We recorded the neural activity using two main paradigms with chronic arrays, the non-match-to-goal (NMTG) task and the object-in-place (OIP) scenes learning task, recording from the Frontal Pole.
We trained the monkeys to perform a social variant of the non-match-to-goal task (NMTG), designed to study the ability of monkeys to monitor human agent choices and computer choices generated by a cursor. Under experimental control, the monkey and a human agent alternated each other’s roles as actor and observer. As actors, the monkeys had to choose between two peripheral targets (PT) displayed on the two right and left sides of the touchscreen. They had to follow a non-match-to-goal rule, discarding the target chosen in the previous trial for the alternative one. As an observer, they had to observe the human agent or the computer agent in a controlled condition who followed the same rule, and to monitor his/its choice to perform its own future choice correctly. The computer controlled cursor acted similarly to the human agent.
With this paradigm, we found that both during the monkeys' trials and the observation of the human agent FPC neurons represented the spatial goal (right/left) around the end of the movement that is when the choice is made explicit and the goal is accomplished.
By comparing human and computer trials we found that immediately after the movement separated populations of FPC cells were dedicated to the representation of the action goals in the human (social) and the computer (nonsocial) indicating an important difference between the interaction with living and non-living agents. Importantly, the fact that this prefrontal area encoded differently human and computer spatial goals shed light on the role of the area in the representation of other’s goals once these are accomplished. These results were presented in a symposium that organized by Genovesio at the 2020 FENS and by Genovesio at International Symposium on Biology of decision making (SBDM). Nougaret presented the results in a symposium at the SINS meeting (
http://www.sins.it/medias/943-final-program.pdf(si apre in una nuova finestra))
Second, we analyzed the sessions from the object in place (OIP) paradigm in which the monkeys learned the task in two modalities: alone or through observation. We found that during the delay period the neuronal activity underwent the largest changes between the first and the second presentation of a given problem. This result can support the hypothesis that FPC is involved in one trial learning. This result shows for the first time a neural substrate of the one-trial learning, a process in which area 10 appears to play a pivotal role, as reported by lesion studies. We found also a learning effect at the level of the direction signal, the direction tuning increased with learning moving between the first presentation of the scene(first run) to the last (last run). These results were presented by Nougaret at the Neurofrance conference and at SFECA conference.
Besides the recording and the electrophysiological analysis, we published five articles. In one article in Behavioral Brain Research (Ferrucci et a., 2019) we compared one-trial learning in the OIP task within different reward contexts. Our results showed that the amount of upcoming reward could speed the learning.
In another second article, we provided evidence that monkeys can show one-trial learning in the OIP paradigm both by observing a human agent and a ghost agent represented by a cursor controlled by a computer.
These two papers paved the way to address the question from a neurophysiological point of view in the Frontal Pole.
