Dual credit assignment processes underlie dopamine signals in a complex spatial environment. Krausz et al
Description
This dataset was collected to test how expectations of future reward are updated through experience. By measuring nucleus accumbens dopamine (dLight fiber photometry) while rats foraged for reward in a complex decision-making task, we found that dopamine scaled with the values of maze locations. We then identified two distinct algorithms used to update these place values: progressive propagation over space, and inference using maze knowledge. We are sharing four tabular datasets from our photometry recordings. The first, “photLevelDf,” is a dataframe where each row consists of a photometry measurement sampled at 250Hz (processed using isosbestic correction and z-scoring) along with the associated relevant behavioral and task variables (i.e., the rat’s location, speed, etc.). “hexLevelDf” is a dataframe where each row corresponds to the photometry signal along with relevant task variables binned by the occupancy of each distinct maze location (“hex”) on each run through the maze. “sessionTable” lists the experimental subject and recording information associated with each recording session, whose identifiers are identical to those in the other shared dataframes. Finally, “trialLevelDf” contains a table where each row consists of aggregated information about each individual trial (i.e., which port the rat entered, whether it was leftward or rightward choice from the prior port, whether the rat received reward, etc.) to help with port-level choice analyses. Further information about each dataset, along with code to help analyze the data, can be found at https://github.com/Berke-lab/DA_maze