In this study, we established a novel spatiotemporal head-fixed Pavlovian reversal learning task for mice. In the fixed-time schedule task with a single licking spout, mice showed anticipatory licking toward the timing of reward delivery after training, suggesting that mice learned to predict the timing of the reward, as reported in our previous study [21]. After mice learned the fixed-time schedule task, we trained them on the fixed-time reversal learning task with two drinking spouts. After every 10 trials, the rewarding side was reversed in the reversal learning situation. Mice quickly switched the anticipatory licking to the rewarding side of the spouts, even on the first or second day of training. Errors in anticipatory and consummatory licking decreased during training. This result indicates that mice can learn this head-fixed Pavlovian reversal learning task quickly. Licking responses on this head-fixed Pavlovian reversal learning task can be accurately quantified and show the hallmarks of the results of the reversal learning task in free-moving operant conditioning. After reversal of the contingency between the stimuli and outcomes, a substantial error rate was observed, which decreased as the trials progressed.
The results obtained from the head-fixed reversal learning task showed the hallmarks of the results of the traditional reversal learning tasks. After reversal of the rewarding licking spout, high error rates were observed in both anticipatory and consummatory licking. The overall error rates decreased with training. Errors of consummatory licking were stable in the fifth session of the training, but errors of anticipatory licking were decreased from the fifth to seventh session of the training, suggesting that the effect of learning is relatively slower in anticipatory licking. This head-fixed procedure allows us to analyze (1) conditioned and unconditioned responses as anticipatory and consummatory licking, (2) acquisition and maintenance of the temporal prediction of the timing of the reward, (3) acquisition and maintenance of reversal learning, (4) errors soon after reversal, and (5) maintenance of the response after reversal. In particular, reversal learning and maintenance of reversed responses are dissociated in different brain areas. Chudamasa and Robbins [16] compared the effects of excitotoxic lesions of the OFC and infralimbic cortex (ILC) in the visual discrimination reversal learning task in rats. When the stimulus–reward contingencies were reversed, more errors were observed in both the OFC and ILC lesion groups, but only the OFC lesion group was unable to suppress the previously rewarded responses, committing more “stimulus perseverative” errors. In contrast, the ILC group showed a pattern of errors that were attributable to “learning” than perseveration. Although animals need to learn from the reward omission in the trial after reversal, they should maintain the newly learned stimulus–reward contingency throughout other trials. These two learning processes are distinct, suggesting that the corresponding neurobiological mechanisms are dissociated. Therefore, our spatiotemporal Pavlovian head-fixed reversal learning task could be a useful behavioral approach to uncover the psychological and neurobiological mechanisms of behavioral flexibility.
This spatiotemporal Pavlovian head-fixed reversal learning task can be extended in the future. One direction is to manipulate the task difficulty. The difficulty of this reversal learning task can be manipulated by changing the temporal interval between reward delivery. This manipulation may be important when considering the effect of working memory load on reversal learning performance. Another direction is the implementation of outcome probabilities. If the reversal learning task is not a probabilistic design, organisms can use a win–stay and lose–shift strategy. This can be simple discrimination learning triggered by extinction or outcome omission (IF there is no reward, THEN do another response). To exclude such a possibility, making the task probabilistic may be useful for studying behavioral flexibility in detail. The third direction of the extension is to invent the head-fixed Pavlovian set-shifting task based on the current reversal learning task. By adding visual, auditory, and/or other modalities, the current head-fixed Pavlovian spatiotemporal reversal learning task can be extended to the set-shifting task as a Wisconsin Card Sorting-like task. For example, the contingency between the stimuli and outcome may be based on visual, auditory, and spatial cues. As previously demonstrated, the head-fixed Pavlovian procedure can reduce the amount of training compared to free-moving operant conditioning [21]. Comparing the results of the head-fixed Pavlovian task with those of the existing head-fixed operant task may be important for understanding the psychological and neurobiological mechanisms of behavioral flexibility.
Recent technical advances in molecular biology and machine learning have allowed us to investigate the relationship between behavior and its neurobiological correlates. The behavioral task developed in this experiment can be combined with calcium imaging techniques. Because of the advantage of the head-fixed condition, the location of the brain is constant while the mice perform this reversal learning task. The experimental setup requires minimal apparatus with no usage of external cues; thus, this setup is easy to equip under the microscope, including the two-photon microscope. In addition, the head-fixed experimental setup can use traditional or latest behavioral and physiological measurements, such as pupil size, eyelid size, facial expression, heart rate, and respiration. Using these techniques with this spatiotemporal Pavlovian head-fixed reversal learning task will pave the way for understanding behavioral flexibility.
In summary, we established a novel spatiotemporal Pavlovian head-fixed reversal learning task for mice. Licking responses on this Pavlovian head-fixed reversal learning task can be accurately quantified. Mice showed the hallmarks of the results of the reversal learning task in free-moving Pavlovian and operant conditioning. This novel head-fixed reversal learning task is a useful approach for studying the neurobiological mechanisms of behavioral flexibility as it can provide an informative framework for understanding the mechanism at the levels of genes, cells, neural circuits, behavior, and its computations.