A self-initiated cue-reward learning procedure for neural recording in rodents

Single-unit recording in Pavlovian conditioning tasks requires the use of within-subject designs as well as sampling a considerable number of trials per trial type and session, which increases the total trial count. Pavlovian conditioning, on the other hand, requires a long average intertrial interval (ITI) relative to cue duration for cue-specific learning to occur. These requirements combined can make the session duration unfeasibly long.

To circumvent this issue, we developed a self-initiated variant of the Pavlovian magazine-approach procedure in rodents. Unlike the standard procedure, where the animals passively receive the trials, the self-initiated procedure grants animals agency to self-administer and self-pace trials from a predetermined, pseudorandomized list. Critically, whereas in the standard procedure the typical ITI is in the order of minutes, our procedure uses a much shorter ITI (10 s).

Despite such a short ITI, discrimination learning in the self-initiated procedure is comparable to that observed in the standard procedure with a typical ITI, and superior to that observed in the standard procedure with an equally short ITI.

The self-initiated procedure permits delivering 100 trials in a ∼1-h session, almost doubling the number of trials safely attainable over that period with the standard procedure.

The self-initiated procedure enhances the collection of neural correlates of cue-reward learning while producing good discrimination performance. Other advantages for neural recording studies include ensuring that at the start of each trial the animal is engaged, attentive and in the same location within the conditioning chamber.