Computational Neuroscience szeminarium

Kedves Kollegak Az MTA KFKI RMKI Biofiziak Osztaly Computational Neuroscience szeminariumainak soraban szerdan, majus 13.-an a kovetkezo eloadasra kerul sor. Az eloadasok kezdete 15.15 , helye KFKI RMKI, III-as epulet tanacsterem Minden erdeklodot szerettel varunk. udv ep ******************************************************************* MODELING NEURAL MECHANISMS OF SELECTIVE VISUAL ATTENTION USING BIOLOGICALLY-PLAUSIBLE COMPUTER SIMULATIONS Dr. John Bickle and Ms. Marica Bernstein Focused Research Program in Computational Neuroscience East Carolina University Greenville, North Carolina, U.S.A. Vision provides an excellent sensory modality for computer simulation in biologically plausible neural network. We know more about the cell properties and connectivities in the visual system than for any other sensory modality. Yet deep puzzles remain about how these cell properties and connectivities combine to produce visual phenomena. One puzzling phenomena remains selective attention: what are the neural mechanisms by which the visual systems in mammals focuses in on particular aspects of the entire range of input, maintains focus on salient aspects, and regains focus following peripheral distractions? In this talk we will present two of our computer simulations. Our NISSL model mimics cell properties and connectivities within and between lateral geniculate nucleus (of the dorsal thalamus (LGN), cortical vertical columns of primary visual cortex (V1), and reticular nucleus of the ventral thalamus (TRN). We adapt cell properties and connectivities within these structures in an interactive activation and competition (IAC) neural network. Experiments with the model suggest that this thalamocortical system is part of a mammalian mechanism for stimulus-driven selective visual attention. Our second model (the MOUNTCASTLE model) is a bit more ambitious. Using the GENESIS program, we are modeling cell properties and connectivities within the lateral intraparietal area of posterior parietal cortex (area LIP), frontal eye fields of premotor cortex (FEF), and intermediate layers of midbrain superior colliculus (ISC). This structure contains neurons with response properties necessary for implementing vector subtraction, which enables the saccade generating system to compute a sequence of saccades from an initial fixation point. Recent physiological discoveries suggest that FEFs also contain neurons capable of sending a 'redirect' message to the saccade generating system, enabling the system to control when visual attention returns to original fixation point following peripheral distraction. Initial results with our proto-MOUNTCASTLE model suggests that this system (LIP-FEF-ISC) might be an important component of the mechanism by which we "choose" to redirect attention following distraction or continue to explore the distracting stimulus. ***********************************************************************