Fig. 6.

Influence of a secondary common synaptic input on EMG-EMG coherence. A: 2 simulated pools of MN receive 3 sources of synaptic input. The 1st is common to all motor neurons in both pools and is a cortical input, whose power spectrum is shown in B. The 2nd source simulates afferent input to each motor neuron pool. This input is assumed to be independent for the 2 motor neuron pools but common to all motor neurons within each pool. The 3rd source of input is an input independent for each neuron. The common afferent inputs and the independent inputs were simulated as colored noise processes with bandwidth 0–100 Hz. Each motor neuron was simulated as a perfect integrate-and-fire system. Each motor neuron pool innervates a different muscle, whose surface EMG signal was simulated using the volume conductor model described in Farina et al. (32). The EMG-EMG coherence between the 2 surface EMG signals of the 2 muscles was computed in 4 levels of common afferent input to 1 muscle (0, 10, 20, and 30% of total power). The total power of the 3 sources of input to each motor neuron pool was constant for all simulations and was normalized to 100%. The relative power of the cortical input was constant across all conditions and was set equal to 70% by altering the power of the independent inputs. For example, the 1st condition corresponded to 70% cortical input, 0% afferent input, and 30% independent input, whereas in the 2nd condition the powers were 70, 10, and 20%, and so on. Therefore, the strength of the common cortical input to the 2 muscles was identical for all conditions. Despite the invariance of the common cortical input, peak coherence ranged from 0.1 (30% afferent input) to 0.6 (0% afferent input) (C). Variation in the afferent input to the other muscle would have similar effects and concurrent variation of both afferent inputs would have additive effects.