Large-scale imaging to inform neuro-computations for 3D sonar object representation
Abstract
"A description of this effort can be found under DURIP in PE0601103N. We request funds to purchase major equipment to conduct transf,ormative research and-training in areas that align with a key mission of the ONR, namely to advance sonar-technology for detecting,, discriminating and recognizing 3D objects, particularly in the-presence of clutter. These advances rely on innovations in distribut,ed computation, which-will be informed by novel investigations of the bat sonar receiver. The bat exhibits-extraordinary computation,al capacity to processes complex and dynamic echo patterns,-which inform split-second decisions to track targets or steer around obs,tacles. How does-the bat accomplish these daunting tasks? Answers will emerge from new knowledge-about signal processing in the bat?,s sonar receiver, i.e. its brain.-Past research shows that the bat brain is comprised of different compartments that each-process so,und, but this knowledge is critically incomplete, as it comes from studies of-small numbers of neurons within an isolated brain regi,on. This patch-work approach has-failed to investigate how different brain compartments collectively extract high-level-sonar featur,es. Our project will overcome these limitations by exploiting new technology,-a two-photon mesoscope. Mesoscopic imaging will allow, us for the first time to study-sound-evoked activity in thousands of neurons with single cell resolution, across an ultrawide-field, of view that encompasses several interconnected compartments of the bat?s-sonar receiver. The ability to monitor the activity of th,ousands of neurons in separate-brain regions, in a nearly simultaneously manner, will allow us to unravel the neural-computations th,at give rise to 3D sonar object recognition, figure-ground segregation and-attentional modulation of these processes. The requested, equipment will enable us to-compare the auditory receivers of bats and mice to identify specializations for 3D sonar-signal process,ing. Graduate students and postdocs working on this project will receive-training in cutting-edge neuroscience research techniques a,nd computational-modeling. The scientific merit of mesoscope imaging data is tied directly to the wide-range of possibilities for ne,w computational models and algorithms that will advance-sonar technology.-In summary, innovative research with a two-photon ultra wi,de-field mesoscope will-contribute key advances in the ONR priority of sensing and sense-making, with a focus-on cognitive neuroscie,nce of perception and attention and biological and physiological-monitoring and modeling. Use of this equipment will broaden the sci,entific scope and-impact of research we are conducting as part of the ONR MURI N00014-17-1-2736 and-will continue through future DoD,-funded research. Experimental work conducted with-the mesoscope will generate unprecedented data sets of broad importance to the Do,D,-including applications in sonar technology, signal processing, object representation,-neural plasticity, artificial intelligence, and scene understanding."--
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 08, 2022
- Source ID
- N000142212695
Entities
People
- Cynthia F. Moss
Organizations
- Johns Hopkins University
- Office of Naval Research
- United States Navy