Gaia Tavoni, a postdoctoral fellow of the Computational Neuroscience Initiative, has been named a Swartz Foundation Fellow for Theory in Neuroscience for her research proposal suggesting pathways to investigate the brain mechanisms involved in learning and memory.

Tavoni said she is fascinated by the deep connections between learning and memory.

“Much progress has been made in understanding the process of memory consolidation,” Tavoni said, “but how memories are transformed to abstract representations of the world from our experiences, and in goal-directed learning to drive behavior, are largely open questions. We know that reward, or the emotional valence of experiences, elicits powerful signals in the brain able to drive informational associations between the sensory input and our decisions and behavior outputs. But the specific mechanisms by which these favored pathways of neural activity are formed are largely unknown.”

After earning a Ph.D. in statistical physics at the École Normale Supérieure in Paris working at developing model-based methods to detect effects of learning in neural data, Tavoni came to Penn and started studying olfaction. Olfaction is an interesting world, she said, “especially in relation to associative learning and memory.

“It has happened to me several times that an odor triggered a memory of an entire experience,” she said. “I'm always impressed by how instantaneous and unintentional, implicit in a way, this recall process can be. For instance, the smell of maritime pine can instantaneously evoke in me joyful memories of camping holidays near the sea. If associated to experiences of our lives, odors can trigger some strong and very rich memories of these experiences.”

Tavoni believes this is possible because the brain’s olfactory bulb, the first structure processing incoming odor information, is placed in a favorable position to trigger formation and recall of implicit and strongly connoted memories.

“It’s placed in the oldest part of our brain,” she said, “directly connected with the centers of emotion and memory, and its network and neural properties can sustain highly synchronized oscillatory states, which, propagating to downstream areas, may give rise to new associations through neural plasticity.”

Tavoni said that a critical factor in the generation and amplification of synchrony and periodic oscillatory activity in the olfactory bulb is feedback connectivity, which can convey high-level information, such as salience of an experience, from cortical areas all the way back to the sensory periphery.

Together with Vijay Balasubramanian, Cathy and Marc Lasry Professor in the Department of Physics and Astronomy in the School of Arts and Sciences, and David Kersen, a Ph.D. student, Tavoni is building spiking network models of the olfactory system which can be simulated under different conditions to investigate the role of feedback and the interrelations between feedback, oscillatory activity and learning.

“Understanding these interrelations,” she said, “will also provide a coherent interpretation of disparate experimental results, collected in the past 30 years, about contextual, non-purely sensory modulation of the activity at the periphery of the olfactory system.”

Tavoni is also working with Joshua Gold, a professor of neuroscience in Penn’s Perelman School of Medicine, and other researchers to investigate how people make inferences and predictions about the world based on past experiences. This part of Tavoni's research has recently led to interesting theoretical predictions on the role of working memory in predictive inference under different task demands.

“In neuroscience,” Tavoni said, “it's important to have both components: theoretical and computational advances and experimental evidence. We have to find a good way of building a connection between these two things to make significant progress.”

Tavoni said that being awarded the Swartz fellowship is exciting because it will allow her to stay and work at Penn for another three years, giving her a concrete opportunity to pursue this research.