japan）