Representation of information in brain signals

Physics is unique among the sciences for its outward-looking orientation. Defining it by what it studies orphans many of its achievements. Physics is best defined by what it produces. Physics produces paradigm shifts and actionable measurement schemes for other sciences that are hobbled by ambiguity. A recent study of methods in neuroscience revealed that identifying network connections from correlations, mapping activity patterns to behaviors, and characterizing the response triggers for network elements, would fail to explain a well-known system: an old Atari. The brain is nothing like an Atari, but the point remains that a clarified approach is needed. After decades of data analysis advances, studying computation in nature, and articulating information as a fundamental quantity of our physical world (perhaps the most fundamental), the hope is that neuroscience is full of low-hanging fruit for a physicist. I'll cover a few attempts at clarifying computational neuroscience being pursued in the Wessel lab, focusing on representation. Information does not align well with lay usage or preconceptions of computation, rather representation is closest to what we mean when we say "X carries information about Y." Information cannot be created or destroyed but it can be represented in more or less useful ways. We tested new representations of old brain signals based on first principle insights into brain function and anatomy. We demonstrated improved clarity and simpler models of brain function.