Article ID Journal Published Year Pages File Type
9653510 Neurocomputing 2005 8 Pages PDF
Abstract
We study a canonical model of type-I membranes subject to correlated fluctuating input currents. We present a semi-analytical approach for calculating the response of this neuron model to time dependent inputs both in the input current and the noise amplitude using a novel sparse matrix representation of the systems Fokker-Planck operator. It turns out, that the maximum stimulation frequency which can be transmitted through this model neuron is approximately given by the stationary firing rate. Our results agree well with the behavior of a conductance-based model-neuron but are in qualitative disagreement with key response properties of leaky integrate-and-fire neurons.
Related Topics
Physical Sciences and Engineering Computer Science Artificial Intelligence
Authors
, , ,