This seminar will be held in the Jerome L. Greene Science Center on Columbia's Manhattanville campus (9th floor lecture hall). Columbia University's Intercampus Shuttle Service is the best way to travel between campuses.

Finding optimal stimuli that drive neurons has been at the core of the quest to decipher information processing in the brain. This effort has led to seminal discoveries in the field such as Barlow’s “fly detector” cells in the retina, Hubel and Wiesel’s orientation-selective cells in primary visual cortex and Gross’s “face cells” in inferotemporal cortex. However, finding the optimal sensory inputs is difficult due to the high-dimensionality of the search space and because sensory information processing is nonlinear. To mitigate this problem, Dr. Tobias' and his team have developed inception loops: a closed-loop optimization technique that combines large scale in vivorecordings with in silico deep learning modeling. They call them "inception" because they allow us to implant a desired activity pattern in the brain ("Inception" à la the movie by Nolan), and "loops" because in one pass of our protocol we start with in vivo experiments, optimize in silico responses and return to in vivo experiments in the same animal. They applied inception loops in the visual system and found that the optimal stimuli in mouse V1 exhibited complex, high spatial frequency details such as sharp corners, checkerboard patterns, irregular pointillist textures, and a variety of curved strokes, that deviate strikingly from the currentde facto standard model of V1 as a bank of Gabor filters. Inception loops are a widely applicable, potentially paradigm-shifting tool for neuroscience research. They enable, for the first time, a systematic inquiry into the functional organization of the neural mechanisms of sensation and cognition. The researchers are also reverse engineering the functional organization principles we learn from the brain into AI models.  This provides a powerful platform to test our understanding of brain function under natural complex tasks and develop the next-generation of less artificial and more intelligent algorithms.

Dr. Tolias’ research goal is to decipher brain’s mechanisms of intelligence. He studies how networks of neurons are structurally and functionally organized to process information. Research in his lab combines computational and machine learning approaches with electrophysiological (whole-cell and multi-electrode extracellular), multi-photon imaging, molecular and behavioral methods. He earned his PhD from MIT in Systems and Computational Neuroscience. The focus of research in his lab is to reverse engineer neocortical intelligence. To this end his lab is deciphering the structure of microcircuits in visual cortex (define cell types and connectivity), elucidate the computations they perform and apply these principles to develop novel machine learning algorithms.

Those who wish to meet the speaker during their visit should contact Tristan Geiller, PhD (Losonczy lab). For general inquiries please contact [email protected].

The Columbia Neuroscience Seminar series is a collaborative effort of Columbia's Zuckerman Institute, the Department of Neuroscience, the Doctoral Program in Neurobiology and Behavior and the Columbia Translational Neuroscience Initiative, and with support from the Kavli Institute for Brain Science.