The neural circuits that underpin spatial navigation in mammals, birds, and reptiles are well known, but it was unclear whether similar structures also existed in fish. This knowledge would allow us to build a more cohesive picture of how spatial navigation systems evolved.
To investigate whether fish have similar spatial navigation systems to terrestrial species, researchers from the University of Oxford tested whether goldfish (Carassius auratus) could perform a task central to spatial mapping-distance estimation.
The study, led by Dr Adelaide Sibeaux, trained nine goldfish to travel a distance of 70 cm within a narrow tank covered with a repeating pattern of vertical stripes every 2 cm. On reaching the set distance, the fish were prompted by an external cue to turn around and swim back to the start position.
The researchers then tested whether the fish would swim the same distance if the external cue was removed, and the starting position was shifted. They also tested if goldfish would swim the same distance when the background pattern was manipulated.
- Most goldfish (eight out of nine) accurately estimated the set distance when the external cue to turn around was removed. On average, the fish swam for a distance of 74 cm.
- The fish continued to swim the 70cm distance even when their start position was shifted 20 or 40 cm forwards.
- When the background was altered to a pattern of vertical stripes every 1 cm (doubling the frequency of spatial information), goldfish overestimated the distance they travelled by 36%. This meant they turned before reaching the target distance (47.5 cm on average).
- The fish swam a slightly shorter distance (65 cm on average) when the background pattern was changed to one of horizontal lines. Furthermore, the fish became more inconsistent in their estimation of distance, showing twice as much variation compared with when the background had a pattern of vertical stripes.
- There was no difference in the distance swam when the pattern was changed to a 2 cm checked pattern, compared with the 2 cm vertical stripes pattern. These two patterns had the same frequency of special information. Consequently, this result indicates that it was the change in the frequency of the spatial information that affected the fishes’ distance estimation, and not just a change of pattern.
According to the researchers, the results indicate that goldfish estimate distances by visually streaming the apparent motion patterns of objects in the environment (called ‘optic flow’). Many terrestrial species are known to use optic flow to estimate distance, but goldfish appear to process the information differently.
Terrestrial animals, including humans, ants, wolf-spiders, and honey bees, estimate distances by measuring how the angle between their eye and surrounding objects changes as they travel. Goldfish, on the other hand, appear to use the number of contrast changes experienced en-route.
Lead author Dr Adelaide Sibeaux (Department of Biology, University of Oxford) said: ‘We present robust evidence that goldfish can accurately estimate distance and show that they use optic flow to do so. These results provide a compelling basis to use goldfish as a model system to investigate the evolution of the mechanisms that underpin spatial cognition in vertebrates.’
A similar study by the same research group, led by Dr Cecilia Karlsson, recently demonstrated that Picasso triggerfish (Rhinecanthus aculeatus) were also able to accurately estimate distances. The fish reproduced the learned distance of 80 cm with an average underestimate of only 4 cm .
 Karlsson, C., Willis, J., Patel, M. and de Perera, T.B., 2022. Visual odometry of Rhinecanthus aculeatus depends on the visual density of the environment. Communications biology, 5(1), pp.1-7.
Materials provided by University of Oxford. Note: Content may be edited for style and length.