Non-necessary neural activity in the primate cortex

Abstract

When neuroscientists record neural activity from the brain, they often conclude that neural responses observed during task performance are indicative of the functional role of the brain area(s) studied. In humans and nonhuman primates, it is often hard to combine recordings and causal techniques within the same experiment, leaving the possibility that the activity recorded may be epiphenomenal rather than reflecting a specific functional role. Currently, the prevalence of epiphenomenal neural activity in the cortex is unknown. To estimate the extent of such activity in primates, we chronically recorded neural activity in the prefrontal cortex of the same monkeys using the same neural implants during the performance of four different cognitive tasks. The four tasks were carefully selected such that only one of them causally depends on the brain area recorded, as demonstrated by previous double dissociation studies. Using the four most common single neuron analyses methods in the field, we found that the prevalence and strength of neural correlates were just as high across all four tasks, including for the three tasks that do not depend on this brain area. These results suggest that the probability of observing epiphenomenal activity in primate cortex is high, which can mislead investigators relying on neural recording or imaging to map brain function. One-Sentence SummaryTremblay, Testard and colleagues show that inferring a brain areas function from neural recordings alone could be misleading.