Figure 3. Optimal Tuning Curves for Downstream Functions That Have Both Peaked and Monotonic Components.

(A) Four examples of functions F→ obtained by combining a localized oscillatory function (with a Gaussian envelope) and a saturating monotonic function. Such functions represent hypothetical motor responses to stimuli at various binocular disparities. The inset depicts the correlation matrix obtained from 5000 functions.

(B) Optimal sets of two, four, and eight sensory tuning curves obtained with low noise.

(C) As in (B), but with high noise and high power cost. In all plots, the x-axis represents binocular disparity.