Figure 5.

Control architecture of the Bioinspired BMI (B-BMI). The SNN consists of two MSNs, which receive simulated excitatory synaptic inputs from the extracellularly recorded motor cortex units. The MSNs are reciprocally connected to each other through inhibitory synapses. The prosthetic control module monitors the spiking activity of the MSNs and determines the one with the highest firing rate as the winning neuron. The action corresponding to the winning neuron is applied by the digital base servomotor (Joint 1) of the robotic arm in one-dimensional space. A global reward signal modulates the weights of the excitatory synapses between the motor cortex and medium spiny neurons through a reward-modulated spike-timing-dependent plasticity rule.