Human automatic postural responses: responses to horizontal perturbations of stance in multiple directions

The effect of the direction of unexpected horizontal perturbations of stance on the organization of automatic postural responses was studied in human subjects. We recorded EMG activity from eight proximal and distal muscles acting on joints of the legs and hip known to be involved in postural corrections, while subjects stood on an hydraulic platform. Postural responses to horizontal motion of the platform in 16 different directions were recorded. The amplitude of the EMG responses of each muscle studied varied continuously as perturbation direction was changed. The directions for which an individual muscle showed measurable EMG activity were termed the muscle's "angular range of activation". There were several differences in the response characteristics of the proximo-axial muscles as opposed to the distal ones. Angular ranges of activity of the distal muscles were unipolar and encompassed a range of less than 180 degrees. These muscles responded with relatively constant onset latencies when they were active. Proximo-axial muscles, acting on the upper leg and hip showed larger angular ranges of activation with bimodal amplitude distributions and/or onset latency shifts as perturbation direction changed. While there were indications of constant temporal relationships between muscles involved in responses to perturbations around the sagittal plane, the onset latency relationships for other directions and the response amplitude relationships for all directions varied continuously as perturbation direction was changed. Responses were discrete in that for any particular perturbation direction there appeared to be a single unique response. Thus, while the present results do not refute the hypothesis that automatic postural responses may be composed of mixtures of a few elemental synergies, they suggest that composition of postural responses is a complex process that includes perturbation direction as a continuous variable.