Kim, Bomin. (2020). Machine learning based classification of prefrontal cortex activation based on severity of Alzheimer’s disease with portable fNIRS device. doi: 10.22677/thesis.200000322121
Alzheimer’s disease (AD) is an infamous degenerative brain disease accompanied with severe cognitive decline. By the average life expectancy of humanity increasing, the number of potential and actual patients of AD has been rising. Still a fundamental treatment of AD is lacking, medication which can delay the degradation of cognition is available. For patients’ and their caregivers’ well-being, early detection of AD, which makes achievable of fast medication slowing down the cognitive decline, is needed. This study gives the positive feasibility of simple and easily accessible AD diagnosis using portable fNIRS device which can cover the prefrontal cortex of user’s head. With participants undergoing working memory (WM) based task, we observed the distinctive prefrontal cortices’ activation by detecting amount of oxy-hemoglobin (HBO) and deoxy-hemoglobin (HBR). Collected data were analyzed by functional connectivity analysis. By machine learning based classification, we were able to find the possibility of diagnosing people by prefrontal cortex observation with mild cognitive impairment and AD from healthy people, which normally done by using data from magnetic resonance imaging (MRI) and positron emission tomography (PET).
Table Of Contents
Ⅰ. Introduction ··· ··················································································· 1 II. Methods and materials 2.1 Participants ··················································································· 6 2.2 Working memory task based experiments ··············································· 6 2.3 Portable fNIRS device ······································································ 8 III. Data analysis and classification 3.1 The modified Beer-Lambert law ························································ 10 3.2 Signal preprocessing ······································································ 11 3.3 Functional connectivity ·································································· 11 3.4 Machine learning based classification ·········································· 13 VI. Discussion ··················································································· 23 V. Conclusion ··················································································· 24 References ··················································································· 25
