Broadly speaking: assigning a numerical score to the “unusualness” of data relative to a specific theory. In this entry: computing a p-value for a statistical hypothesis test with emphasis on spike-train resampling.
Deciding whether the patterns observed in experimental data support or challenge existing theories is a recurring theme in data analysis. Assigning a numerical score, or significance, is a central part of the decision process. Hypothesis testing is the classical statistical approach with p-values being the most common (but not the exclusive) method of communicating statistical significance. This entry is about basic statistical hypothesis testing with a focus on the connection between hypothesis testing and the so-called resampling methods that are gaining widespread popularity in the statistical analysis of neural firing patterns.
A comprehensive statistical analysis rarely...