|
Abstract:
|
This research provides a new approach to analyzing the statistical properties of limited data sets. A lack of sufficient data is a pressing issue in the neuroscience community, where researchers want to understand the statistics of interactions be tween neurons. The methods used in this work involve an information-theoretic criterion for determining a class of models that can accurately describe the available neural data's statistics. These methods combine principles from Minimum Description Length coding and an approach based on the Kullback-Leibler distance. In the context of limited data, this research provides a less forced or error-prone view of the statistical properties of the data. This provides researchers in the neuroscience community with a more robust approach to data analysis. |