I would be grateful for any suggestions for a statistical test for a change in information. For example, taking a sample of 100 neurons with responses that have a small amount of information about a pair of stimuli (maximum = 1) and re-testing them with the same stimuli but perhaps at a higher stimulus intensity might give a slightly different value, perhaps a change from a total of 0.25 to 0.30 or some small change of this kind. As information in this situation has an upper and lower limit, this would pose considerable constraints in choosing a proper statistical test, even leaving other considerations aside. Are there any suggestions or is there any literature from physics or biology or some other research field on how the statistical significance of such a change might be evaluated?