A Better Way to Study Social Behavior in Autism Mouse Models

Why haven’t They Done That Yet?

I was thinking over the past few weeks about research into autism and what steps remain to develop a valid behavioral model of autism in mice. This is a problem since there is no “autistic mouse” that can serve as a baseline for comparison, so all of the interpretations of behavioral oddities in mice as “showing autistic like reductions in social behavior” is based on conjecture at best. As always, I have been pondering why a well controlled, clear, and surprisingly simple behavioral study that was reported widely in the media (even featured on an episode of NOVA) in 2011 has been passed over, forgotten about, and then never applied toward studies evaluating mouse models of disease.

The paper containing the behavioral task I am describing in this post is “Empathy and pro-social behavior in rats” by Inbal Ben-Ami Bartal, Jean Docety, and Peggy Mason that appeared in Science Magazine in 2011.

What Should be Studied and Why?

The reason I am interested in this question is that I have always been hesitant to accept the current measures of social behavior in mice, namely the three-chambered social novelty/preference test and social dyad experiments (see Here and Here for specific critiques of the current methodology). My belief is that if a given strain of mice are not ostracized or preferentially beat up by the other mice they are grouped with, then gross-level social behaviors cannot be all that disrupted, just different than what was expected. Also, anecdotally, even the nonsocial mice always seem to sleep together cuddled up in a corner, suggesting some social behaviors are intact. As such, we have to look at a much closer level at what may be abnormal in these animal’s social behaviors. Sated another way, I think looking at “social behavior” is like studying “learning and memory”, there are so many categories within the term that just saying “social” or “memory” become meaningless.

So to break social behaviors down, I want to know if the drive to engage I social behavior is different across models and if that drive can be used as a proxy measure for social behavior. My thinking is thus: why do we not evaluate the valance of social stimulus to mice as a measure of social behaviors? At least it would allow us to get one step closer to seeing if different mouse strains at least attribute different levels of value to social or nonsocial stimulus.

My reason for wondering about all of this is that it is not only individuals with autism that show abnormal social behaviors. There is a wide range of social deficits reported in genetic disorders, and not even all individuals with autism present clinically with social avoidance behaviors. Individuals with Angelman syndrome and William syndrome are typically hyper social (i.e., seek out social contact more than a typical child), as are a subset of individuals with autism. On the other hand, their is a population of individuals with autism that appear to lack social approach behavior, and the same goes for individuals with fragile X syndrome.

A Proposed Approach

Instead of looking at social approach behaviors themselves, I suggest quantifying the valence or affective reward value each mouse places on social interaction may actually provide better data. Specifically, I think it would be quite informative to evaluate whether different mouse strains in themselves show differing levels of behaviors congruent with empathy. In this case, how quickly do they rescue either a cage mate or stranger mouse from a plexiglass enclosure, and whether they prefer a food reward over “rescuing” the other mouse. The reason this task evaluates empathy or valence of social stimuli is that freeing the other mouse takes some effort; therefore, a decision has to be made to expend the energy and time to free the other mouse rather than on other tasks.

This experiment, to the best of my knowledge, has not been performed in mice. It has, however, been demonstrated quite convincingly in rats. Bartal and colleagues showed that rats not only would actively work to free other rats from restraint, but they would share food with the newly freed rat (I other words, they did not eat all the food first, they only ate roughly half).

What made this experiment great was the fact that there was no aversive component. What I mean by that is that in previous experiments probing empathy in rodents, the tasks involved inducing fear to a rat or mouse while others watch. The experimenter then evaluates “fear-like” behaviors in those that watched the fear conditioning. An experiment requiring a rat to free another rat can more easily be interpreted as one rat considering the other and acting directly on the other rat’s behalf without any self interest.

Short video of rats performing this task. Instructions on how to make the restraint box are found (Here)

Potential Interpretation

So assuming these experiment work, which is not trivial since I have not piloted any of this, the patterns of results would lend themselves to easy interpretation. If a mouse ignores the other mouse entirely, even in the absence of any other distracting rewards it would suggest that social stimuli show little to no valence to the mouse-or it has no interest in social stimulus. An intermediate phenotype would be to rescue the trapped mouse, but only after consuming some of the food or exploring the space thoroughly first. This, based on the rat study, would encompass “normal” social behavior. On the other end of the spectrum, if a mouse frantically tries to free the other mouse at the cost of consuming reward or exploring the environment that would suggest that social stimuli are highly valanced-or that the mouse is hyper social.

I think that data like these, particularly if a spectrum of behaviors can be obtained among strains, would grant researchers a stable baseline against which to compare the social behaviors of mice. Similar experiments then could be performed using different mouse disease models to see where they fit within the spectrum of behaviors. Additionally, crossing the mutant mice onto different backgrounds, and thus potentially modulating baseline social behavior may serve to further elucidate the role of given mutations or diseases on social behaviors, or at least social behaviors that reflect empathy.

As always, I am also looking at the translational implications of data collected using tasks such as these. Developing a collection of social phenotypes across the associated mouse disease models would not only be useful to define social behavior itself, but also such a matrix of social behaviors would serve useful as outcome measures for treatment or intervention studies.


I would love to hear your thoughts on this!

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s