Towards modeling closed-loop sensorimotor integration in tactile sensing systems


The current view in neuroscience holds that the brain, together with its sensory and motor structures and the environment, form a closed-loop system – a sensorimotor loop – in which the brain receives information from the environment and converts it into a motor response while simultaneously making predictions about future sensory events based on the current action. Thus, it is crucial to study the brain in the context of the animal’s body and its ecological niche. However, conducting experiments in freely behaving animals is very challenging, and it is often impossible to measure the sensory information real animals acquire. Therefore, neuroscientists often break the loop between the sensory system and the motor system and replace the environment with controlled stimuli and motor tasks. This approach has major drawbacks, including impoverished stimuli and motor actions, forcing the brain to operate outside of its natural context, and breaking the external link between sensory and motor systems. This thesis attempts to address these limitations through simulation and modeling. First, we take a thorough look at the sensorimotor loop to examine how the flow of signals within the loop may affect the animal’s ability to predict sensory consequences of motor actions for different modalities. We introduce more precise language to characterize the conceptual differences between “active sensing” modalities, and show that tactile sensing (touch) is unique and therefore especially well-suited to examine the sensorimotor loop. We then use this insight to develop a model of the rat vibrissal (whisker) system that allows us to simulate the complete vibrissotactile sensory input during active sensing behavior. Lastly, we take the first steps to characterize the environmental constraints imposed on the tactile sensorimotor loop. We use information theoretic methods to analyze the tactile statistics of three-dimensional objects and show that there are close similarities to the statistics of sensory inputs to other modalities such as vision and audition.

Alternate Identifier
Date created
Resource type
Rights statement