This factor is an essential feature of our design and does not constitute a confound. If we had chosen to use a different version of the RFT with no negation in the conditional rule, then the matching heuristic strategy would have been easy to falsify for our subjects because the logical inference did not interfere with matching bias. Additionally, the RFT condition without negation is well known to lead participants to improved performances. Indeed, as the RFT task without negation does not elicit matching-bias strategy, it could not be used as a starting point for a matching-bias inhibition learning. Therefore, both of the used conditional reasoning tasks require the following: 1) inhibition of a heuristic strategy at the attentional level, and 2) access to the analytic strategy, which CYT387 allows participants to give the same logical response. Based on the RFT, the participants learned the irrelevance of cases named in the rule. This acquired knowledge became more ��strict�� when it was supplied as verbal and visuo-spatial executive warnings. We observed that the EL leads to a significant reduction in matching selections on the WST without increasing logical performances. The correct answer to the RFT has the same logical status as the correct answer to the WST, which was presented as the post-test. However, in the case of the WST, the correct answer contains a matched card due to the absence of a negation in the antecedent of the rule. Therefore, reasoning in the post-test task required that the participants understand that it was necessary to only consider one matching card as irrelevant whereas the A card was relevant. Although only EL was effective in permitting the inhibition of the matching card and the activation of the non-matching card, some participants included in this learning group applied the previously acquired knowledge to the RFT and considered all cases named in the conditional rule irrelevant. Alternatively, one could argue that EL allows the reasoners to infer that his/her performance on the reasoning task presented in pre-test was incorrect. However, in both conditions the learning phase began based on the participants’s response on the RFT, for which the Bortezomib experimenter explicitly explains that it is false. By subtraction, only the logical components remain present in the two learnings. Then, the experimenter explains logically to the two groups why the given response on the RFT is false. Therefore, all participants could infer that his/her performance on the reasoning task presented in pre-test was incorrect. Our results are consistent with the negative priming effect of EL.