Movement and non-movement related pallidal unit activity during bar press feeding behavior in the monkey

Activity was recorded from 358 neurons in the globus pallidus (GP) of monkeys (Macaca fuscata) during an operant feeding task consisting of 3 stages: (1) food or non-food presentation (1st stage); (2) bar pressing (2nd stage); and (3) food acquisition and ingestion (3rd stage). There were two kinds of neurons, one with high and the other with very low (almost silent), spontaneous firing rates. Two hundred and four neurons (57%) responded in one or more of the feeding stages. Of the 21 neurons which responded in the 1st stage, two responded selectively to food presentation, and 19 responded to both food and non-food visual presentation. One hundred and seventy-four neurons (49%) and 107 neurons (30%) responded in the 2nd and 3rd stages, respectively, and 106 (30%) of these were directly related to specific feeding motor acts such as arm extension, flexion, bar pressing, grasping, chewing etc. Both high and low firing neurons responded to motor acts with sharp or gradual onset. More than half of those that responded to arm extension showed laterality (contra or ipsi)- and function (extension or flexion)-dependent responses. The incidence of the motor related neurons was higher in the caudodorsal part of the GP. On the other hand, about one third, especially in the rostroventral part of the GP, showed dissociating responses in that they responded during bar pressing for food or during ingestion in an operant task, but not during bar pressing for non-food or during forcible ingestion. The magnitude of firing changes during arm extension and bar pressing depended on the nature of the food. Moreover, in trials using new food or false (model) food, firing changes during bar press appeared or disappeared within a few trials with no correlation to bar press movement. These data suggest heterogeneous functions within the GP; the caudodorsal part is strictly concerned with motor execution and preparation, while the rostroventral part is not related to motor function directly, but may rather be important in coupling internal, motivational information to the motor system.