Australasian Science: Australia's authority on science since 1938

Recognising Nemo

Credit: Stephen Coburn/adobe

Credit: Stephen Coburn/adobe

By Ulrike Siebeck & Guy Wallis

Recent studies are helping to dispel the myth that fish have a 3-second memory. In fact, some species of reef fish can even recognise human faces.

Do you remember Dory, the well-meaning surgeonfish with the 1-minute memory in the film Finding Nemo? Although the film portraits her as unusual, in many ways she plays to the common conception that fish only have a very limited memory. One joke even concludes that there’s no need to worry about the fish tank becoming boring because every time the fish turn around they think: “Oh, that’s nice, haven’t seen that before!”

But like a lot of our preconceptions, there’s precious little evidence to support this notion. Indeed, goldfish have been tested on some tasks and can recall them months and even years later.

In fact, work in our lab has revealed that some species of reef fish can recognise faces belonging to other individuals of their own and other species. How is this possible in an animal with a 3-second memory and a very small brain?

Our findings challenge assumptions about the limited capabilities of the fish brain. Beyond that, they are informing us about what is, and what is not, special about the human cerebral cortex.

So why our special interest in face recognition? Many neuroscientists argue that the human brain possesses innate, specialised face discrimination circuits that are central to basic face recognition tasks. The task is complicated by the presence of other very similar-looking objects – other faces – and made even harder by changes in hair, clothes, facial expression and viewing direction. Our capacity to do this is effectively at the limit of our visual abilities. A bank robber or a superhero only needs to wear an eye mask to obscure their identity.

How do humans do it? What features are they using? How can they recognise a face when a person turns their head or walks into a room with different illumination?

One clue lies in the fish we have been studying. If, as we have shown, fish can tell individual faces apart as easily as they can tell other objects apart, it appears that some aspects of face recognition do not require specialised cortical circuitry after all. These findings are motivating us to delve deeper into the ability of fish to learn to discriminate objects they see on a regular basis.

So how do we test face and object discrimination in fish? It all started with our discovery that some damselfish have faces decorated with complex patterns that uniquely differ between individuals, much like our own fingerprints. What makes these patterns even more interesting is the fact that we cannot see them because their colours lie in the ultraviolet range of the electromagnetic spectrum.

Earlier work in our lab had already established that, unlike humans, these fish can see light in the UV range so we selected one such species, the ambon damselfish, as our model. One of the many useful things about these fish is that they are territorial, and thus aggressive towards intruders. They also adapt quickly to being housed in fish tanks, which they rapidly adopt as their new territory.

To human eyes, ambon damselfish look very similar in colour as well as shape to another reef fish species, the lemon damselfish, which also have faces decorated with ultraviolet patterns. We wanted to know whether territorial ambon damselfish males would be able to tell the difference between an intruder of their own species and an intruder belonging to the lemon damselfish.

Since the main visual difference between the two species lies in their facial features, we displayed the intruders inside containers that manipulated the visibility of these features. One container allowed the fish to see the facial patterns while another did not as it filtered out ultraviolet light.

We found that the ambon damselfish could only tell the species apart when the facial features were visible. As long as the UV patterns were visible, the fish warded off intruders of their own species but showed relatively little interest in the other species. However, as soon as the UV signal was removed they approached both equally. From this we learned that the ambon damselfish do indeed use the ultraviolet patterns for species discrimination.

We then changed our methodology as we wanted to control our stimuli more carefully by controlling for the brightness of the stimuli or the degree or type of motion of the target fish. Instead of displaying live fish in containers, we photographed the fish and created stimuli that only displayed the faces. This was not an easy feat because we needed to capture the faces in a spectrum, including ultraviolet light, and to create stimuli that resembled the original faces.

We found that the best way to do this was to use an ultraviolet-reflecting white substance on an ultraviolet-absorbing background, and spent hours cutting out the patterns and filling the stencils with the ultraviolet-reflecting material. In the end we had stimuli that looked plain white to the human eye but displayed an ultraviolet pattern on a white background to the fish.

Having created the stimuli, all that remained was to see if the fish could discriminate the two types of face. Sounds easy, but how do you do it?

We developed a set of behavioural training methods that allow us to test visual learning and discrimination abilities in various fish species. Our basic method relied on the fish learning to associate a visual stimulus with a food reward. The stimulus would be presented to the fish on a computer screen, or as a laminated paper print-out, and the fish would swim up to the stimulus and bite it. As soon as it had done this it would receive a food reward.

When the fish had made this association and performed the task quickly and reliably, a second distracting stimulus was introduced that differed from the first. The fish’s task was then to identify the original stimulus in order to get the food reward. More distracting stimuli could be added to increase the power of the experiments, and the stimuli could be made more or less similar to test the sensitivity of the fish to slight changes.

The amazing thing is that the fish learned to associate a food reward with a visual target within just days of capture, and they learned the task so well that it became possible to deliver the food with a delay and in a different part of the aquarium to where the stimuli were displayed. This anticipatory behaviour indicates higher learning capabilities, as the stimulus is not treated as a substitute to food but as part of a procedure that in the end will lead to the supply of food.

So what did we find out? We trained several ambon damselfish to associate a food reward with a face stimulus belonging to another ambon damselfish individual, and trained several other ambon damselfish to associate a food reward with a face belonging to a lemon damselfish individual. Both were then tested against novel ambon damselfish faces. We were excited to find that all fish were able to reliably discriminate species as well as individuals from the same species on the basis of their facial patterns alone. So fish, like us, use facial features to discriminate individuals!

Our next question was which features were important for face discrimination. Could the fish only perform the task when the patterns were displayed in ultraviolet colours?

To test this, we repeated this experiment using black and white patterns instead of the ultraviolet patterns, and found that the important features were the shape of the patterns rather than their colour. From this point onwards we have been working with black and white or greyscale face images, which makes our lives a lot easier as we can now display the faces on a computer screen (which cannot produce ultraviolet colours) rather than having to print them out and go through the labour-intensive process required to produce ultraviolet stimuli.

This series of experiments, together with technologies derived from machine learning and computer vision, has set the scene for a large number of questions we have been addressing over the past few years. We developed morphing software that allows us to create smooth continua of intermediate stimuli between any two original faces. This allows us to test the limits of the discrimination ability of the fish, as we can make the distracting stimuli increasingly similar to the original stimulus.

At present we are investigating which features within the facial patterns the fish are using. Humans pay most attention to the eyes, nose and mouth when performing human face discrimination. We are using a combination of machine learning, computer vision and behavioural experiments to test if this is similar for fish. Our initial results show that the eyes are particularly important for fish as well.

Another line of our work is to test whether ambon damselfish and archerfish can discriminate human faces as well as their own faces. Our initial results show that this is indeed the case.

Overall, our work to date suggests that fish have the ability to learn visual discrimination tasks within just a few days from capture, and to a high level of accuracy (>80%). They learn to discriminate fish faces and human faces as fast and as well as they learn to discriminate objects such as squares and triangles, or differently coloured shapes. They show anticipatory behaviour, and have the capacity for flexible decision-making depending on the specific conditions of a particular experiment.

Our results demonstrate that fish have much better cognitive abilities than is generally assumed, which makes them a great model for the study of visual learning, processing and perception in general and specifically for the evolutionary origins of visual abilities that are thought to require a cortex, such as face recognition.

So next time you pass a fish in a tank, just recall that they may well be watching you.


Ulrike Siebeck is a senior research fellow at the University of Queensland’s School of Biomedical Sciences, where she heads up the Visual Neuroethology Laboratory. Guy Wallis is an academic in the Centre for Sensorimotor Behaviour in the School of Human Movement and Nutrition Sciences at the University of Queensland.