Optimal feedback control (OFC)

Optimal feedback control is a theory of biological motor control in which the nervous system is modelled as computing a control policy that minimizes a cost combining task error and effort, using continuous state feedback rather than a pre-planned trajectory. Its signature prediction is the minimal intervention principle: the controller corrects only deviations that matter for the task and leaves task-irrelevant variability uncorrected, which accounts for the structured trial-to-trial variability seen in natural movement. Under noise and delay, an OFC also maintains an internal estimate of state (an observer) that it drives with feedback.

OFC enters BCI in three ways. As a descriptive model, it captures how a user actually controls a cursor or prosthesis under feedback, including which errors they bother to correct. As a design principle, decoders and assistive controllers can be built to assume, or to cooperate with, an optimally controlling user. And as an analysis tool, it frames the closed loop as controller-plus-plant so that user variability and correction bandwidth become interpretable rather than nuisance. The caveat is that fitting an OFC requires committing to a cost function and noise model that are rarely known, so OFC accounts are often qualitatively insightful but quantitatively underdetermined.