Fig. 8: Effects of OFC inactivation on the learning of orthogonal rules.

A Representative photomicrographs of viral transfection in hM4d and control rats. comparable viral expression was observed in all animals (hM4D: n = 5; control: n = 5). B Curriculum overview of the inactivation experiment. DREADD agonist was administered only before sessions in the Non-match B problem. C Probability of responding ‘go’ across all trial types (different odors in match/non-match configurations) for hM4d group. D Same as (C) but for control group. E Accuracy in problem non-match A under the non-match (blue) and cue-identity (yellow) rules (data are shown as mean ± SEM with individual data points overlaid). Solid lines represent the hM4d group (n = 5 rats), while dashed lines represent the control group (n = 5 rats). Non-match accuracy increased with training (main effect of Session in two-way ANOVA: F(4,40) = 3.79, p = 0.011). Cue-identity accuracy showed no significant effects (see Supplementary Table 14). F Same as (E), but for problem non-match B (red title indicates this is the only problem when DREADD agonist was administered). Controls maintained higher non-match accuracy than the hM4d group (main effect of Group in two-way ANOVA: F(1,40) = 31.16, p < 0.001), with no training-related change. Cue-identity accuracy showed no significant effects. G Same as (E), but for cue-identity B problem. Non-match accuracy decreased with training and was lower in the hM4d group, with a significant Session × Group interaction. Cue-identity accuracy increased with training, with lower accuracy in the hM4d group and a significant Session × Group interaction. H Same as (E), but for the cue-identity A problem. Non-match accuracy decreased with training and was lower in the hM4d group. In contrast, cue-identity accuracy increased with training and was higher in the control group (all ANOVA results are in Supplementary Table 14).