The motor system estimates uncertainty and higher order statistics for the control of grip forces.

Successful manipulation of an object requires exerting grip forces (GF) sufficient to prevent slippage. To prevent slip in more uncertain environments, GF would need to increase. Here we investigate the brain's ability to efficiently control grasp by producing GFs that correspond to confidence estimates of uncertain environments that are characterized by probability density functions of different variances and higher order moments. We found that GFs increased dramatically with the level of environmental uncertainty, and even when environmental uncertainty was held constant while higher order moments were varied, GFs changes in a way that was appropriate for kurtosis.