Coupling of antagonistic ankle muscles during co-contraction in humans

In 35 healthy human subjects coupling of EMGs recorded from the tibialis anterior (TA) and soleus (Sol) muscles during voluntary co-contraction was analysed in the time and frequency domains. Two patterns were observed in different subjects or in the same subject on different occasions. One pattern consisted of central peaks in the cumulant density function of the two signals, which was often accompanied by coherence in the 15-35 Hz frequency band. The other pattern consisted of a central trough in the cumulant density function, which was mostly accompanied by coherence around 10 Hz. When this was the case oscillations were usually observed in the cumulant density function with time lags of 100 ms. Both patterns could be observed in the same subject, but usually not at the same time. Coherence around 10 Hz associated with a central trough in the cumulant density function was less common during weak than during strong co-contraction. The central peak with coherence in the 15-35 Hz frequency band in contrast tended to be most common during weak contraction. There was a tendency for the 10-Hz coherence with central trough to occur when the contractions had been maintained for some time. Both patterns could be observed when sensory feedback in large diameter afferents was blocked by ischaemia. When a central peak with coherence in the 15-35 Hz frequency band was observed for paired TA and Sol EMG recordings (10 out of 19 subjects), a coupling in the same frequency band was also observed between the EMG activities from the two muscles and the EEG activity recorded from the leg area of the motor cortex. When the central trough and the coherence around 10 Hz was observed for the EMG recordings (8 out of 19 subjects), no significant coherence was observed between EEG and EMG in 7 of the 8 subjects. In the last subject coherence around 10 Hz was observed. It is suggested that these findings signify the existence of two different central input systems to antagonistic ankle motoneurones: one input activates one muscle while depressing the antagonist and the other coactivates antagonistic motoneurones. The data suggest that at least the latter input depends on motor cortical activity.