You warily walk into a dark compartment, wondering if there is food inside. Suddenly there is a loud tone and you feel an uncomfortable surge of electricity through your feet. This goes without saying, but it won’t take long before you will learn to be afraid of that tone. However, over time, you hear the tone without the shock, and slowly (foolishly??) accept that the previous connection may no longer hold.
Or perhaps you are extremely motivated to work for food, given that in your home area, nutrition has been sparse and hard to come by. You see millet seeds seemingly just within the reach of your fore-limb. Though not a typical movement for you, you reach for it. In another instance, you find a different type of food that is difficult to handle. However, it is nourishment nonetheless, so you will learn the required motor skills.
SPOILER ALERT: In each of the above cases, you were a mouse the whole time (I know!) But this is a neuroscience blog, not M.Night Shyamalan’s IMDB page, so perhaps we should focus on what was taking place in the brain as each scenario played out. In both of the cases above, learning was occurring, with new information stored away within the appropriate neural connections of particular brain areas. These situations are on display in a pair of new(ish) papers out in Nature, exploring the structural substrates of such learning and identifying patterns underlying the observed structural changes as learning occurred.
Most excitatory connections between neurons are made onto the ends of small structures called dendritic spines. Almost ten years ago, a splash of papers, including these two in Nature, kicked off an era of in vivo structural imaging that has remained a hot topic and fascinating subfield since. The early work, most of which was conducted in mice, took advantage of transgenic lines in which only a small subset of neurons expressed a fluorescent protein, allowing for little overlap between cells and making for exquisite images. Researchers mainly peered into sensory areas, since external factors like light (visual system) or object proximity (somatosensory) are easy to manipulate, providing a means to induce plasticity and watch how the morphology of a neuron might change in response. And scientists did indeed see some spines being lost, while others formed de novo following such sensory manipulations. The interpretation was that new synapses were being made or broken in response to the external manipulation. Of course, just to make sure, it was determined in the earliest studies that observed new spines DID make new synapses, based on correlative electron microscopy in which newly-formed spines in vivo were later identified in tissue sections and visualized at the ultrastructural level.
Later work began to explore the arguably more complex issue of whether learning can be “‘seen” in neurons through changes in morphology. It would be exciting if indeed the same types of morphological responses to new sensory experience also occurred in response to more abstract experience, such as learning a new skill. The labs of Wenbiao Gan and Yi Zuo (a former post-doc of Gan’s) have indeed taken in vivo structural plasticity to this next level, starting with back-to-back papers in Nature two years ago, where each group demonstrated a selective stabilization of new spines in response to motor learning. In short, skills and experiences are likely to be retained through a vast network of new (and old) connections made between cells. This may sound obvious, but presuming something and actually observing it in the head of a learning rodent are different things entirely.
In their latest papers, Gan and Zuo explore this concept of structural synaptic plasticity in response to varying experience at a deeper level. Gan and colleagues investigated how neural circuits were modified by fear learning and extinction. Pairing an auditory cue with a foot shock in a particular chamber taught their rodent subjects to fear the tone and in frontal association cortex, spine elimination was enhanced. In contrast, extinction protocols designed to make the mouse “unlearn” the association between the tone and footshock increased the rate of spine formation. The degree of this remodeling corresponded well with the behavioral expression of the fear learning and extinction. Interestingly, the formation and elimination of spines during the two learning paradigms often occurred on the same dendritic branches, with spines that were gained during extinction growing out in spots within 2 μm of spines previously lost during fear learning. Also, when fear learning was reconditioned in these animals, the same spines that grew during extinction were subsequently lost. Thus, the authors could almost track these behavioral traces in the structural changes of individual neurons.
Zuo and colleagues explored the related theme of whether neighboring synapses on specific dendritic branches are involved in encoding the same information or experience. The authors watched dendritic spine dynamics in motor cortex as mice learned novel forelimb motor skills, specifically a reaching task or the “capellini task” (the latter was adapted for mice from the better-known “vermicelli task” in rats. Apparently, rodents have a difficult time holding pasta, requiring them to learn new motor skills to adapt.) As the motor skills were learned, a third of new spines formed in clusters near pre-existing spines. Although there is always a basal level of spine turnover regardless of ongoing learning or experience, spine clusters that emerged as mice learned the reaching skills were more likely to persist over time. Interestingly, and adding to the authors’ interpretation regarding the importance of newly emerging spine clusters to skill learning, under control conditions where no motor learning was taking place, new spines seemed to be preferentially added away from pre-existing stable spines, rather than in clusters. Thus, repetitive actions led to motor learning, with structural traces of the learning seemingly left in the clustering of dendritic spines (synapses) involved in the operation.
Now you may have already noticed that although each set of papers by the above authors were published in the same issue of Nature both times, the submission and acceptance dates don’t match up. As you can imagine, for every time authors learn about other labs conducting similar experiments and subsequently coordinate submission, there are many more times when authors (whether they are aware of competition of not) submit similar studies within weeks to months of each other. How do we deal with such issues at Nature? First, each paper must stand on its own, meaning that we initially consider a manuscript on its own merits to ensure that each study can individually demonstrate the final conclusion without support of another study to produce confidence in the results. Once similar papers have passed this test, editors invariably find overlapping referees in order to have some consistency within the review process. However, it is rare that all refs overlap, since often times, each study may include a unique type of experimentation that requires additional expertise. Non-overlapping refs also provide some balance to the process, since it is only natural for a reviewer to try and “rank” each paper drawing a similar or related set of conclusions, perhaps biasing his/her reports accordingly. Sometimes such a ranking is justified, while other times it may be unfair to the authors (but that’s why we have editors, right? To look out for everyone involved??) Finally, once decisions are made on publication, editors coordinate the publications while making every attempt to maintain confidentiality for the authors. In the above cases, I believe that both sets of the authors did not know of the other study until publication. As long as the editor is looking out for the best interests of the authors, both by not delaying one study unfairly, or “scooping” another by a matter of weeks, it is surprisingly easy to reach a satisfactory resolution in such cases.
I lay out a typical “everyone ends up happy” scenario above, but many authors will be well-aware of (and perhaps have experienced) alternative outcomes. Since every one of these situations is unique, it will be difficult to speak more generally on the darker side of failed joint publication. With that in mind, I invite you to ask your questions about co-submission or joint publication as well as to share your experiences in the comments below. I’ll do my best to provide an editorial opinion on each case and perhaps give you an insider’s view into why your decision went down the way it did. Let the joint publication discussion begin!