Fig. 4: Developmental trajectory of visuospatial learning on the rodent touchscreen two-object sPAL test.

A Touchscreen pretraining for instrumental conditioning. Mice were allocated a maximum of 8 days for pretraining and progressed through Stages 1 and 2 within comparable numbers of sessions to criterion. Female mice at all three ages, but particularly early-adolescence (3.5 W), required more training on Stage 3 compared to age-matched male mice. Data are average number of sessions at each stage (see also Supplemental Table S1). 4.5 W: n = 13 male, n = 11 female; 6 W: n = 13 male, n = 12 female; 12 W: n = 13 male, n = 12 female. B and C Visuospatial learning and memory in B female and C male mice on the two-object sPAL task, illustrating response accuracy (% correct) across 19 sessions. Data represent mean ± SEM. Gray dotted line indicates performance at the chance (50% accuracy). D and E Mixed-effects logistic regression analysis; Point estimates are shown with 95% CI (see Supplemental Table S2 for complete statistics). Regression weights of key biological (age, sex) and task (e.g. session) variables (denoted in filled circles), and their interaction effects (denoted in italics text, open circles) on correct responding were estimated and expressed as odds ratios. An odds ratio >1 indicates an increased likelihood of correct responding, and <1 indicates a decreased likelihood of correct responding. For regression analyses, 12 W adult mice were used as the reference age thus the effect of age (e.g. 6, 4.5 W) reflects the performance of 6 or 4.5 W mice relative to adult mice. Similarly, female mice were used as the reference sex, thus the effect of sex (e.g. male) reflects the performance of male mice relative to female mice (4.5 W: n = 13 male, n = 11 female; 6 W: n = 13 male, n = 12 female; 12 W: n = 13 male, n = 12 female). D There were no significant differences due to sex, nor any sex × age interactions on sPAL response accuracy. Relative to 12 W adult mice, both 6 and 4.5 W adolescent mice showed a tendency for decreased accuracy, but only 4.5 W mice were significantly different. At both 6 and 4.5 W adolescent ages, male mice tended to show better accuracy than female mice (6 W × Sex (male), 4.5 W × Sex (male) interaction effects), but this was not statistically significant. **P < 0.01. E The effect of session was used as a proxy for the rate of task acquisition to examine sPAL learning trajectories. Mice at all three ages showed changes in their learning rate across sessions, indicating acquisition of sPAL (Session (12 W), Session (6 W), Session 4.5 W)). However, relative to the 12 W adult learning trajectory, both adolescent ages showed slower rates of improvements in response accuracy (Session × 6 W, Session × 4.5 W interaction effects). **P < 0.01, *** P < 0.001. F–G Cumulative number of trials (pseudorandom first-presentation trials and correction trials) completed across sessions by F female and G male mice. Data represent mean ± SEM. Male and female data visualized separated only for clarity. Two-way ANOVA on total trials completed showed no differences due to sex (F_{(1, 67)} = 0.18, P = 0.670), or any sex × age interaction (F_{(2, 67)} = 1.78, P = 0.177), but a significant main effect of age (F_{(2, 67)} = 3.97, P = 0.024), with Bonferroni’s multiple comparisons indicating 4.5 W mice completed significantly more trials relative to 12 W animals (P = 0.020).