Model comparisons using the LLR test for decorrelated sub-sets of the data.

LLR test for the comparison of the power law with the exponential distribution, and of the exponentially truncated power law with the lognormal, the gamma and the inverse Gaussian distribution in all data sets (i.e., 7 data sets recorded in vitro, 7 in vivo under anesthesia, and 2 in vivo in awake monkeys). Data sets were decorrelated by choosing an avalanche lag for which the autocorrelation showed small values after the initial drop (cf. Figure 3B); the chosen avalanche lags were 4–26 (in vitro), 5–14 (in vivo, anesthesia), and 10 (monkey X and Y). LLR values for the model comparison in the decorrelated sub-sets are reported as meanSD. Statistical significance (Eq. 18) is reported for the sub-set that had the largest p-value; pppp0.01).