Healthy cerebral cortex function requires population neural activity that is coordinated across a wide range of timescales. Neural circuits must operate with quick, momentary sensorimotor responses (~10 ms), but also maintain persistent activity for short-term memory (~10000 ms), and many other timescales between these extremes. Depending on behavioral context, faster or slower timescales may be more useful, but there is no single most-important timescale. The status quo in computational neuroscience is to pick one particular function (usually associated with one, or a few particular timescales) and then build theory and experiments which address it. While useful, this cottage industry approach leaves open an important knowledge gap; we lack a unified theory that accounts holistically for the many timescales that coexist in real cortical circuits. How can a single, local population of cortical neurons generate dynamics with a broad repertoire of coexistent timescales? How is the repertoire of timescales tuned to accommodate behavioral context? Here we propose to develop fundamental theory and data analytic tools to answer these questions and to directly test important hypotheses in awake mice. Our preliminary theoretical efforts and experimental data strongly support the following hypotheses which frame our proposed work. Central hypothesis 1: Neural populations in awake cortex manifest with a very broad repertoire of timescales because they operate near criticality, which is a dynamical regime defined by an infinite range of scale-invariant timescales. Central hypothesis 2: Changes in arousal and behavior tune the cortical operating regime around criticality, allowing the system to shift the repertoire of timescales to be faster or slower. Our proposed work will be organized into four specific aims. In Aim 1, we will develop a temporal renormalization group theory (tRG) for describing and explaining the origins of multiple timescales in local cortical population spike recordings. Our theory will build upon traditional RG from physics, but specifically modified for spatially local, high-density recordings with single cell resolution. In Aim 2, we will generate an easy-access data analysis toolbox for investigating the timescale repertoire of spike recordings. In Aim 3, we will determine how timescale repertoire shifts with vigilance and behavioral context. In Aim 4, we will identify biophysical mechanisms which impact the timescale repertoire. Our theory and tools will perform fair comparisons of multiple competing theories (criticality and others) based on a holistic view of cortical timescales. We expect our results to unify and differentiate the zoo of different possible explanatory models for the timescale repertoire of cortical population dynamics.