Figure 8.

Performance comparison among models in trajectories with perfect inertia. All models presented possess the basic skeleton from the main Kalman filter model in Experiment 1b. SWM refers to the spatial working memory model. AP-AP refers to the model in Experiment 1b. The first position in each model name refers to the variable used to address correspondences and the second refers to the variable that is combined with assigned observations to infer new posterior estimates. AP refers to adjusted prior, and P refers to unadjusted prior. Figure 9 summarizes these results in terms of performance collapsed across target load.