Figure 3.

The consequences of sensory uncertainty. Simulation of inference on a binary hidden state x₁ (black dots) using a three-level HGF under low ($\widehat{\pi }$_u = 1000, top panel) and high ($\widehat{\pi }$_u = 10, bottom panel) sensory uncertainty. Trajectories were simulated using the same input and parameters (except $\widehat{\pi }$_u) in both cases: μ⁽⁰⁾₂ = μ⁽⁰⁾₃ = 0, σ⁽⁰⁾₂ = σ⁽⁰⁾₃ = 1, κ = 1, ω = −3, and ϑ = 0.7. Decisions were simulated using a unit-square sigmoid model with ζ = 8.