The proposed project aims at developing a model of cortical oscillation in the brain as they are believed to encode in a precise way all the different brain-state dynamics. Nowadays, we are able to obtain measurements of neural activities at multiple scales starting at the molecular level with detailed chemical interactions, to cellular levels with localized recordings of synaptic currents and firing dynamic, to networks with recordings from neuronal assemblies, to systems with recordings of brain activation from multiple cortical areas. The model will have the following attributes: 1) it is consistent with physiological hypotheses and experimental recordings under specific brain states, 2) it is hierarchical, describing the brain activity at multiple resolutions as well as describing brain state transitions as information-state refinements and 3) it is amenable to efficient simulations. This model will further our understanding of several phenomena evident in the cortical dy- namics of the brain: 1) it will explain the recruitment process the brain undergoes in terms of synchronization, 2) it will explain the robustness of the synchronization process, 3) it will explain the smoothness in the dynamics of brain-state transitions, 4) it will explain the relationship be- tween the number of synchronized cells and the frequency of synchronization, 5) it will explain the dynamics and integration of the top-down dynamic evolution based on internal stimuli and the bottom-up processing of sensory information, and 5) it will explain the timing properties and efficiency in execution. The derived detailed cortical model will then be integrated with our existing system-level model of the motor control system. In particular, we aim to study the role of oscillations in information transfer across constituent cortical subsystems, and to subsequently identify unique signatures of motion planning in scalp EEG recordings. Models of cortical oscillations will have utility in various ways. On one hand, if used as predictors, they can point to important experimental activities designed to invalidate the models, which will result in a substantial progress of our understanding of the brain processes. On the other hand, these models are critical as a diagnostic tool, for example, such models can be used to detect seizures and brain tumors, to calibrate drug action during anesthesia, to study cognitive task signatures, and to study mental retardation. Finally, they can provide the appropriate dynamic description for a proper design of neural prosthesis. Intellectual Merit: The proposed research provides a bridge between neurophysiology, neu- roanatomy, and engineering through the development of a consistent computational model of cortical activities. The research is conducted by a multi-disciplinary group with expertise in both modeling, analysis and design of systems in general, as well as modeling and analysis of the nervous system in particular, with a track record of collaboration through NSF support. The proposed research will be augmented with an experimental component through our ongoing motor system control collaboration with workers in Massachusetts General Hospital. Broad Impact: Beyond the utility mentioned above, this research is systematically closing the gap between the fragmented research on understanding the nervous system conducted by biolo- gists, statisticians, computer scientists and system theorists. With this unification, computational biology developments will aid in understanding and potentially curing much of the neural diseases known. Progress in this research removes the boundaries between the strongly segregated fields such as sciences and engineering. Such multi-disciplinary research requires complete interactions between different groups with different domains of expertise. This research will result in PhD theses cov- ering both such fields as demonstrated by our KDI project. In addition, courses across disciplines can be structured to provide better multi-disciplinary training at the undergraduate level.
