(a) In the novelty place preference (NPP) test, Igf2-P0 KO mice (N=19) spent less time in the unfamiliar and more aversive novel chamber than WT controls (N=22), Igf2-total KO mice (N=9) and WT controls (N=18) (GENOTYPE, F2,64=12.15, P<0.001, eta2=0.16; post-hoc, F1,57=7.55, P<0.05). Although all mice were more wary of entering the novel chamber if it was white in colour (rather than black), the overall statistical pattern remained consistent regardless of chamber colour. (b) Representative video-tracking data illustrating the hesitancy of the Igf2-P0 KO mice (bottom) to enter the novel chamber (grey: habituated/familiar side, black: novel chamber) compared with Igf2-P0 WT mice (top). (c) Further evidence for selective increases in acute anxiety in the Igf2-P0 KO mice (N=36) indexed by increased food neophobia (that is, decreased preference for a novel foodstuff) (GENOTYPE, F2,125=6.59, P<0.03, eta2=0.07; post-hoc, F1,101=4.67, P<0.01), WT controls (N=37), Igf2-total KO mice (N=20) and WT controls (N=30). (d) Igf2-P0 KO mice (N=21) also showed selective enhanced reactivity to stressful acoustic startling stimuli (GENOTYPE, F2,80=7.53, P<0.01, eta2=0.2; post-hoc, F1,67=17.08, P<0.001). WT controls (N=37), Igf2-total KO mice (N=15) and WT controls (N=11). The data were subjected to further statistical analysis, as co-varying for variance in the data due to differences in the litter-mate WT groups was not significant in the initial analysis (ANCOVA, Supplementary Table S3). Using two-way analysis of variance with factors STRAIN (Igf2-P0, Igf2-total) and GENOTYPE (WT, KO), significant interactions were found for the NPP and startle data, but not preference for a novel foodstuff (c), although there was a strong tendency for a significant interaction (see Supplementary Table S4). Subsequent post-hoc pairwise analysis showed that Igf2-P0 KO mice significantly differed from the other three groups of mice (Supplementary Table S7–S10). Data are mean+s.e.m., *significant difference from control (α<0.05). Note, data show nominal P-values; these P-values remained significant, taking into account potential false-discovery rates due to multiple comparisons across all the separate tests of anxiety, as they did for the additional statistical analysis (see Supplementary Tables S2 and S4).