Neither me nor thee: the fish in the mirror

If you want to get a Siamese fighting fish (Betta splendans) like this, show it another fish...

...or its own reflection.

Famously, very few animals understand reflections as representing themselves. There are veritable reams of papers written about the “mirror test” and how much it suggests about animal self-awareness, and therefore, animal consciousness.

Whether you agree with great apes or dolphins showing self-recognition in the mirror test, pretty much everyone agrees that fishes do not. The example of the betta doing a threat display to a mirror goes back in the scientific literature over 75 years.

Desjardins and Fernald looked at this question in an African cichlid, Astatotilapia burtoni, rather than a betta. I suspect that the reason for using this particular fish is that they know a little more about the cichlid’s genetics than the betta’s.

The experiment is dead simple. Take fish, put them in two adjacent tanks with a divider, and pull back the divider to reveal an empty tank, a mirror, or another animal. After 20 minutes of video recording the behaviour, the fish are sacrificed. Blood samples are taken to measure hormone levels (as things like testosterone are associated with aggression), and brain tissue is analyzed for gene expression.

The gene expression in the brain is different depending on whether the fish sees a reflection or another animal.

That caught me, and the authors, flat-footed. Are all those decades of observations made by who know how many fish owners wrong? What is going on?

The genes they examined were c-fos and egr-1, both of which can be “switched on” rapidly. c-fos is more interesting to me, because lots of c-fos is thought to be correlated of neurons generating lots of action potentials. The pattern was the same in three of the four brain regions they examined: both genes were expressed way more if the animal saw a mirror than of the fish saw another fish or a control. In the fourth area, both reflection and foe caused higher levels of gene expression than an empty tank.

The authors found no difference in behaviour or hormone levels depending on whether fish saw a mirror or another animal. This is good, because it replicates the older results: there is no behavioural difference between mirror and opponent.

But the gene expression indicates that something different is going on in these two circumstances. It’s not at all clear what that might be yet. Desjardins and Fernald suggest the mirror might be inducing more “fear” than an opponent. This word has a little more cognitive baggage than I would like, personally, but they provide some supporting references that suggest egr-1 is induced during stress responses.

The follow-up experiments that I would want to do would be to use a video system to start messing with the reflection. Delay the image, so that the movement of the reflection doesn’t perfectly track the fish. Flip the image horizontally or vertically. Stickleback have been shown to respond to video images many times (e.g., Rowland 1995). If these fish have better visual discrimination than stickleback, maybe HDTV will work: earlier this year, Pronk and colleagues showed octopuses would respond to HD TV images. I think these would work.

While I am a great believer in the power of a behavioural analysis, this paper points out that there may be more subtle things going on in the brain. One doesn’t necessarily mirror the other.



Reference

Desjardins J, Fernald R. 2010. What do fish make of mirror images? Biology Letters 6(6): 744-747. DOI: 10.1098/rsbl.2010.0247

Bradbury J. 2005. Social Opportunity Produces Brain Changes in Fish PLoS Biology 3(11) DOI: 10.1371/journal.pbio.0030390

Pronk R, Wilson D, & Harcourt R (2010). Video playback demonstrates episodic personality in the gloomy octopus Journal of Experimental Biology, 213 (7), 1035-1041 DOI: 10.1242/jeb.040675

Rowland W (1995). Video playback experiments on stickleback mate choice: female motivation and attentiveness to male colour cues Animal Behaviour, 49 (6), 1559-1567 DOI: 10.1016/0003-3472(95)90077-2

Betta picture by calwhiz on Flickr; used under a Creative Commons license. Astatotilapia burtoni picture from Bradbury.