Skip to main content
. 2009 Nov 4;29(44):13883–13897. doi: 10.1523/JNEUROSCI.2457-09.2009

Figure 6.

Figure 6.

Quantal amplitude and CV changes in NL2 KO mice can be predicted by a uniform scaling of the unitary IPSC amplitude distribution from NL2-expressing mice. A, Quantal amplitude increases with increasing IPSC amplitude for both NL2[+/+,+/−] and NL2 KO mice (correlation: NL2[+/+,+/−] r = 0.54, NL2 KO r = 0.63; both p < 0.0001 when compared with zero; n = 43, 38; Fisher's r to z.) (A–D) are data replotted from Figure 4. B, Average quantal amplitude is calculated for each IPSC amplitude interval. Quantal amplitude increases with IPSC amplitude identically in both NL2[+/+,+/−] and NL2 KO mice (2-factor ANOVA, p < 0.0001 for unitary amplitude groups; no difference based on genotype, comparing only 4 leftmost points; for two NL2 KO points farthest to the right, n = 1). C, The frequency distribution for each IPSC amplitude interval is plotted, and a new distribution is derived by scaling down the NL2[+/+,+/−] distribution to 52% (gray circles). D, Average quantal amplitude is calculated for NL2[+/+,+/−] and NL2 KO data by weighting the amplitude (B) by the frequency (C) at each interval. Then, a prediction for the NL2 KO value is calculated by weighting the NL2[+/+,+/−] amplitudes (B) by the “52% scaled NL2[+/+,+/−]” distribution (C). The prediction error is <2%. E–H, The same calculations are performed for the CV. Data are replotted from Figure 5. E, The CV decreases with increasing IPSC amplitude for both NL2[+/+,+/−] and NL2 KO mice (correlation: NL2[+/+,+/−] r = −0.55, NL2 KO r = −0.53; both p < 0.0001 when compared with zero; n = 83, 85; Fisher's r to z). F, The CV decreases with IPSC amplitude identically in both NL2[+/+,+/−] and NL2 KO animals (2-factor ANOVA, p < 0.0001 for unitary amplitude groups; no difference based on genotype, comparing only 5 leftmost points; NL2 KO point farthest to the right, n = 1). G, As in C, the frequency distribution of the unitary amplitude is plotted for this dataset. H, Using the same calculation as in D, the CV in the NL2 KO can be predicted by a 52% scaling of the NL2[+/+,+/−] distribution. The prediction error is <3%. All data shown are means ± SEMs. The genotypic group NL2[+/+,+/−] includes both homozygous and heterozygous mice. See supplemental Figure 6, available at www.jneurosci.org as supplemental material, for log–log plots of A and E. Amp, Amplitude.