Cortical activity predicts good variation in human motor output.

Human movement patterns have been shown to be particularly variable if many combinations of activity in different muscles all achieve the same task goal (i.e., are goal-equivalent). The nervous system appears to automatically vary its output among goal-equivalent combinations of muscle activity to minimize muscle fatigue or distribute tissue loading, but the neural mechanism of this "good" variation is unknown. Here we use a bimanual finger task, electroencephalography (EEG), and machine learning to determine if cortical signals can predict goal-equivalent variation in finger force output. 18 healthy participants applied left and right index finger forces to repeatedly perform a task that involved matching a total (sum of right and left) finger force. As in previous studies, we observed significantly more variability in goal-equivalent muscle activity across task repetitions compared to variability in muscle activity that would not achieve the goal: participants achieved the task in some repetitions with more right finger force and less left finger force (right > left) and in other repetitions with less right finger force and more left finger force (left > right). We found that EEG signals from the 500 milliseconds (ms) prior to each task repetition could make a significant prediction of which repetitions would have right > left and which would have left > right. We also found that cortical maps of sites contributing to the prediction contain both motor and pre-motor representation in the appropriate hemisphere. Thus, goal-equivalent variation in motor output may be implemented at a cortical level.

The motor cortical representation of a muscle is not homogeneous in brain connectivity.

Functional connectivity patterns of the motor cortical representational area of single muscles have not been extensively mapped in humans, particularly for the axial musculature. Functional connectivity may provide a neural substrate for adaptation of muscle activity in axial muscles that have both voluntary and postural functions. The purpose of this study was to combine brain stimulation and neuroimaging to both map the cortical representation of the external oblique (EO) in primary motor cortex (M1) and supplementary motor area (SMA), and to establish the resting-state functional connectivity associated with this representation. Motor-evoked potentials were elicited from the EO muscle in stimulation locations encompassing M1 and SMA. The coordinates of locations with the largest motor-evoked potentials were confirmed with task-based fMRI imaging during EO activation. The M1 and SMA components of the EO representation demonstrated significantly different resting-state functional connectivity with other brain regions: the SMA representation of the EO muscle was significantly more connected to the putamen and cerebellum, and the M1 representation of the EO muscle was significantly more connected to somatosensory cortex and the superior parietal lobule. This study confirms the representation of a human axial muscle in M1 and SMA, and demonstrates for the first time that different parts of the cortical representation of a human axial muscle have resting-state functional connectivity with distinct brain regions. Future studies can use the brain regions of interest we have identified here to test the association between resting-state functional connectivity and control of the axial muscles.