Learning the hierarchical organization of the frontal lobe with differential causal effects.

In this video article, accompanying the paper "An approach to learning the hierarchical organization of the frontal lobe", we discuss a data driven approach to learning brain connectivity. Hierarchical models of brain connectivity are useful to understand how the brain can process sensory information, make decisions, and perform other high-level tasks. Despite extensive research, understanding the structure of the prefrontal cortex (PFC) remains a crucial challenge. In this work, we propose an approach to studying brain signals and uncovering characteristics of the underlying neural circuity, based on the mathematics of Gaussian processes and causal strengths. For discovering causations, we propose a metric referred to as double-averaged differential causal effect, which is a variant of the recently proposed differential causal effect, and it can be used as a principled measure of the causal strength between time series. We applied this methodology to study local field potential data from the frontal lobe, where the interest was in finding the causal relationship between the medial and lateral PFC areas of the brain. Our results suggest that the medial PFC causally influences the lateral PFC.