Fig. 1: Neonatal manipulations and experimental timeline.

a Cross- and in-fostering procedures: 1 day after birth, dams remained in their home cages while offspring were transferred from their original cages to those housing their foster dams. At the end of fostering procedures, litters consisted of wild-type (WT) and knock-out (KO) mice in a 1:1 ratio. b Experimental time-schedule and allocation of the four study cohorts to the different assessments. After the cross- and in-fostering procedures, experimental subjects were allocated to two study cohorts. Subjects in cohort 1 (upper line, N = 10 per group) were tested for neurodevelopmental milestones (Supplementary Fig. 3), object recognition memory, T-maze, prepulse inhibition, general locomotion and were then used for tissue collection, specifically brain and gut samples (prefrontal cortex, hippocampus, striatum and cecum content, respectively) at 25 weeks of age. An independent cohort was used to collect samples at eye-opening for brain RNA-seq, plasma metabolic phenotyping and fecal microbiome analyses (N = 12 per group). Subjects in cohort 2 (lower line, N = 10 per group) were exposed to the Barnes maze, to the attentional set-shifting task, and then used for tissue collection specifically brain and gut samples (prefrontal cortex, hippocampus, striatum, and cecum content, respectively) at 25 weeks of age. An independent cohort (N = 3–5 per group) was used to evaluate long-term potentiation (LTP) in hippocampal slices (electrophysiology experiment). To avoid litter effects, each group in each cohort consisted of mice born to different dams; to limit test battery effects, the presentation of test paradigms was scheduled based on invasiveness-level considerations, wherein the most invasive test was performed at the end of the battery. Active maternal care was assessed, during the first 10 days of life, in dams of the cohort 2.