VALPROIC ACID RAT MODELS OF AUTISM SPECTRUM DISORDER SHOW NO BEHAVIORAL FLEXIBILITY DEFICITS IN A VARIABLE INTERVAL REVERSAL LEARNING TASK

Aims: To evaluate whether prenatal exposure to valproic acid impairs behavioral flexibility by slowing reversal learning in rats.
Method: Two Sprague Dawley dams were exposed to 600mg/kg of VPA on embryonic day 12.5 and other two dams were exposed to saline. Offspring began operant training at PD37 receiving water reinforcement. After magazine training, rats acquired lever pressing on a variable-interval 15 s (VI-15) schedule (20 sessions; R1 reinforced). During the reversal session, reinforcement contingencies were switched to the opposite lever (R2), while R1 responses were extinguished. To prevent inadvertent reinforcement of response chains, R2 presses were reinforced only if they occurred at least 5 s after an R1 response. A reversal index was computed for each minute of the session, ranging from −1 (exclusive R1 responding) to +1 (exclusive R2 responding), generating a minute-by-minute reversal learning curve. Reversal curves were analyzed by a generalized additive mixed model fitted by a quasibinomial distribution, testing the effects of VPA exposure and sex.
Results: Substantial variability across subjects was accounted for by random effects, with no significant main effects of prenatal VPA exposure or sex, and no significant interaction between them.
Conclusions: Prenatal valproic acid exposure did not impair reversal learning, as exposed rats were as flexible as controls and did not exhibit greater perseveration on R1. Although deficits in behavioral flexibility would be expected to support the face validity of the valproic acid model of autism spectrum disorder, no such deficits were observed under the present testing conditions. Funding: FAPESP (2024/04091-5).