Seeing is such a big part of everyday life that about half the brain is involved in the process. Shockingly, however, the part of vision with the highest fidelity occupies only 1% of our visual field. The remaining 99% falls on our peripheral vision. Our brain must use sophisticated “algorithms” to piece together vision despite not having all the information needed. Dr. Ruth Rosenholtz, of Massachusetts Institute of Technology, is fundamentally rethinking vision from the ground up, with the goal of gaining a better understanding of both underlying mechanisms and the strengths and limitations of our visual processing. For most of us, vision is a big determinant of how well we do a wide range of daily tasks; it affects our quality of life and ability to be independent. Dr. Rosenholtz’s research may eventually help better understand disorders such as dyslexia, age-related macular degeneration, and cognitive impairments as we age. In addition, her basic research aids in increasing our knowledge of how normal vision works, and therefore may also lead to the better design of tools for those with normal vision and cognition, such as maps, web pages, cell phones, and GPS systems.

While in the last decades neuroscientists and psychologists have learned more about vision than about many other neural processes, the field has still been stuck due to a lack of computational tools and theories. Dr. Rosenholtz’s commitment to computational thinking helps to answer many of the questions that remain unanswered. She and her team use state-of-the-art computational techniques from artificial intelligence, computer vision, machine learning, and image processing to understand human vision. This movement towards computational tools has allowed for the development of new models that explain vision and the brain in ways that are both precise and testable. Thus, Dr. Rosenholtz is pushing the fields of neuroscience and psychology past old models of the brain and towards novel and exciting research that makes clear predictions and true progress.

Current research includes:

• Peripheral Vision: While peripheral vision is worse than central vision, its capabilities are incredibly important determinants of how well we can perform visual tasks. Peripheral vision seems to encode its inputs in a way that makes it impossible to perfectly reconstruct the information coming into the eye. Dr. Rosenholtz is working to makes sense of the implications of this encoding and the loss of information for real-world visual tasks like navigating, recognizing our surroundings, reading, or using a map.

• Paying Attention: What does it mean to “pay attention”? Dr. Rosenholtz is studying the impact of paying attention on performance of visual tasks. What tasks require you to pay attention, and what are easy to do even when distracted?

• Decision-Making: Dr. Rosenholtz is interested in elucidating a new understanding from computer and human vision on how the brain might represent the scene in front of it. She and her team are using this understanding to shed light on a diversity of cognitive processes, such as visual memory and decision-making.