Optimality of the vestibulo-ocular reflex

• Applicants: Professor Dr.-Ing. Stefan Glasauer; Professor Dr. Erich Schneider
• Subject Area: Cognitive, Systems and Behavioural Neurobiology ; Sensory and Behavioural Biology ; Medical Physics, Biomedical Technology
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 545300591

Continuous and accurate visual perception during self-movements requires the minimization of image drift on the retina through gaze stabilization. This is achieved by compensatory eye and/or head movements generated by the concerted action of multisensory signals and intrinsic motor efferent copies. Gaze stabilization is a ubiquitous and well-developed motor behavior that is highly conserved evolutionarily and occurs in all vertebrates. The vestibulo-ocular reflex (VOR) consists of eye movements that are triggered by a short-latency neural circuit. For perfect gaze stabilization, the compensatory eye movement should perfectly cancel out the head rotation in terms of amplitude and dynamics. However, in many vertebrate species, but also in humans in old age or after inflammation of the vestibular nerve, the eye movement is significantly less than required for perfect stabilization. However, the reason for this apparent suboptimality is unclear. We suspect that the sensory, motor or neuronal elements responsible for gaze stabilization have reached a performance limit, which manifests itself as inaccuracy in the execution of eye movements. Due to these signal-dependent inaccuracies or noise, a smaller compensatory eye movement may be more favorable for gaze stabilization. We will test this hypothesis using a combination of theoretical and experimental methods in amphibians and healthy humans. The amphibians we use represent a perfect model system to investigate open questions in sensory-motor control, as the animal model is relatively simple and there is a large body of knowledge about VOR circuitry in these animals. Based on the assumption that the VOR is an optimized sensorimotor response, we use optimal control theory as a normative model to derive conditions under which seemingly suboptimal gaze stabilization is the best choice for optimal reduction of retinal image drift, or alternatively optimal retinal image fixation. We then test these predictions experimentally in humans by comparing the strength of the VOR with the variability of eye movement, and in amphibians by manipulating the signal noise in the different neural elements of sensorimotor processing. Comparing experimental data with predictions of optimal control and a neural network model of VOR circuitry will help answer the question of why the eye movement is smaller than required for perfect gaze stabilization. Our project will therefore help to understand why such seemingly suboptimal eye movements can also occur in humans.

• DFG Programme: Research Grants
• International Connection: France
• Partner Organisation: Agence Nationale de la Recherche / The French National Research Agency
• Cooperation Partner: Dr. Francois Lambert