Spike count data for studying dimensionality, information and learning in prefrontal cortex
Description
Learning leads to changes in population patterns of neural activity. We wanted to examine how these changes in patterns of activity affect the dimensionality of neural responses and information about choices. We carried out high channel count recordings in dorsal-lateral prefrontal cortex (dlPFC; 768 electrodes) while macaques performed a two-armed bandit reinforcement learning task (see https://doi.org/10.1523/JNEUROSCI.0631-17.2017). The high channel count recordings allowed us to study population coding while the animals learned choices between actions or objects. This dataset includes spike counts from time bins (250ms, consecutive) around the onset of the cue that instructed the animals to respond for the recorded neural population. In addition, behavioral choice data used to model behavior is included. The dataset consists of 8 recording sessions, there is one MAT file per session. Using this dataset, we have found that the dimensionality of neural population activity was higher across blocks in which animals learned the values of novel pairs of objects, with respect to the dimensionality across blocks in which they learned the values of actions. The increase in dimensionality with learning in object blocks was related to less shared information across blocks, and thus to patterns of neural activity that were less similar compared to learning in action blocks. Furthermore, these differences emerged with learning, and were not a simple function of the choice of a visual image or action. Therefore, learning the values of novel objects increases the dimensionality of neural representations in dlPFC.