Much of the research done in sensory neuroscience is founded on the assumption that "sensory" function can be adequately characterized without knowledge of response dynamics, trial-to-trial variability, between-neuron interactions, or stimulus-response relationships. My lab's research demonstrates that single-neuron taste responses in gustatory cortex (GC) in fact contain dynamics that reflect tight perception-action coupling: across 1.5 sec, these responses progress from first "coding" the presence of taste on the tongue, then the identity of that taste, and finally the taste's palatability. In this talk, I will describe the tests that we have done to relate these response dynamics to changes (attentional, motivational, and learning-related) at longer time-scales, and our evidence that they reflect coherent, attractor-like processes emerging from interactions among local and distributed networks of neurons.