BI 178 Eric Shea-Brown: Neural Dynamics and Dimensions

Eric Shea-Brown joins Paul Middlebrooks to examine what dynamical systems and dimensionality reveal about neural computation. They compare task-related activity with less constrained behaviour, consider how network connection motifs shape collective dynamics, and discuss high-dimensional activity embedded within lower-dimensional descriptions. Research examples connect predictive learning, latent representations and rich versus lazy learning regimes. The conversation also examines how theoretical neuroscientists choose useful abstractions and which assumptions about network structure and function deserve closer scrutiny.