Theoretical & Computational Neuroscience
Rava Azeredo da Silveira’s lab focuses on a range of topics in theoretical and computational neuroscience and cognitive science. These topics, however, are tied together through a central question: How does the brain represent and manipulate information?
Among the more concrete approaches to this question, the lab analyses and models neural activity in circuits that can be identified, recorded from, and perturbed experimentally, such as visual neural circuits in the retina and the cortex. Establishing links between physiological specificity and the structure of neural activity yields an understanding of circuits as building blocks of cerebral information processing. On a more abstract level, the lab investigates the representation of information in populations of neurons, from a statistical and algorithmic-rather than mechanistic-point of view, through theories of coding and data analyses. These studies aim at understanding the statistical nature of high-dimensional neural activity in different conditions, and how this serves to encode and process information from the sensory world.
In the context of cognitive studies, the lab investigates mental processes such as inference, learning, and decision-making, through both theoretical developments and behavioral experiments. A particular focus is the study of neural constraints and limitations and, further, their impact on mental processes. Neural limitations impinge on the structure and variability of mental representations, which in turn inform the cognitive algorithms that produce behavior. The lab explores the nature of neural limitations, mental representations, and cognitive algorithms, and their interrelations. Whereas most of the lab’s projects are theoretical and computational, some projects involve designing and conducting behavioral experiments in humans.
The lab moreover collaborates extensively with experimental neurobiology labs. While diverse approaches and research methods are used, they are driven by a common objective: to uncover mathematical principles and structures that capture apparently disparate facts or suggest new ways of thinking about a set of phenomena.