Empathy and Altruism in Rats?

A recent paper in Science discussed behavioral data in rats suggestive of empathically motivated behavior. This is a potentially very important report for two major reasons. First, a deep understanding of the mental and psychological abilities of rats, and other species, is a crucial goal for comparative psychologists, evolutionary biologists and other basic scientists. Second, the autism spectrum disorders are characterized by atypical reciprocal social interactions, and difficulty with experiencing and understanding the emotions of others appear to contribute; therefore, an animal model system in which we can learn how the brain responds to and processes the emotions of others is crucial to progress in this area. For these reasons, the experiments address a very significant question.

The experiment consisted of having a rat placed in an arena (the free rat) who is able to see and interact  with a companion that is trapped in a cylindrical restrainer with a door (the trapped rat).  It was found that the free rat learned over time to free the trapped rat by intentionally opening the door.  In control experiments, rats did not open empty tubes or ones containing an inanimate object.  When given a choice between getting access to chocolate and freeing the trapped rat, they would often free the rat even before eating the chocolate, suggesting that the motivation to liberate its companion trumped even its desire for the chocolate, a potential sign of altruism.

The authors concluded that “the free rat was not simply empathically sensitive to another rat’s distress but acted intentionally to liberate a trapped conspecific.”

The media reported on the finding by declaring science has shown altruistic behavior in rats.  Some media titles include “Rats: Holiday spirit in rodent form”, “If someone calls you a rat, take it as a compliment”, “Rats kind-hearted, generous creatures”, “Rats show Empathy and Altruistic Behavior”, “Rats are as compassionate as humans” and so on.

It appears that both the press, and perhaps even the authors, interpret the findings as implying the following:

  1. The free rat has a mental state that represents the well-being of a conspecific.
  2. This representation generates a distressful response in the free rat.
  3. The free rat learns it can act in a way to relieve the distress of the caged rat by opening the door of the cage.
  4. The rat intentionally acts to relieve the caged rat from distress even when there it has nothing to gain from the action.

Dr. Daniel Povinelli, in a Nature coverage of the paper, had a different view, saying that “This work is not evidence of empathy — defined as the ability to mentally put oneself into another being’s emotional shoes.”

Though the view that rats exhibit empathic behavior may be consistent with the data, we must ask if there could be alternative, simpler explanations that do not necessarily involve invoking assumptions 1-4, above.

One possibility is that the trapped animal is generating an alarm signal, either in the form of vocalizations or pheromones, that generates stress in the free rat.  The free rat may then learn it can stop the distressing signal by opening the door (so-called negative reinforcement).  In acting in such a way, the free rat would then be relieving its own distress rather than the perceived and shared stress of a conspecific.

Is this possible?

The authors did not measure chemical signals but did measure vocalizations during their experiments and found that “significantly more alarm calls were recorded during the trapped condition (13%) than during the empty and object conditions.”

So this alternative scenario is, in principle, a possibility.  The authors dismissed this alternative explanation because the rate of alarm calls was relatively low and yet they remained open to the possibility when they concluded:

Thus, the most parsimonious interpretation of the observed helping behavior is that rats free their cage-mate in order to end distress, either their own or that of the trapped rat [...] This emotional motivation, arguably the rodent homolog of empathy, appears to drive the pro-social behavior observed in the present study.

This is a bit confusing and requires clarification.

There are at least two different interpretations of the data.  Not one.

Either the rat is freeing the companion to end its own stress (caused by an alarm signal) or it is doing it to end the perceived stress of the caged rat.   The interpretation of a pro-social, empathically motivated, altruistic behavior is only applicable to the second interpretation and not the first one.

To differentiate among these possibilities one can conduct some additional control experiments.  One could, for example, just play alarm calls that are stopped once a rat presses a lever once placed in the arena.  Or we could use chemical signaling if we learn the behavior is mediated by pheromones and identify the pheromone in question. One could have offered the free rat the option to leave the arena to a dark, quiet place, potentially ending its own distress and leaving the companion trapped.  Or the free rat could be offered the possibility of a “personal sacrifice” (such as a mild shock) to free the other rat, thus paying a price to help his companion.  These are all doable experiments that would help tease apart the different interpretations of these data.

Another potential explanation of the data is raised by video records of these experiments provided as part of the Science article shown below.

In this example, taken after the rat has learned to free its counterpart, we see the free rat going right into the restraint immediately after opening the door.  Why would the rat enter the tube if it truly felt and understood the distress the other rat experienced by being confined?

If one has ever seen rats at the pet store, you know that you will often find them snuggled up together in tubes and tight spaces because they apparently enjoy the safety and security of these types of experiences. This view was raised in an online discussion of the data:

Rats enjoy access to tight enclosures.  We routinely put plastic tubes in home cages for “environmental enrichment” and the rats are often found “snuggled” together in them, especially when resting – presumably an inherent protective response.  In fact, if you try to grab a rat in a cage with a tube, the rat will immediately go for the tube and try to stay in it.  Thus the “trapped” rat could also be seen by the “free” rat as enjoying a protected situation, and the free rat could in fact be displaying “envy” by freeing his companion so that he can enjoy the same protection and/or being motivated for social reasons to have a companion to “snuggle” with.  Indeed, the first thing the free rat did in the video after opening the enclosure was to go right into the tube with the other rat! 

So the basic question is, does the free rat want to get in, believing that his cagemate enjoys the privilege of a protected space, or does he fear for his cagemate and want to release him?   

Again, only additional experiments can address this. Resolution of these alternative views is crucial in terms of both of the prevailing motivations for conducting the study. Either rats are acting to relieve their own distress, or that of another – the difference bears strongly on our understanding of their mental abilities. In addition, if the former, but not latter, phenomena is correct, the value of studying the biology of empathy using rats is significantly challenged.

Still, we are left with a provocative phenomena —  rats freeing one another, invoking similarities with human behavior. There are plenty of other examples in nature where individuals of a species cooperate and interact in ways that could be described in terms of our own (human) mental states as altruistic or empathic behavior.  The examples range from bonobos, to bats, to even single-cell organisms, such as social amoeba (see here and here.)  The behavior is essentially the same across all these species and yet one would be hard pressed to argue that single-cell organisms have a notion of altruism and empathy in the same sense humans do.

Our brains (including those of scientists) are wired in such a way that they readily interpret the behavior of others in terms of our own mental states.  Such ability is useful in many situations, form navigating daily social interactions and even in the description of scientific data.  Care must be exercise in descriptions based on our own mental states when the outcome can have clear moral and scientific consequences.

Scientists must always keep an open mind.  But before rushing to declare that humans must seek moral guidance from rats, we should pause and try to understand exactly what the data say.  As new experiments are done and more information is available, we will surely be able to discern which of the alternative explanations is the correct one. If additional work confirms the (premature) conclusions of the authors, it will lay the ground work for developing new animal models for human psychological disorders, which will be a welcome development. For now, however, we must await that conclusive work.

J. David Jentsch and Dario Ringach

6 responses to “Empathy and Altruism in Rats?

  1. Daniel Povienlli

    Brilliant! This is about headlines, not science.

  2. I have not yet read this paper so I’m making this comment based on other media reports and this article alone.

    “Rats enjoy tight enclosures.”
    I’d actually say it depends on what circumstances. Rats are by nature agoraphobic, but also curious. Rats also do not like to be restrained. Forcing them into a restrainer tube is stressful for them, based on behavioural (struggling) and endocrine (cort rising, etc). But giving the rat the control, so they are ‘escaping’ the human predator and the bright lights and enter the tube themselves has the opposite effect, from my own experience. This is essentially what happened here. The stimulus rat was forced there and could not escape when he wants. The test rat, however would want to enter that tube because he’s in an open area and there is a stranger rat in there too.
    From the video, it could be argued that the test rat wants to investigate the stranger rat in the tube (a very common assumption in many behavioural experiments) and this particular experiment shows this is actually rewarding. However what is interesting is that the rat will let the stimulus rat out into a different compartment. I’ll need to read further to see if the test rat will get access to the tube or could make the assumption that it will have access to the stimulus rat. I’d also like to see more direct comparison of the effect of the stimulus rats distress (this may have been done; I need to read the paper). I imagine there is variability in the alarm calls made by the stimulus rat and even how distressing it feels to each rat. Do high levels of alarm calls correlate with faster release times? Are there any other measures of distress aside from alarm calls from the stimulus rat and did they measure audible or ultrasonic? Where they actually alarm calls not other types of communication?.

    “significantly more alarm calls were recorded during the trapped condition (13%) than during the empty and object conditions.” Does that refer to alarm calls made by the test rat or the stimulus rat assuming they kept a rat in the tube and put an object or had nothing in the arena? I might as well stop here and continue after I’ve read the darn thing.

    Empathy and altruism are very difficult concepts to define, therefore hard to design experiments to test them, even in humans. For example, giving money to charity may be considered altruistic, but it may also be considered selfish because you do it for the ‘warm glow’ effect or whatever. And also, what level of cognition is required for altruism? For example, why do prey animals let off distress signals when hunted? One reason (aside from the obvious scaring the predator off) is to warn other conspecifics that there is a predator so they can get away which could be considered as altruism. But it’s very unlikely those thoughts are running through the preys mind just like humans who have automatic fear reactions. Like when screaming is done out of fear, even though it might help scare of the attacker, warn others or get help but that may not be the conscious intention.

    Empathy is similar. I define it in two different ways, which are “understanding another’s emotional state” (emotional theory of mind) and “experiencing another’s emotional state”. Neither necessarily lead to altruism but there still needs to be some rudimentary understanding that the other is distressed. Which in that case, we’d also call reading a fear facial expression or hearing ‘I am scared’ as empathy in humans. The key word I guess ‘understand’ as there can be different levels of understanding here.

  3. The alarm call was once identified as coming form the caged rat but in other cases the authors could not identify the source.

    “Empathy and altruism are very difficult concepts to define, therefore hard to design experiments to test them…”

    I agree with you… but then,, if we are to write a manuscript saying that we have shown empathy and altruism in some species we ought to have these concepts clearly defined to begin with.

    I think the four points above is what the authors had in mind. At least they appear to be what the press had in mind while reporting on these findings. Yes, such interpretation is consistent with the data, but there are alternative explanations as well that could have been ruled out by other experiments as proposed above.

    I’d be curious to get your feedback again once you’ve read the paper.

  4. May I suggest a further resource to learn more about empathy and compassion.
    The Center for Building a Culture of Empathy
    The Culture of Empathy website is the largest internet portal for resources and information about the values of empathy and compassion. It contains articles, conferences, definitions, experts, history, interviews,  videos, science and much more about empathy and compassion.

    More articles on the rat empathy study are here in ther
    Empathy and Animals Magazine
    New International News Page about Empathy and Compassion with, by and for Animals.

  5. Thank you Edwin.

    Your website lists multiple definitions of empathy and, as you can probably guess, not all of them are amenable to scientific investigation. A definition of empathy like “imagining how another is thinking and feeling” is not scientifically useful, as we cannot determine what a non-verbal animal is imagining. And, of course, nobody here is denying the value of empathy or compassion. Indeed, empathy and compassion is what drives scientists working to advance medicine and human (and animal) health.

  6. a similar experiment like this was run 50 yrs ago. RM Church (1959), J. comp. physiological psychology, 52. Wonder if they were cited? Also see: “Wild Minds,” by Marc D. Hauser, NY, NY; Holt (2000). nothing new under the sun?