A neurally-inspired approach to enhance perception and performance in novel visual environments via a generative network enabled Virtual Reality
Abstract
Neuroplasticity is critical to allow sensory systems to operate effectively within diverse viewing conditions. In the visual system, processes of adaptation continuously recalibrate sensitivity in response to stimulus changes in the environment. These adjustments allow the system to optimize performance for the current stimulus context. Considering the diverse operational environments in which the military must conduct missions, these adjustments are likely to be critical – individuals gain more information and process it more efficiently when they are adapted to their environment. However, the biological processes of adaptation usually take significant time, during which performance will be impaired. This proposal will explore ways to speed up adaptation so that individuals can adjust rapidly to make optimal perceptual decisions in novel environments. In our first aim, we will explore the consequences of adapting images rather than observers. Increasingly, the exploration of new environments involves inspection of images through remote sensing. We will use generative adversarial networks to transform these images so that they are optimized for the current adapted state of the observer, by modifying images from novel environments until they cannot be discriminated from the observers’ current environment. This will eliminate the need for the observer to adapt by instead adjusting the images to match the adapted state of the observer. We will then characterize how well observers can extract information from the adapted images compared to the original novel environment. To better understand the dynamics of adaptation and to identify protocols that allow observers to adapt more quickly-effectively to different environments, in our second aim we will develop paradigms for adapting observers to novel image statistics in virtual reality (VR). We will also explore the transfer of this adaptation to real-world scenarios to determine whether individuals can be pre-adapted with VR regimens prior to deployment to varying visual environments.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 07, 2023
- Source ID
- FA95502110207
Entities
People
- Fang Jiang
Organizations
- Air Force Office of Scientific Research
- Nevada System of Higher Education
- Office of the Secretary of Defense