Working on my PhD dissertation at the UNIVERSITY OF TEXAS
under the supervision of Prof. Risto Miikkulainen
Using Neural Networks for Prediction and Interpretation of brain activity during sentence comprehension
For years, many scientific communities have been trying to figure out how human brains represent and organize semantic knowledge. Recent advances in functional Magnetic Resonance Imaging (fMRI) provide a new approach by directly observing brain activity during word and sentence comprehension. When humans listen or read sentences, they simulate seeing the scenes and performing the actions that are described using their motor, perceptual and possible other brain systems. Subsequently, parts of the brain that control these actions light up during the fMRI experiments.
I am working on some ideas using feedforward and recurrent neural networks to map a concept representation (meaning) to its corresponding brain activation patterns in order to accomplish the following tasks: Prediction of the brain activity for novel concepts and Interpretation of the impact that context has on the representations.
Nora's Home Page
Last update: October 25, 2002.