We are excited about the next steps, such as determining causality and the circuit mechanisms of SWR-DA coupling, as well as whether replay content of different valences is evaluated differently. :) (9/9)

Building on earlier findings (Gomperts et al., 2015), our work shows that DAergic transients known to drive learning in response to online experienced events are also present following offline replayed events. (8/9)

To characterize the SWR-DA, we fit LMMs. We found that the SWR-DA size could not be explained by the differences in track running speed (a proxy for motivation) or by the magnitude of the prediction error coding during the task. (7/9)

Averaging across all sessions and mice, we found an increase in DA peaking ~0.3 s after SWRs! The SWR-DA transient was about 24% of the size of a putative positive RPE. (6/9)

Next, during a rest period before or after the task, we found SWR-triggered DA transients (SWR-DA). Here are some examples from three different sessions and mice (MUA: black tick marks; local field potential traces: blue line; GRABDA2m fiber photometry traces: green line). (5/9)

Consistent with previous work, we found a clear suppression of the dopamine concentration ([DA]) signal following reward omission relative to the expected (medium) amount, and a clear increase in [DA] following larger-than-expected reward. (4/9)

First, we established what putative VS DA prediction errors look like in our hands by measuring the GRAB-DA signal on a probabilistic reward task. Mice ran on a linear track where rewards of different sizes were delivered at both ends. (3/9)

Leading computational theories (e.g. Mattar & Daw, 2018) suggest that replayed experiences drive learning via a teaching signal (e.g.prediction errors). To test if this may be true within the brain, we investigated the link between putative dCA1 replay events (SWRs) and VS DA teaching signals. (2/9)