Unraveling the dynamical mechanisms of motor preparation based on the heterogeneous attractor model

Different types of inhibitory interneurons play a crucial role in regulating the reaction time and accuracy of voluntary movements. To explore the neurodynamic mechanisms underlying this regulation, we construct a cortical inhibitory microcircuit model, comprising excitatory preparatory and executive nuclei as well as three inhibitory interneuron nuclei. It replicates the neural activity patterns in the motor cortex during movement preparation and execution observed in physiological experiments, as along with activity changes induced by learning. We analyze the effects of inhibitory synaptic strength and inhibitory neuron self-firing rate on voluntary movement in the model. Our findings reveal that the inhibitory synaptic strength of somatostatin (SST) and parvalbumin (PV) neurons on pyramidal cells (PC) can significantly affect reaction time. By regulating their firing rates, the ratio of the inhibitory effects of SST and PV can improve the response speed and the accuracy of motion selection. Further analysis indicates that SST-dominated selection leads to quicker but less accurate responses, whereas PV-dominated selection produces slower but more precise outcomes. Finally, using a mean-field approach, we find that the stabilization points and attraction domains of the system are different in the preparation and execution phase. Learning expands the attraction domain for choosing correctly in the preparation and execution phases and the equilibrium point for choosing failures disappears, improving the speed of reaction and the rate of correct choices. Our model gives new insights into the dynamics of inhibitory networks in voluntary movement control and learning.