Neurons exhibit a great diversity of homeostatic mechanisms, including a large number of K+ channel subtypes that maintain neuronal excitability. Consistent with principles of “efficient coding” (Barlow, 1961), it has been proposed that the diversity of K+ channels allows neurons to predict and counteract diverse temporal patterns of synaptic excitation, a phenomenon which could be called “predictive homeostasis” (Fiorillo, 2008). According to this theory, the particular K+ channels expressed by a neuron should have properties that are matched to the statistically common temporal patterns of synaptic excitation that the neuron experiences. As a result, the neuron’s output would correspond to “prediction error,” and the neuron’s information about its stimulus would tend to be maximized. If the theory is correct, then it should be possible to predict the properties of K+ channels given knowledge of common patterns of synaptic excitation. We focus here on A-type K+ channels, which are distinguished by their strong and rapid inactivation.