NICOP - How the brain builds visual object detectors
Abstract
The mammalian visual system is organized in a hierarchical manner, with the complexity ofneuronal computations increasing as you move from the retina to higher brain regions. However, littleis known about the functional wiring of brain circuits whereby complex feature detectors, such asvisual object selective neurons, are built from simpler feature detectors. In other words, it is unclearhow the brain is wired together such that it can detect specific objects in the visual world from thevast diversity of incoming visual information. This limited knowledge of how the brain is functionallywired restricts our ability to develop effective strategies for repairing damaged and diseased brains,and also hinders the development of computer vision systems that possess human-like abilities. Visualobject selectivity is classically studied in humans and non-human primates, but these models do notpermit for detailed studies of the circuit computations underlying the construction of complex featuredetection. Fortunately, behavioral studies indicate that rodents also possess relatively complex formsof visual object recognition. With the powerful genetic, viral and imaging tools available in rodents,scientists can now examine in detail the structure and function of neuronal circuits underlying visualobject recognition. In the proposed project, we will use state-of-the-art techniques to record theactivity of visual object selective neurons in mice during the presentation of visual stimuli, and we willuse cutting-edge methods to trace the circuits belonging to these cells. Finally, combining the datafrom our different experiments, we will produce a computational model regarding how visual objectselective responses are built in the brain. This work will be innovative in that it will push forward theuse of in vivo imaging and circuit tracing techniques into less explored brain regions, and the resultswill provide insights into how brain circuits are functionally wired together to generate complexresponse properties.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Oct 17, 2018
- Source ID
- N629091812178
Entities
People
- Stuart Trenholm
Organizations
- Office of Naval Research
- United States Navy