How to sense and correct for self-motion

During locomotion, the visual system is challenged by continuous movement of the visual scene that sweeps over the retina. How do animals sense and correct for visual motion during locomotion? We have identified various types of neurons that respond to visual motions, some of which might act as candidate circuit components to drive different compensatory behaviors.

HFSP Long-Term Fellow Fumi Kubo and colleagues
authored on Fri, 09 May 2014

Whenever we are walking or sitting in a moving car, we experience the image that we see moving across our visual field. This flow-like movement of the large field of view resulting from our own movements is called “optic flow”, and we are equipped with the ability to sense and compensate for the self-motion using optic flow information. The optic flow can be largely divided into two different features: rotation (moving in a circle) and translation (moving in a straight line). In fish, in particular, these two types of optic flow are tightly linked to different behaviors, and therefore need to be clearly distinguished from each other. Namely, the rotational optic flow triggers eye movements (optokinetic response, OKR) that follow the perceived motion to negate it, whereas the translational optic flow triggers a swimming response (optomotor response, OMR) that allows the fish to move in the same direction as the perceived motion when it is swept away by the current so that it can maintain its original position in space against the current. One strategy that an animal can take to distinguish between rotation and translation is to compare information coming from the right and left eyes, but how and where in the brain this computation is achieved is poorly understood.

Figure: In vivo Ca2+ imaging revealed neurons that responded to optic flow in the pretectum of zebrafish larvae. The rotating visual stimulus was presented on the arena that surrounded the zebrafish (bottom). Neurons that respond to clockwise and counter-clockwise optic flow are shown in red and blue, respectively (top).

To understand how the brain senses and distinguishes between rotational and translational optic flow, we monitored the activity of a large population of neurons in zebrafish larvae using in vivo two-photon Ca2+ imaging, in collaboration with Aristides Arrenberg and Wolfgang Driever at the University of Freiburg. We focused on a visual area, the pretectum, which receives direct inputs from the retina. Concomitant with the Ca2+ imaging, we presented a set of visual stimuli that make fish perceive rotational or translational optic flow. Our systematic “whole-pretectum” Ca2+imaging revealed thousands of neurons that responded to various stimulus combinations. We therefore classified neurons into functional types based on their response patterns and identified 11 neuron types that were frequently represented in the pretectum. Four of them responded to optic flow moving in one direction (inward or outward) in one eye (right eye or left eye), as reported in previous studies. However, the 7 remaining classes showed much more complex response patterns that have not been described in previous studies. Interestingly, these 7 classes include neuron types that respond to translation but not to rotation. This suggests that these neuron types can selectively respond to rotation, and therefore could act to distinguish between rotation and translation.

Our findings not only identified new neuron types that process optic flow, but also allowed us to make predictions about how identified neuron types can interact with each other to compute binocular optic flow. The predicted circuit suggests connections between different neuron types with excitatory and inhibitory synapses. Future work will be needed to uncover the connectivity of the proposed circuits.

Reference

Functional architecture of an optic flow-responsive area that drives horizontal eye movements in zebrafish. Fumi Kubo, Bastian Hablitzel, Marco Dal Maschio, Wolfgang Driever, Herwig Baier and Aristides B Arrenberg. Neuron, 2014 Mar 19; 81(6):1344-59.

Pubmed link

Max Planck Institute Press Release