Active sensing in echolocating marine mammals
Abstract
Echolocating animals effortlessly navigate, hunt, and interact with their environment, despite cluttered and noisy echo returns, wi"th performance that in some circumstances exceeds that of sonars engineered by humans. The ultimate objective of this project is to" understand the behavioral strategies and the cognitive and neural mechanisms that allow toothed whales (odontocetes) to detect, loc""ate, and identify objects through echolocation, and to relate these mechanisms to those allowing humans to make sense of sound in ev"eryday settings. Our teamaims to identify the neural mechanisms that extract echo-acoustic information and the brain networks that" build robust, invariant representations of auditory objects in complex auditory scenes. To this end, we will study the mechanisms a""nd strategies both odontocetes and humans use to hear (i.e., process acoustic energy emitted by sound sources) and how this relates"" to how odontocetes echolocate. Our hypothesis is that, despite having specialized sensory processing to extract echolocation-specif""ic acoustic information from echo-like sounds, the high-levelcognitive processes that control active sensing during odontocete echo"location are likely homologous to those engaged during human hearing. Work is divided into three interrelated thrusts. Thrust I uses a combination of behavioral and neural measures to explore how dolphins and porpoises hear and echolocate. We will study how odonto"cetes extract echo-acoustic cues to form invariant representations of targets, allowing them to generalize recognition of a target a"cross different aspect-dependent views. The combined movement and acoustic behaviors offree-swimming odontocetes will be quantified" as they solve different echolocation tasks, so that their echolocation strategies can be compared to those predicted by theoretical" models. We propose to use non-invasive electroencephalography (EEG) and to develop the capacity to use functional near infrared spectroscopy (fNIRS) and functional magnetic resonance imaging (fMRI) methods to measure spatial and temporal patterns of neural activity across brain regions associated with hearing and echolocation. Thrust II studies the brain networks controlling hearing in humans" by combining behavioral performance and neuroimaging methods (EEG,magnetoencephalography (MEG), fMRI, and fNIRS). The goal is to u"nderstand the neural mechanisms of active sensing of acoustic inputs. We will integrate information from M/EEG and fMRI/fNIRs in order to investigate the neural dynamics of mechanisms important for processing complex auditory inputs. Thrust III develops informatio"n theoretic models for echolocation, positing that echolocators strive to maximize the information available about an unknown scene."" We will formulate models for how echolocators accrue information about the number, location,and structure of objects by moving thr""ough space, changing orientation to targets, directing beams, and modifying transmitted signals, incorporating data about how echolo"cators~ attention and prior information alters the processing of observed echolocations signals. Insights from this project promise" to lead to more robust, more autonomous sonar systems that achieve detection and discrimination performance closer to that of biolo""gical echolocators.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Dec 20, 2017
- Source ID
- N000141812062
Entities
People
- Peter Tyack
Organizations
- Office of Naval Research
- United States Navy
- University of St Andrews