Functional Connectivity Inference Using Time-delayed Mutual Information

Interneurons are important for computation in the brain, in particular, in the information processing involving the generation of theta oscillations in the hippocampus. We propose time-delayed mutual information to characterize information flow between a special class of interneurons (theta-driving neurons in the hippocampal CA1 region of the mouse) and theta oscillations, and show that information flows from the activity of theta-driving neurons to the theta wave. Via realistic simulations of a CA1 pyramidal neuron, we further demonstrate that theta-driving neurons possess the characteristics of the cholecystokinin-expressing basket cells (CCK-BC).

We formulate a new sampling scheme motivated by physiological receptive field structure, localized random sampling, which yields significantly improved CS image reconstructions. For each set of localized image measurements, our sampling method first randomly selects an image pixel and then measures its nearby pixels with probability depending on their distance from the initially selected pixel. We compare the uniformly-random and localized random sampling methods and show that, higher quality image reconstructions can be consistently obtained by using localized random sampling.