Computational Models of Neural Representations in the Human Brain

Abstract For many centuries scientists have wondered how the human brain represents thoughts in terms of the underlying biology of neural activity. Philosophers, linguists, cognitive scientists and others have proposed theories, for example suggesting that the brain organizes conceptual information in hierarchies of concepts, or that it instead represents different concepts in different local regions of the cortex. Over the past decade rapid progress has been made on the study of human brain function, driven by the advent of modern brain imaging methods such as functional Magnetic Resonance Imaging (fMRI), which is able to produce three dimensional images of brain activity at a spatial resolution of approximately one millimeter. Using fMRI we have spent several years exploring the question of how the brain represents the meanings of invididual words in terms of patterns l activity observed with fMRI. The talk accompanying this abstract will present our results, and the use of machine...

• Type: Conference
• Year: 2008
• Where: ALT
• Authors: Tom M. Mitchell
• Keywords: ALT 2008, Brain Activity, Fmri Data, Machine Learning, Neural Activity