Objective Latent inhibition (LI) denotes that a prestimulus with repeated presentation but without reinforcement (i.e.pre-expose) will interfere with the subsequent learning task related with this stimulation and its change is one important manifestation of schizophrenia.This work aims to explore whether behavioral training of reversal learning could interfere with LI of adult rats.Methods Twenty adult male Sprague-Dawley rats (body weights; 100-120 g each) were randomly divided into two groups, pre-expose group (PE) and non pre-expose group (NPE).The experimental protocol includes 3 stages: First, Gemini avoid system was used to observe LI phenomenon of rats, in which the number of active avoidance, no response, and avoiding latency were detected by the system.The LI test procedure contained adaptation phase, pre-expose phase, and conditioning phase.Second, the followed intervention training of reversal learning was conducted in the Habitest radial-arm maze system (8 arms).The process included adaptation phase, acquisition phase, and reversal learning phase, in which correct percent of running into arms of rats were observed.Third, after the training of reversal learning, LI was tested in the same approach once again.Results (1) In the first LI test, the difference of no response number, and the avoiding latency was statistically significant (two-sample t test, P < 0.05) between the PE and NPE groups, while the difference in the number of active avoidance was not significant (P > 0.05).These results mean that PE group shows latent inhibition.(2) In the intervention training of radial-arm maze, the difference in correct percentage of running into arms in two groups at the final day of acquisition phase (A8 day) and the first day of reversal learning phase (R1 day) was significant, respectively (paired t-test, P < 0.001), indicating the occurrence of the reversal learning.(3) In the second LI test, the number of active avoidances or no responses, and avoiding latency did not differ between rats in the PE and NPE groups, suggesting the disappearance of LI.(4) The comparison of LI results before and after the training of reversal learning showed that no significant difference in the number of active avoidance or no response, and avoiding latency in NPE rats was observed, but difference in three parameters in the PE group was significant (P < 0.05 or P < 0.01).Conclusion These results suggest that the reversal learning in radial-arm maze may affect latent inhibition of adult rats.