Optimizing machine learning algorithms for cognitive impairment classification

 
   Lead Investigator:    Bratati Kahali
   Institution      :    Indian Institute of Science
   E-Mail           :    bratati@iisc.ac.in
   Proposal ID      :    1152



   Proposal Description:
   Cognitive assessment plays a crucial role in detecting loss of cognitive functions and change in behavioral and functional state compared to normal conditions. These tests can measure different cognitive domains (e.g., language, learning, and memory) and subdomains (e.g., long- term memory and recognition memory). Identifying the best combination of tests to be used to classify and diagnose AD, is still a matter of debate, and a large amount of subjectivity also lies in the diagnostic process. Our aim is to assess the        potential of different machine learning approaches in quantifying the cognitive decline and optimizing/reducing the number of neuropsychological tests used to classify AD patients, also at an early stage of impairment. Using machine learning algorithms, we classify data into normal, MCI, and dementia, while taking into account demographic data such as patient?s sex, age, and education level. However, the identification of all attributes which are in some circumstances relevant for classification, is a crucial problem. Our goal is to find all relevant attributes, instead of only the non-redundant ones (the correlated tests only), to answer this obstacle. In particular, this will help  us in understanding mechanisms related to the classification process for          cognitively healthy or impaired, instead of merely building a black box predictive model. We will build the feature selection process such that it takes into account the fluctuations of the mean accuracy loss among trees in the forest.                                                                       Specific aims:                                                                1. Assess the potential of machine learning approaches in quantifying the cognitive decline.  2. To use neuropsychological (cognitive tests results) and demographic data to predict cognitively healthy, MCI, AD through the implementation of machine learning models.