Gustavo SudreGraduate student
Carnegie Mellon University
Center for Neural Basis of Cognition
Advisor: Tom Mitchell
I am currently a graduate student working with Dr. Tom Mitchell investigating knowledge representation in the human brain. More especifically, we look for semantic information in the brain while subjects perform several cognitive tasks. We use a multi-modal imaging approach (fMRI, MEG, EEG, ECoG) paired with machine learning techniques to classify different brain states. We also investigate generic models of cortical functions.
Previously, I worked at the Human Rehab Neural Engineering Lab studying the control of movement to develop new technologies for assisting and restoring motor function after nervous system injury and limb loss. The focus of that project was how the nervous system processes proprioception, and I analyzed magnetoencephalography (MEG) data to investigate the influence of somatosensory input in motor cortex activity. By using MEG to visualize brain processes, the hope was to not only assess the interaction among brain regions processing sensory information related to visually-guided movement, but also to identify optimal regions that show significant signal modulation that can be used to control brain-machine interfaces in real-time.
My ultimate research goal is to understand how the brain stores information. I intend to use this knowledge to develop truly intelligent machines and use their capabilities to help people. I believe that understanding and modeling how the human brain works is crucial to construct interfaces that will be able aid individuals in several tasks, such as object and speech recognition. Neural prosthesis is another area that can make use of such knowledge, especially in the realm of artificial stimulation of sensory areas of the nervous system to restore function. More specifically, I intend to focus on how the brain performs tasks that involve sensory decoding, learning, and memory, which can eventually be used in developing such intelligent machines.
I'm also involved in several projects in which I help with MEG data acquisition and analysis. With Dr. Anto Bagic, we look at the use of MEG for different clinical applications. With Dr. James Becker we investigate the use of MEG as a biomarker for Alzheimer's disease, and use MEG to research human memory disorders. With Dr. Rich Randall we use MEG to study expectation in musical sequences. Finally, I'm one of the authors/developers of rtMEG, a tool to access MEG data from Neuromag devices in real-time (please check the link for the appropriate publication to be cited).
Keywords: Computational Neuroscience, Knowledge Representation, Magnetoencephalography, Functional Imaging, Machine Learning, Pattern Recognition, Neural Networks, Bayesian Networks, Brain-machine Interfaces, Neural Prosthesis, Theoretical Neuroscience, Human Cognition, Cognitive Computing, Neural Engineering