Neural Correlates of Internal States that Capture Movement Variability

The brain lacks the ability to perfectly replicate movements. In particular, if specific movements are cued sequentially, how you perform on past trials may influence how you move on current and future trials. Past trial outcomes may, for example, modulate motivation or attention which can play a significant role in how one moves, yet variability due to such internal factors are often ignored when modeling the sensorimotor control system. In this study, we wish to extract such internal factors by modeling variability in movements during a motor task riddled with unpredictable perturbations. Four subjects performed the task, and we simultaneously obtained Local Field Potential (LFP) activity from nonmotor brain regions via depth electrodes implanted for clinical purposes. We first show that motor behavior depends not only on current trial conditions, but also on internal state variables that accumulate past outcomes involving movement performance, movement speed, and whether or not perturbations have occurred. We further show that these internal states modulate with beta band activity in specific brain regions on a trial-by-trial basis. These results suggest a nontraditional role of nonmotor brain regions and prompt a need for further exploration.