Neuronal population dynamics and network organizations reflect motivational context in monkey premotor cortex during movement execution and inhibition

Actions require constant updating of response preparation, which may involve suppressing or executing the action depending on context. Many studies have shown that the dorsal premotor cortex (PMd) is a key area for controlling the level of movement preparation. However, it is still unclear how contextual information is integrated to regulate movement preparation. To address this issue, we investigated the neuronal population dynamics and network organization in different motivational contexts. We recorded neuronal activity from PMd of two monkeys (Macaca mulatta), performing a a stop-signal reaching task. The task required to respond to a Go signal as fast as possible (Go trials), and to inhibit the response if an unexpected Stop signal (Stop trial) was presented. Before each trial, a Cue signal indicated in which motivational context (Go+: higher reward for correct Go than Stop trials; Stop+: higher reward for Stop than Go trials; Neutral: same amount for both correct trials) the current trial would run. We used spike density function (SDF) to perform neuronal analysis on well-isolated single unit activity. We extracted the neuronal dynamic by mapping the neuronal population activity in a Low-Dimensional State Space using a Principal Component Analysis (PCA). We also investigated the multiscale network topology by using the node-based multifractal analysis framework (NMFA) and minimal spanning tree analysis (MST). Behavioral results show that in both animals (M1 sessions=2; M2 sessions=5) the motivational context affected motor preparation by lengthening response times (RT) and increasing the ability to inhibit in the Stop+ condition compared to the Go+ condition. Analysis of the neuronal state space showed that in Go trials of the Stop+ and Neutral conditions, the neural trajectories from the Go signal to motion generation evolved similarly, while in the Go+ condition, the trajectories followed a different evolution. The functional network analysis revealed that PMd has a more complex organization when deciding to stop than when deciding to move in the Go+ condition. However, this difference in complexity wasn’t present in the Stop+ and Neutral conditions. The MST identified the topological backbone of the network revealing a network endowed have with hubs present the Go+ condition in both trial types, absent in the other conditions with a more dispersed functional communication between neurons. These results indicate that the motivational context influences the movement preparation in PMd. At neuronal level, these influences can be detected as changes in the neuronal dynamics, as well as in the complexity and topological organization of the network.