I'm not familiar with any studies where multiple animals are introduced to the same NOR task simultaneously. That said, your question stimulates a multitude of questions that might be interesting to ask, depending on the research question at hand.

Nevertheless, I would suggest avoiding such a scenario due to a multitude of potential confounding variables, most notably the influence of dominance hierarchies. Given that the NOR is predicated on the notion that in order to identify whether an animal truly recognizes new vs old objects, one must remove any potential distractors. It is for this very reason that habituation periods are necessary, to reduce the inherent distractions of a new environment. In that regard, if we were to consider the confound of dominance statuses alone, it is plausible that a more subordinate animal would avoid interacting with a novel object because it is more focused on the dominant animals' behaviors. It follows then that conclusions drawn would be depend more on the behaviors of the dominant animal, rather than on mechanisms of memory or recall. Another issue would be physical space surrounding the objects. Can all of your animals even interact with the objects equally across the duration of the test? Can you control for that?

With all of this said, the more important question is, why would you like to put more than one animal in at a time? If it's to reduce experimenter burden, then I would strongly advise against it unless you plan to control for the myriad of confounds. If it's to potentiate exploration/reduce stress and anxiety, I advise introducing longer habituation periods. But if you'd like to investigate the interactions between memory and dominance, crowding, or social buffering, then yours would be a paper I'd be very interested in reading. I see many challenges though.

Dear Alex,

Thanks for your thoughts. The initial idea came to me when I realized that habituation is so different if littermates are included. The animals behave completely different in the arena afterwards and they are habituated to the environment much faster. Event this seems very interesting to me.

I am aware of the potential problems and I believe the whole thing should be carefully examined in detail prior to any conclusions. However, the possibility of various software packages to simultaneously track myriad of behavioral phenomena could allow us to control at least for some of the confounders and maybe discover unexpected ones also. Based on the data the experiments could maybe be done with groups of animals forming different social structures and this could provide even more data. The reason why this seems interesting to me is that even though single animal experiments allow for controlled conditions, the control itself induces certain unrealistic outcomes (eg. in papers where certain animal models are more eg. stressed, and seem to also be cognitively impaired.. are we really looking at cognitive impairment or behavioral change due to different stress response. In this case, would the results for (eg. NOR) look different if the habituation period for all animals was much longer (so all the stressed animals from the experimental group are as calm as the controls in the NOR test)?). I think the amount of confounding factors in currently used controlled experiments is also very often underestimated, and that great amount of behavioral results are biased because of complex interactions with all the behavioral outcomes different tests measure.

Group testing if done the right way, as you describe it, would allow for a new set of data that could help us sublimate some of the prominent behavioral clusters. Of course at the cost of inducing new bias as you point out (maybe less anxiety, maybe more social influence). Still, it seems to me this would be interesting to explore. And very demanding indeed :)

You raise many great points. A major problem with behavioral neuroscience is that the rigor of the "behavior" component often gets sacrificed in favor of more impactful or technologically advanced "neuroscience". There are far too many published papers where behavior is used either as a treatment or dependent variable, yet the conclusions published are weak given poor control of the behavior or consideration of its ethological validity.

I will add that the advancements in tracking software are tremendous, but that may not defray the risks to any conclusions alone, just give more data to make sense of. While you would have more information about where each individual spends their time, I fear that if it were me, I'd struggle to know whether animal X has poor recognition (and thus spends little time with the novel object) or not. Still, it seems you're accounting for that, which is great.

As far as the confound of stress in preexisting studies where NOR is run with an individual: You're absolutely right that there may be stress introduced, but I would venture to say that with sufficient habituation, that stress is at least reduced and, moreover, that it's theoretically reduced for all individuals. More importantly, by approaching it this way you're removing that dynamic of social stress, which could compound with the inherent stress of being placed in the NOR arena. Once all animals are put in the arena together, you reintroduce that social stress which could interact. I say this as a stress researcher though, where many of my experimental treatments are stressful. So for me, deselecting stressful experience is paramount given that I want to hone in on particular stress epochs.

Obviously, this discussion is all theoretical for me at this point, but in keeping with that, if I were reviewing the paper and the rationale for grouping the animals was to reduce stress and encourage more exploratory activity, I would suggest drawing blood to ensure that those that explore the novel object less didn't have a greater CORT response than those that explored the object more (within one group and across all groups). I would also encourage the author(s) to determine if there was any relationship between CORT response and overall exploration in the NOR test, but also (and probably more importantly) the habituation test too.

Another alternative would be to treat the animals as a unit based on which animals were in the NOR test at the same time. With this comes the inherent individualized variability within each "cohort"/"cage"/"group" or whatever you want to call them. This would theoretically control for social dynamics impeding exploration within the test when you compare across groups. Of course, you would need to treat each member of the group identically, and thus have "control groups" and "experimental groups". I'd bet you'd have to run more animals though. And there would be the inherent question of how much your experimental conditions changed the social dynamic - but I'd wager in an early stage experiment such as this theoretical one, you'd be able to look at the social dynamics somewhat superficially, with more in depth analyses as a future direction (given that, at this point, we're not even sure that social dynamics would be a factor). It would be cool to identify the dominance hierarchy within each group member though and determine whether that influenced their interaction times with the objects. This could be done using in-cage observations, tube test, or the warm spot test. There are a few others as well.

Still that doesn't acknowledge your last point - what about interactions between treatment, stress, and cognitive performance? That's a tough one that I've battled with myself. I did a study looking at a nootropic's effect on memory (and effects of stress) on transgenic Alzheimer's mice. I battled with knowing whether the AD mice had more stress associated with the plaque buildup and, rather than cognitively perform in the test like the wild-types or AD mice with the nootropic, these AD "controls" just withdrew as if helpless (i.e. and just sit in the arena with minimal movement). All you can really do in that scenario is incorporate the stress component into your model or find elegant ways of extricating it. In the end, we tried to add a less stressful memory test with long habituation periods. In this case, the added test was the NOR. So we had two cognition tests.

Definitely doable - but like you said, "demanding indeed". :) Best of luck!

We used adult mice and pups in the open field:

Psychology

• Published 1997
• DOI:10.1111/1467-9450.00017

OPEN-FIELD NIDUS AND PULLUS INDICES IN MUS MUSCULUS L.

• S. Béatrice M. Ewalds-Kvist, R K Selander

Swiss albino mice (Mus Musculus L.) were studied in an open field (OF). The question whether laboratory rodents were more attracted to a nest equipped with familiar odoriferous cues from the home-cage bedding or to signals of an unfamiliar newborn pup inserted in the OF was addressed. The results of the three quantitatively recorded OF behaviors, ambulation, retrieval and radial latency, indicated that in females both ambulation and retrieval were enhanced by the presence of a pup in a nestless… CONTINUE READING