To be able to extract most of the information possible from the fish, we wanted to use a moving camera with a small field of view to track the fish across the tank at all times. To do this, we used a camera mounted to the moving motor of a 3D printer that we had dismantled.Figure 1. The rig being built and our optic flow chamber.
For the optic flow experiment, the whole setup was lighted from below by infrared light. To project stimuli on both sides of the tank, we used mirrors that reflected the visual stimulus from the projector, which was situated at the very bottom of the setup.
For the rheotaxis experiment, the tank was left open on the top and water was pumped in sideways into the tank from one side and pumped out on the other side so that we had steady flow in the tank.
We used Bonsai to present square gratings of different velocities on both sides of the tank. We had envisaged four kinds of visual stimulation: 1) control condition: where the gratings are stationary 2) coherent gratings: where the gratings on both sides move with the same speed in the same direction 3) opposite gratings: where the gratings on both sides move with the same speed in the opposite direction and 4) shear gratings: where the gratings on one side moves thrice as fast as the gratings on the other side, both in the same direction. Bonsai was also used to track the fish after the experiments.Figure 2. Example fish tracking with Bonsai
For the rheotaxis experiments, we wanted to see where in the open-lid tank the fishes preferred to swim since they now face a different water velocity profile as compared to the closed tank. We observed that with the open-lid tank, zebrafish larvae like to swim in the bottom or on the surface of the tank so as to have less water disturbance. They perform rheotaxis i.e. they swim upstream while trying to reach the bottom or the surface of the tank. To look at the wavefront pattern of the created water flow, we then did some simple particle imaging as well in the chamber. We observed that water is more stable and slow on the surface of the tank as expected.Figure 3. Imaging the flow profile in our rheotaxis chamber.
For the optic flow experiments during the coherent gratings, the larva seemed to prefer the edges of the tank when it was swimming along the direction of the optic flow. In particular, it swam along the edge closer to the side on which the optomotor stimulus was being presented. When the larva swam opposite to the direction of the moving gratings, it swam in the center of the tank.Figure 4. Edge preference during optic flow experiments.