Fig. 2: Inter-test correlations and behavioral predictions as a function of variable complexity.

A Inter-test correlations between increasingly complex variables of different anxiety tests. The correlation coefficients between OF and EPM (green), EPM and LD (red), and LD and OF (blue) were plotted at different variable combinations of these tests. The #n of test events and the #n of test variables refers to how many repeated test events (1 or 3 repeats) were included and the number of different variable types (time, frequency, latency) that were used to create summary variables, respectively. SiMs are single variables, for example time spent in the open arms of an EPM, while SuMs consist of either multiple samplings of one variable type and/or multiple variable types measured in one or multiple events/repeats. B, C Network plot of inter-test correlations between different tests revealed by SiMs (simplest model) and SuMs (reduced) in the rat (B) and mouse (C) experiments. * p < 0.05 significant correlation D Behavioral predictions of anxiety measures in an aversive environment. Correlation coefficients were measured between identically composed variables in baseline and aversive contexts. SiMs represent one, and SuMs represent multiple testing events, while COMPs include SiMs or SuMs of all test types. E Correlations between increasingly complex anxiety measures (SiMs, SuMs, or COMPs) from the baseline sampling and their aversive counterparts in rats (left) and mice (right). F Behavioral predictions of anxiety measures in the fear conditioning paradigm. Correlation coefficients were measured between baseline anxiety measures (SiMs, SuMs, COMPs), freezing in the fear conditioning paradigm, and startle in the acoustic startle paradigm. G Time spent with freezing (%) in the conditioned fear paradigm in the contextual reminder (conditioned) and safe (different) contexts. Freezing was sampled two times in the safe context, from which an average freezing value was used for predictions. H Correlation of increasingly complex anxiety measures (SiMs, SuMs, or COMPs) with freezing behavior in a novel context following fear conditioning (left) and startle response to acoustic stimuli (right). COMP represents the composite of all anxiety tests and measures. Red dashed lines represent the threshold of significance on the scale of correlation coefficients. Additional numbers on the plots show the smallest and largest coefficients. Figures A-E show male Wistar rats and male C57BL/6J mice, while Figures F-H display results from experiments with male Long-Evans rats. All statistical results are presented in Fig. 5A). Statistical parameters of Fig. 2G) can be found in Table 1.