Interneurons are believed to realize critical gating functions in cortical circuits, but it has been difficult to ascertain the underlying type of interneuron and the content of gated information in primate cortex. Here, we address these questions by characterizing subclasses of interneurons in primate prefrontal and anterior cingulate cortex while monkeys engaged in attention demanding reversal learning. We find that subclasses of narrow spiking neurons exert a net suppressive influence on the local circuits indicating they are inhibitory. These putative interneurons encoded area-specific information showing in prefrontal cortex stronger encoding of choice probabilities, and in anterior cingulate cortex stronger encoding of reward prediction errors. These functional correlations were evident not in all putative interneurons but in one of three sub-classes of narrow spiking neuron. This same putative interneuron subclass also gamma – synchronized (35-45 Hz) while encoding choice probabilities in prefrontal cortex, and reward prediction errors in anterior cingulate cortex. These results suggest that a particular interneuron subtype forms networks in LPFC and in ACC that synchronize similarly but nevertheless realize a different area specific computation. In the reversal learning task, these interneuron-specific computations were (i) the gating of values into choice probabilities in LPFC and (ii) the gating of chosen values and reward into a prediction error in ACC. This finding implies that the same type of interneuron plays an important role for controlling local area transformations during learning in different brain areas of the nonhuman primate cortex.
bioRxiv Subject Collection: Neuroscience