Predicting the Presence of Multiple Sclerosis Using Semantic Categories and Logistic Regression
Mentor:Mortaza Jamshidian, Professor of Mathematics, California State University Fullerton
Multiple Sclerosis is a disease characterized by brain damage. Diagnosis and confirmation of MS involves investigation and discovery of that damage. An approach to this discovery is to take advantage of known facts about how the brain processes information. Research has shown that the brain processes visual information in an organized manner. In particular, many studies have concluded that the human brain organizes visual information by using the three semantic categories of Animal, Fruit, and Object. There is evidence to support that MS patients are unique in that their judgment is impaired in all three categories. Taking advantage of this information, the goal of this study was to build a predictive model for estimating the probability of a subject having MS using gender and response times to visual stimuli in the three semantic categories. Patients from a neurological research institute who met criteria for clinically definite MS volunteered to take the cognitive test. Control participants, having demographics similar to patients, were recruited from the community. The logistic regression model constructed is capable of predicting probability of MS, given the predictors. Internal validation of the model was performed using bootstrap and thresholds were determined according to desired sensitivity and specificity. Finally, details surrounding the specifics of semantic organization are implied in the interpretations of the figures produced from the logistic regression model.