SCS Undergraduate Thesis Topics
|Jay Pujara||Tom Mitchell||Machine Learning Classification of fMRI Data in Semantic and Syntactic Tasks|
The advent of functional Magnetic Resonance Imaging has been an incredible boon for cognitive scientists in the last decade. By applying the power of computational processing to the tremendous quantity of data generated by fMRI studies, brain activity can be analyzed at a finer granularity than previously possible. My current research involves using the naive Bayes classifier to classify fMRI data between semantic and syntactic tasks. The goal of this research is to highlight the critical differences between the activity patterns seen when classifying isolated words as opposed to simple sentences. Through the analysis of classifier performance, I hope to lend new insight to the current understanding of linguistic comprehension. At the same time, methodology for such analysis must be constantly refined to eliminate confounding influences and lend credence to the results of classification.