Fig 5

(A) Mean prediction errors for the single factor model fitted to mean swim speed (left group of bars) and to bout measures (mean bout rate and intensity) (right group). Fitting to swim speed failed to predict the bout measures; fitting to bout data reduced prediction errors for bout measures but produced less accurate swim speed predictions. (B) Enhanced bout generator has separate inputs for bout intensity and bout initiation and the latter is subject to motor inhibition. (C) Bout rates for the single factor model fitted to bout measures. The model failed to predict that bout rates are sensitive to height in the OMR regulation procedure (left) but not the baseline flow one (right). (D) Bout rates predicted by the enhanced bout generator when fitted to bout data and constrained to reproduce the observed bout intensities. It was now possible to reproduce the bout rate patterns observed in the two experiments. (E-G) Dual factor model variants. Each variant has an enhanced bout generator with separate bout intensity and initiation inputs. Optic flow is integrated only for the intensity input. (E) Variant A: the overall optic flow is integrated. (F) Variant B: input to the flow integrator is a linear combination of the forward and backward components of optic flow. (G) Variant C: only the forward component of the optic flow participates in control of the translational OMR.