Fig. 8

IntelliCage multivariate behavior reveals repetitiveness and compulsive licking. ChR2-flfl and Avil-ChR2 mice were exposed to blue light in a chamber on postnatal day P1-P5 together with the Cre-negative blue-insensitive mother. IntelliCage observations started at 30 weeks of age and lasted 9 weeks. The experiment included n = 15 ChR2-flfl and n = 16 Avil-ChR2 female mice. A Frequency of rapid corner re-entries with short Inter-Visit-Interval (IVI < 60 s) in learning tasks with either three correct corners (NP 3-corner, NP3c) or one correct corner (PPL, PPL-reversal). The frequency of short IVI's is increased in Avil-ChR2 mice suggesting repetitive behavior. B Repetitiveness describing the ratio of early observed versus expected returns to the same corner during different tasks (FA, NP3c, PPL, PPLrev, overall mean). The box shows the interquartile range, the line is the median, whiskers show minimum to maximum, the scatters are individual mice. Data were compared with 2-way ANOVA for the factors “Tasks” X “genotype” and posthoc comparison for “genotype” with adjustment of alpha according to Šidák.C Analysis of multiple dimensions of IntelliCage behavior represented by 29 different behavioral parameters. For each mouse, “per-day-values” of each parameter were averaged for the duration of each task (FA, NP3c, PPL, PPLrev) and for the toral observation time (overall mean). Hence, five mean values were obtained for each parameter for each mouse, and therefore each mouse is represented by five scatters for each parameter. Data were compared with 2-way ANOVA for the factors “IC-parameter” X “genotype” and posthoc comparison for “genotype” with adjustment of alpha according to Šidák. Avil-ChR2 mice show higher repetitiveness, as revealed by a higher frequency of fast returns to the same corner irrespective of the success in this corner. Avil-ChR2 mice also show a higher frequency of Licks /h and Licking duration /h. D Box/scatter plots show Licks /h during different tasks (FA, NP3c, PPL, PPLrev, overall mean). Statistics as in B. Licks were particularly high in FA where doors were open allowing licking ad libitum. During learning tasks lick duration was restricted by a door closing time of 5 s. E Random Forest importance of behavioral features for prediction of group membership (Please also see Additional file 1: Tables of RF statistics). FA free adaptation; NP3c Nosepoke adaptation with three correct corners; PPL place preference learning; PPLrev place preference reversal learning. Statistics: *P < 0.05, **P < 0.01; ***P < 0.001)