Dr. Anamika Agrawal
Neurons are often viewed as the fundamental units of computation, yet they are also living cells that must sustain this computation within the physical and metabolic limits of biology. In this talk, I present a quantitative framework for understanding neural function as a problem of optimization under constraint: how cell biological organization shapes the energetic and computational capacities of neurons and, ultimately, their susceptibility to dysfunction.
I will begin by describing models of intracellular transport and dendritic scaling that reveal how neurons distribute mitochondria to maintain energetic balance across extended and asymmetric arbors. I will then discuss theoretical work linking dendritic morphology to computational complexity, illustrating how the cellular anatomy of a neuron can also inform the computations neurons can perform. Finally, I will show how embedding such mechanistic principles into statistical inference frameworks provides new leverage for studying neurodegenerative progression, allowing us to extract interpretable dynamics from limited human pathology data.
Together, these studies highlight how mathematical modeling and cell biology can jointly illuminate the design principles, and failure modes, that govern neural function across scales. My research program will develop simplified but principled models that translate cellular physiology into computational function, allowing us to predict how changes in form, at the molecular, organelle, or morphological level, cascade into neural function and dysfunction.