Spike Train Mutual Information
Would be nice to compute the mutual information between multi-neuronal spike trains in Spiking Neural Networks. Perhaps we can learn some low-level spike representation, and relate this somehow to Information Bottleneck in Deep Neural Networks.