Design of Meta-semantic Analysis for Automatic Detection of Alzheimer's Disease

Nowadays, manual diagnosis of early stages of neurodegenerative disorders such as Alzheimer’s disease (AD) has been a challenge. While current neuropsychological examinations often fail to provide satisfactory result in detecting Mild Cognitive Impairment (MC) and linguistic ability has shown to be a good indication of symptoms of AD, in this thesis I examine the semantic linguistic features resulting from verbal utterances of potential patients to distinguish healthy people and people with the disease. For this purpose, I perform statistical and machine learning analysis on a specific language transcript dataset, consisting of 50 healthy people and 50 probable MCIs. Experimental and statistical evaluations suggest that certain patterns and semantic features are effective in helping the clinical diagnosis of MCI.