A sensorimotor paradigm for Bayesian model selection

Abstract

Read online

Sensorimotor control is thought to rely on predictive internal models in order to cope efficientlywith uncertain environments. Recently, it has been shown that humans not only learn differentinternal models for different tasks, but that they also extract common structure betweentasks. This raises the question of how the motor system selects between different structures ormodels, when each model can be associated with a range of different task-specific parameters.Here we design a sensorimotor task that requires subjects to compensate visuomotor shifts in athree-dimensional virtual reality setup, where one of the dimensions can be mapped to a modelvariable and the other dimension to the parameter variable. By introducing probe trials that areneutral in the parameter dimension, we can directly test for model selection. We found thatmodel selection procedures based on Bayesian statistics provided a better explanation for subjects’choice behavior than simple non-probabilistic heuristics. Our experimental design lendsitself to the general study of model selection in a sensorimotor context as it allows to separatelyquery model and parameter variables from subjects.

Keywords

• Sensorimotor control
• structural learning
• sensorimotor integration
• bayesian model selection
• hierarchical learning