Mouse go/no-go: lick-based task
To determine how context affects the behavioral assessment of learning, we first trained mice on an auditory go/no-go stimulus recognition task. Mice learned to lick for a water reward provided through a lick tube after hearing a conditioned stimulus (the ‘target’ tone) and to withhold from licking after hearing an unrewarded (‘foil’) tone of a different frequency. Over the course of learning, we interleaved the reinforced context with a smaller number of trials without reinforcement by removing the licktube (‘probe context’). In the probe context, we removed the licktube for a subset of trials (<40) in order to test whether absence of reinforcement would change the self-report of the mice. First, we focused on a trial block early in learning (trial block 1500-2000) when animals were tone responsive; i.e., they licked indiscriminately to both target and foil tones in the reinforced context, but did not lick during the inter-trial interval (‘reinforced context’). Surprisingly, when we removed the licktube for the probe trials, all mice discriminated between the tones by reliably licking to targets while rarely licking to foils, exhibiting expert performance despite their variable and often poor performance in the presence of the licktube. The improvement of behavioral performance was specific to the probe context, and did not drive improvements in performance in reinforced trials immediately following the probing. Mice therefore appeared to understand the task contingencies many days before they expressed this knowledge in the presence of reinforcement. The below videos provide video evidence of this behavioral phenomenon.
Expert mouse
Performance in reinforced context
Mouse performing at expert levels in the reinforced context after prolonged training. The mouse is trained to lick in response to the target tone and withhold licking to the foil tone
Early in learning
Performance in reinforced context
Video of mouse behaving in the reinforced context (trial block 1500-2000). Note that the mouse appears to lick in response to both the target and foil tones, apparently showing a lack of discrimination.
Early in learning
Performance in probe context
Video of mouse behaving in the probe context (trial block 1500-2000, immediately after Movie S2). After removal of the licktube, the mouse correctly licks to the target tone and withhold licking to the foil tone.
mouse go/no-go: lever-based task
To test whether separating the motor action from the consummatory response would retain (or abolish) the dissociation between task acquisition and expression. In the reinforced context, head-fixed mice were trained to press a lever in response to the target tone to gain access to water reward provided through a licktube . This design added an additional key feature: the licktube was normally absent in the reinforced context and was only introduced for a short period to deliver the water and then immediately retracted. In the probe context, the lever is not advanced for the target trial thus not delivering the water reward. As a result, the sensory environment in the probe and reinforced contexts were identical during tone presentation, removing the possibility of the licktube presence in the reinforced context as an impulsive driver of licking; instead, the possibility of reinforcement was more abstract.
Early in learning
Performance in reinforced context
Video of mouse behaving in the reinforced context with lever (trial block 1500-2000). At this early time point in learning, mice pressed the lever to both the target and foil tones, apparently not discriminating between the two in the presence of reinforcement
Early in learning
Performance in probe context
Video of mouse behaving in the probe context with lever (trial block 1500-2000) immediately after the reinforced trials in Movie S4. As the video shows, the mouse correctly withholds from pressing to the foil tone but still presses vigorously in response to the target tone, revealing that the mouse has learned the task contingencies.