Relationship between the mean of the inverse squared coefficient of variation (CV2) and the mean amplitude of responses in each condition. The gray lines are predictions of what would be observed if the change in Ca2+response amplitude were explained solely by changes in N, p, or q. The Bayesian information criterion (BIC) confirmed that, of the three possible models, variation in q provided the best fit to the experimental data (BIC for both p vs. q and N vs. q > 200 in favor of q).
