Perception is not a function of bottom-up sensory input alone, but involves an interaction between bottom-up signals from the sensory organs and top-down signals from higher brain regions. Top-down connection in higher cortical figures may carry signals indicating the features that higher centers "anticipate" will be present in the bottom-up sensory signals arriving from lower centers. An interaction with Bottom-up/Top-down processes is clearly at work in the higher cortical processes and models the Predictor Corrector System Mechanisms using expectations as predictors and visual observations as the corrector. The input is a noise simulated by the Gaussian Distribution while the running average is the output from the mean of the input when it is offset. When the true average approaches the real value, then your correction weight decreases.
Anastasio, Thomas J. Tutorial on Neural Systems Modeling. Sunderland: Sinauer Associates, 2010. Print.
Researchers should cite this work as follows: