Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
BP: Fachverband Biologische Physik
BP 8: Neurophysics II
BP 8.3: Vortrag
Montag, 16. März 2015, 15:00–15:15, H 1058
A Frequency-resolved Mutual Information Rate — •Davide Bernardi1,2 and Benjamin Lindner1,2 — 1Bernstein Center for Computational Neuroscience, Berlin — 2Humboldt-Universität zu Berlin, Institut für Physik
The information spike trains encode about an external time-dependent stimulus is quantified by Shannon's mutual information rate. However, the numerical estimation of the mutual information rate is demanding and does not reveal which features of the stimulus are encoded. Several studies have identified mechanisms at the cellular and network level leading to low- or high-pass filtering of information, i. e. the selective coding of low- or high-frequency components of the time-dependent stimulus. However, these findings rely on an approximation, specifically, on the qualitative behavior of the coherence function, an approximate frequency-resolved measure of information flow, whose quality is generally unknown.
We developed a numerical procedure to directly calculate a frequency-resolved version of the mutual information rate. This can be used to study how different frequency components of a Gaussian stimulus are encoded in neural models without invoking a weak-signal paradigm or making undue assumptions on the nature of the neural encoding. We demonstrate its application for paradigmatic descriptions of neural firing like an integrator neuron and a simple setup mimicking a coincident detector cell receiving input from two leaky integrate-and-fire neurons.