After a mouse achieved this acquisition criterion, a probe trial was administered in which they began the trial from the previously unused fourth arm. If the mice performed the probe trial correctly, then response discrimination training was completed. If a mouse made an incorrect turn, response training was continued until the mouse made five consecutive correct choices, at which time another probe trial 
was administered. This procedure was continued until the mouse reached the Tulathromycin B criterion and also made a correct choice on the probe trial. In this study, no mouse required more than 2 probe trials to reach criterion. For each mouse, the total number of trials to reach criterion was recorded. Mice with similar test scores were then pairwise assigned to the CIE and air control group, respectively, to reduce variability between treatment groups based on performance in the Response Discrimination phase of the task. In the afternoon of the Response Discrimination day, mice were then subjected to the first of 3 cycles of ethanol exposure as outlined above. After the third cycle of CIE or air exposure animals remained in their home cage for 3 days. We chose 3 days of withdrawal to ensure that the performance of animals in the following behavioral tasks was not affected by symptoms of acute withdrawal. Following this withdrawal period, animals were retested on the response discrimination task to assess retention of the previously learned association. Procedures and criterion for the Retest day were identical to those on the initial Response Discrimination day. The following day, animals were trained to reverse their previous association and to turn to the previously unrewarded side in order to obtain food reward. We analyzed the total number of trials to criterion and the number of probe trials. On the day following turn bias, animals were trained to always turn in the opposite direction of their turn bias to obtain reward as described above. This task also used 3 arms as start arms, which were alternated pseudorandomly to balance their frequency across blocks of 12 consecutive trials. However, in the Attentional Set-shifting task, an additional visual clue was present that had to be ignored by the mouse during the training phase. The visual clue consisted of vertical black and white Gomisin-D stripes on a thin plastic sheet that was attached to the side of one arm opposite the stem arm from which the mouse started the trial. Placement of the visual cue into the right or left arm, respectively, was varied pseudorandomly to balance the frequency of occurrences in each arm across blocks of 12 consecutive trials. Otherwise, training and response criteria on Response Discrimination day were identical to those described for the Reversal Learning task. For each mouse, the total number of trials until criterion was recorded and mice with similar test scores were pairwise assigned to the CIE and air control groups, respectively. In the afternoon of the response discrimination day, mice began the first of three cycles of ethanol exposure as outlined above. After the third cycle of CIE or air exposure, mice were returned to the animal vivarium for a period of 3 or 7 days until behavioral testing began. As described above for the Reversal Learning task, a 3-day withdrawal period was chosen to minimize effects of acute alcohol withdrawal on behavioral performance. A separate group of animals undergoing 7 days of withdrawal was included to provide a measure.