A. Fernández Amil (Adrian) MSc

The brain can be viewed as a multi-level biophysical system operating at different spatiotemporal scales, whereby fast microscopic phenomena both generate and are affected by slow macroscopic dynamics. For instance, synchronized synaptic currents within a population of neurons generate oscillations in the local field potential, which in turn affect the membrane voltages of those same neurons, hence modulating their firing activity across time and thus closing the causal loop. These feedback loops between levels together with the biological nature of the brain’s substrate both constrain and shape the way stochastic processes (i.e., noise) percolate through levels. Therefore, the nature of information itself and the suitable representational formats in biological neural networks are largely determined by these biophysical constraints. In my PhD, I am using computational and theoretical methods to study how noise influences population dynamics, information coding, and learning in cortical and hippocampal circuits. Currently, my focus is on understanding how some physiological features of oscillating neural circuits optimize fundamental tradeoffs related to coding precision, computational capacity, robustness to noise, and energy efficiency. In addition, I am interested in applying these insights to machine learning and neuromorphic systems for robotics and artificial intelligence.