For both MZ and DZ participants in each of the task procedures, correlational analyses revealed that switch rates during blocks 1 and 2 were highly correlated:
rs = 0.72 and 0.78, respectively,
ps < 0.0005, for binocular rivalry;
rs = 0.83 and 0.70, respectively,
ps < 0.0005, for the Necker Cube (
Figure 2). This result indicates high within-subject reliability for switching rate for both binocular rivalry and the Necker Cube perception. Next, using intraclass correlations to index similarity between twin pairs, we found that the perceptual switch rates in binocular rivalry were strongly correlated for MZ twins (
r = 0.54,
F[1, 53] = 3.36,
p < 0.0005) but not for DZ twins (
r = 0.017,
F[1, 46] = 1.04,
p > 0.45; see
Figure 3a). Similarly, we found that the perceptual switch rates for the Necker Cube were strongly correlated for MZ (
r = 0.57,
F[1, 21] = 3.61,
p < 0.002) but not DZ twins (
r = 0.057,
F[1, 16] = 1.12,
p > 0.41;
Figure 3b). Differences in the magnitude of correlation coefficients for MZ and DZ twins were then evaluated for significance by transforming raw intraclass correlation coefficients to Fisher's
Z metric and performing an
F-test of the one-tailed null hypothesis that
rMZ =
rDZ. These follow-up analyses demonstrated that
rMZ indeed exceeded
rDZ for both the binocular rivalry (
p < 0.01) and the Necker Cube (
p < 0.05) tasks. These results provide further support for previous claims that binocular rivalry switching rate is partly heritable (Miller et al.,
2010).