One Self, Too Many Tasks: Bimanual Interference from a Predictive Coding Framework
No Thumbnail Available
In this work we use the predictive coding framework to examine bimanual interference in human hand movements. Based on previous experiments, we hypothesize that bimanual interference can be understood from a similar framework as binocular rivalry, as a bi-stable system created as a Bayesian optimal solution dealing with “un-ecological” conditions under strong hyper priors learnt by the brain in “ecological” conditions. Specifically, we postulate that the layer of the brain in which a single minimal-self is created to predict the correlation of information from the different modalities, includes a hyper prior that only one task goal is possible at any given time. While most tasks require many effectors to work together as one coordinated unit and not interfere with each other, like riding a bike, eating with fork and knife or driving a stick shift car, under usual “ecological” conditions these actions emerge as a solution to a single task goal and individual motion paths are undefined under the task goal. We test this hypothesis by manipulating top down task goal and bottom up visual feedback of subjects’ own hands in an immersive virtual reality environment. We instructed subjects to either follow an avatar’s motion or create a self-motion and manipulated the visual feedback to influence the predictions created by the minimal-self. Our main findings are that providing false visual feedback that is in total opposition to the minimal-self predictions lowers interference for the follow task and increased interference for the self-task. We further discovered that providing a first person view, showing the subject performing bimanual independent movements, increased the interference of the hands despite subjects’ belief that the task is easier. We explain these results using the predictive coding framework and discuss the implications regarding possible rehabilitation programs and notions regarding the relative weakness of the proprioceptive system in comparison to the visual system.
Faculteit der Sociale Wetenschappen