Friday Weird Science: Does your mouse prefer Renoir?

Jun 28 2013 Published by under Behavioral Neuro, Friday Weird Science

…well have you ever ASKED it? Maybe it is more of a Picasso type of mouse.

This seems like a completely absurd question. After all, art is thought to be one of the highest of the human endeavors, making things of beauty may be one of the things that makes us the species that we are. But do other animals have artistic sensibilities? Have we ever asked them?

As it turns out, probably mice don’t care if they’re viewing Renoir or Picasso. But even if they don’t care, they can tell the difference. And that is much more interesting than you might suppose.

451px-Renoir15
(I have seen a lot of Renoir, and I must conclude that, aside from similarities in soft styles and lighting, what Renoir really liked was painting naked women. Source)

Watanabe, S. “Preference for and Discrimination of Paintings by Mice” PLoS ONE, 2013. 10.1371/journal.pone.0065335

The author of this paper wanted to see if mice could tell the difference between different works of art. This isn’t as far fetched as you might believe. Other types of animals “prefer” certain artistic themes, for example, capuchin monkeys like regular patterns, and some birds prefer the hard lines of Cubism to Impressionism. Is this a preference for color? Pattern? Regularity? A rejection of romanticism? That we can’t say, and you can’t really ask the birds.

But while we can’t say what the animals think about art, we can ask whether they can tell the difference. In this case, the author used several artists, and a behavioral paradigm called conditioned place preference.

Conditioned place preference is a behavioral test that is usually used in mice or rats to assess the rewarding properties of things like drugs of abuse. You take two chambers, and distinguish them in some way. Maybe one smells like lemon and the other mint, or one has stripes and the other has polka dots, or there is a difference in the floor. Then, you pair one side (say, the side that smells like lemon) with a drug, and the other side with saline. After a few days of injecting a mouse with drug, and dropping him in lemon, and saline, and then dropping him in mint, you put the mouse in between the two chambers, and see where he spends most of his time. If he found the experience of the drug rewarding, he will spend more time in the lemon scented chamber.

But of course, the key aspect of this is that the mouse can tell the difference between the two chambers. It can tell lemon from mint, or stripes from dots. But can a mouse tell Kandinsky from Mondrian?

The author of this study tested four different painters.

Kandinsky
Wasilly_Kandinsky,_1912,_Landscape_With_Two_Poplars,_78.8_x_100.4_cm,_The_Art_Institute_of_Chicago
(Source)

Mondrian
Mondrian_CompRYB
(Source)

Renoir
800px-Pierre-Auguste_Renoir,_Le_Moulin_de_la_Galette
(Source)

and Picasso
Picasso_three_musicians_moma_2006
(Source)

When the author simply put the mice in between Kandinsky and Mondrian, the mice showed no preference. Perhaps they were not moved by modernism. But when the animals got one artist (not a particular painting, a particular ARTIST, the painting themselves were switched out daily), paired with morphine, they were able to tell the difference.

Screen shot 2013-06-26 at 10.58.27 AM
(Figure 1C from the paper)

You can see above the time spent in the drug paired compartment (“conditioning”) compared to baseline. The animals spent more time in the drug paired compartment, though the differences are small (I tried to find if the error bars here are SD or SEM and could not. I assume they are SD as the data probably wouldn’t be significant otherwise). It appears that, when paired with morphine, the animals could tell the difference between the two artists, associating one with the morphine chamber and one with the saline chamber.

The author also had another set of mice that were trained to lever press for food in the presence of either a Picasso or a Renoir. Though the results don’t look particularly striking, they are apparently significant, that mice can tell to lever press for a Renoir and not for a Picasso.

Does this mean that mice “prefer” impressionism? Probably not. But what it could mean (though the numbers for the lever pressing were low and I could wish for more significance) is that mice can TELL THE DIFFERENCE. These were mixed paintings, more than one for each artist, presented on different days (which may account for the close error bars, asking a mouse to generalize a response to all Picassos or all Kadinskys is a tough call). This means that they may have much more visual acuity than we previously thought, to be able to tell the difference between soft and harsh lighting or soft vs hard lines.

So what’s the point, other than to give the mice a sense of art appreciation? Much as I’d love to see a mouse tour of the Louvre, this mostly means that mice could be much more sensitive to visual cues than we have previously thought. This doesn’t necessarily make them “smarter”, or give them a real appreciation for modernism, but it could make a difference in how people test mouse behavior in future. We may have to give them more credit than for just stripes vs dots.

Next, I want to see whether they prefer Brahams to Beethoven.

7 responses so far

  • dr24hours says:

    Wait: they paired Renoir with MORPHINE? Of course they preferred it! Here's my hypothesis: mice like anything+morphine more than anything else - morphine. WHERE'S MY PAPER?

  • Iddo says:

    Renoir painted naked women, and they researchers paired Renoir with morphine. So mice like nudes (why human nudes is beyond me) & drugs. All that is missing from this experiment is a Pink Floyd soundtrack.

  • Ilovepigenetics says:

    I was going to write something explaining this, but I am having a hard time really understanding what was done. The methods are LACKING to a non-behavioral scientist. Did they properly switch the paintings? They did inject half of the animals in the presence of Mondrian and half in the presence of Kandinsky (or half in the presence of Renoir or Picasso) during the conditioning stage. During the actual experiment, they gave the animals an injection and then waited to see which chamber they spent most of their time in. Watanabe claims that the animals spent more time in the presence of paintings by the artists that are associated with morphine. Although they wiped down the floor with EtOH, how do they know there were not other clues that would tell the animals in which compartment they received the injections? Is the data really significant?

    • scicurious says:

      Good questions. I myself am not sure whether the data is significant, though of course he says it is and I should trust that...it looks really...well. If those are SDs and not SEMs I am more inclined to believe it, but those aren't listed (which is a major issue right there).

      From what I can gather from the methods (though you are right, terribly written), they switched out the paintings daily, so if a mouse was trained on morphine or saline in the presence of picasso or renoir, he got a different picasso or renoir every day, including on the day of testing. During the TEST, there is no injection given, the animal is asked to associate the "desire" for reward (the feeling of being high with morphine) with a given chamber. I've run this test myself a bunch, usually we get MUCH bigger differences than that for a drug like morphine, but then, we aren't switching out the cues every day.

      The usual procedure involves a VERY thorough wipedown with 70% EtOH. As far as other cues, in theory you take care of that with the initial test, which is done BEFORE any drug is given or any training is given, you put them in and see if they spend roughly equal amounts of time exploring each compartment. This shows they have no bias for left vs right (or no inborn bias for picasso), and so you know differences will be due to the drug. As far as other potential cues, I assume they worked hard not to give them any (same floor, same smell, only the paintings differ). Often this test is done with only one cue (I used stripes vs dots), and they have no problem differentiating.

  • Pascale says:

    This sort of study makes our elected representatives really, really angry.

  • [...] Do the mice prefer Manet or Mondrian? We may never know, but we do know that mice can tell one art style from another. [...]

  • [...] Everyone’s a critic. Mice are able to discriminate between artists style. [...]

Leave a Reply to Pascale Cancel reply