Supplementary Figure 2: Evaluating the validity of multiple-probability fluctuation analysis of glutamate imaging.

a Histogram analysis of fluorescence responses of eEOS at a single synapse under conditions of multiple release probabilities (related to Fig. 2c,d). The color coding is the same as in Fig. 2c: blue, 0.5 mM (n = 100 events); cyan, 1 mM (n = 100 events); green, 1.5 mM (n = 50 events); white, 2 mM (n = 50 events); orange, 4 mM (n = 50 events); red, 4 mM [Ca²⁺]_e plus 4-AP (n = 50 events). Dashed lines show binomial release model distributions generated with quantal parameters estimated by multiple-probability fluctuation analysis in Fig. 2d. N _site  = 3, Q = 3.1 %, P = 0.01, 0.04, 0.14, 0.23, 0.51, and 0.83, respectively, Background noise standard deviation (SD) of 0.5 % and quantal variability (CV = 0.28) were also taken into account for the distribution. Cramér–von Mises test was used for testing goodness-of-fit. P = 0.94 (T = 0.039) for 0.5 mM [Ca²⁺]_e, P = 0.073 (T = 0.40) for 1 mM [Ca²⁺]_e, P = 0.46 (T = 0.13) for 1.5 mM [Ca²⁺]_e, P = 0.83 (T = 0.058) for 2 mM [Ca²⁺]_e, P = 0.92 (T = 0.043) for 4 mM [Ca²⁺]_e, and P = 0.59 (T = 0.098) for 4 mM [Ca²⁺]_e plus 4-AP. Panels beneath the histograms depict corresponding averaged glutamate signal images, and arrows indicate the analyzed synapse. b Relationship between expected and observed success rates of glutamate release at the same synapse in a, showing very good agreement between them. The expected success rate is calculated as 1−(1−P)^Nsite