No matter how spontaneous of a person you consider yourself to be, you’d have a hard time being spontaneous if it wasn’t for your body’s uncanny ability to execute the repetitive actions necessary for locomotion—the technical term for a group of behaviors including walking, running, swimming, flying, and any other deliberate movement. On a physiological basis, locomotion can be understood as a set of incredibly stereotyped neuromuscular events— literally, putting one foot in front of the other. While we often take locomotion for granted in our adult lives, all it takes is one look at a baby to realize that we don’t come out of the box with this ability—instead, organisms generally acquire locomotion during early development in a process which involves both anatomical development (think tadpoles growing limbs) and neuronal development (the establishment of neuromuscular circuits that allow for stereotyped movement behaviors). But how, exactly, are these locomotion-enabling neuronal circuits established during development? The short answer is that we don’t quite know, yet. A new publication from researchers in the Bai Lab at Fred Hutch—who study neuron communication in development and disease—describes a significant step forward in our ability to answer this question.
“When we think about studying a big, complex phenomenon like the development of locomotion,” begins Cera Hassinan, a graduate student in the Bai Lab and first author of the study, “the first thing we need is a model system which is simple enough to be tractable, but not too simple as to be irrelevant.” To this aim, the team turned to everyone’s favorite nematode, the roundworm Caenorhabditis elegans (C. elegans). While these microscopic worms may not look like much, they crawl around using stereotyped movements just like you and me. These worms also have several notable advantages when it comes to studying neural development—they’re cheap to rear in large quantities, they’re optically transparent, and most importantly, scientists know the identity (and development) of every single neuron in their nervous system! Not unlike humans, C. elegans larvae are also born without coordinated locomotive capabilities, which they gradually acquire as they develop. “So, the approach seemed clear: if we can figure out how worms develop locomotion, we could gain insights into the fundamental processes that underlie locomotion in more complex organisms,” noted Hassinan. Immediately, however, the team ran into a challenge. “Basically, we have the tools to precisely image the movement patterns of C. elegans larvae as they mature—what we didn’t have is a precise way to detect and quantitatively measure how stereotyped the worms’ movements are.” If you can imagine looking down at a dish of tiny baby worms learning how to crawl, being able to differentiate between random thrashing, coordinated crawling, and the gradual transition between the two is just as important as being able to image the events in the first place.
To address this challenge, Hassinan collaborated with co-author Scott Sterrett—a graduate student in the UW Neuroscience PhD Program—to adapt and implement a statistical method called Principal Component Analysis (PCA) to C. elegans imaging data. PCA is an analysis method usually used on high-dimensional (complex) datasets—in a nutshell, the method tries to simplify a complex dataset as much as possible while retaining the maximum amount of information concerning the relationships between the variables in the dataset. In this specific instance, the team was working with videos of worms moving about their environments—more specifically, for each individual frame in these videos, they (computationally) separated each worm into 11 body segments and recorded the angle of these segments relative to each other. As you might expect, the resulting dataset—consisting of these 10 angles at every frame for every worm over the recording period—is large, complicated, and difficult to interpret at face value (see above).
Here, however, the team was able to leverage the fact that coordinated locomotion, as we mentioned, is stereotyped, consisting of the same few poses repeating over and over again. Thus, if a worm is exhibiting coordinated locomotion, it spends most of its time in a reduced subset of distinct body poses with transitions in between, generating data that looks complex on the surface but actually has a relatively simple underlying structure which PCA can identify. Furthermore, these few poses that the worms mostly adopt have meaning as well—termed eigenworms by the C. elegans community (a reference to eigenvectors, which might stir some unpleasant memories of college algebra classes in some readers). These eigenworms can be used to describe the dataset in a simplified way (if you imagine points on a coordinate plane representing your dataset, then these eigenworms serve a similar function as the x- and y-axes; imagine having to describe points in a space without axes!). “The analogy that we like to use when explaining this is actually Y.M.C.A.” notes Hassinan. “If you imagine someone videotaping a group of people dancing the Y.M.C.A., there’s a lot of complicated motion going on. Nevertheless, a majority of the dancers’ time is spent in four very well-defined poses; PCA basically lets us parse through and quantify the complexity of the overall dataset in terms of these well-defined simplifying characteristics.”
With this method in hand, the team now had a handle on which to pull to analyze the movement behaviors of developing C. elegans. First, they asked a relatively simple question: how many eigenworms did it take to describe the full dataset? In accordance with previous studies, they found that C. elegans behavior was fundamentally low-dimensional (i.e. not so complex); using only four eigenworms, the team was able to account for 97% of the variance in adult worm swimming behavior. Next, they applied their method to weigh on a contentious question in the field: researchers had long appreciated that adult C. elegans exhibit two visually distinct movement patterns, termed ‘crawling’ and ‘swimming.’ But do these two movement patterns actually represent distinct methods of locomotion? By imaging both behaviors and applying PCA, Hassinan and colleagues found that swimming and crawling were associated with unique sets of eigenworms, supporting the notion that these in fact are two fundamentally distinct modes of worm locomotion. Finally, the team turned their method on their initial question—they imaged the behavior of C. elegans larvae at a series of developmental stages and profiled these behaviors using PCA. What they found was surprising: worm behavior in the so-called young L1 larval stage—a stage at which coordinated locomotion has not yet developed—was still characterized by the same four eigenworms as in adult C. elegans. Overall, this result suggests that the development of locomotion in worms rests not in the appearance of the proper movement postures, but in the proper stringing together of these postures to produce efficient movement.
“Overall, we think the real strength of our method lies in its ability to quantitatively describe C. elegans locomotion behavior during development, which not only lets us ask questions about these behaviors and when exactly they develop, but also facilitates the reproducibility of this research between different labs and model systems,” notes Hassinan. “Even with what we’ve shown using this method so far, we were surprised to see that very immature worms were able to reliably produce the same postures that adult worms use to move, even at a developmental stage at which many of the neural circuits implicated in these movements haven’t yet finished developing.” As Dr. Bai adds, “While this study shows a proof-of-concept that this approach can work, we’re excited now to leverage this technology to perturb neurodevelopmental circuits in worms (using drugs, mutant libraries, or other resources) and precisely measure the resulting behavioral effects to piece apart which specific neural circuits are involved in regulating the emergence of these stereotyped behaviors.”
The spotlighted work was funded by the National Institutes of Health.
Fred Hutch/University of Washington/Seattle Children’s Cancer Consortium member Dr. Jihong Bai contributed to this study.
Hassinan, C. W., Sterrett, S. C., Summy, B., Khera, A., Wang, A., & Bai, J. (2024). Dimensionality of locomotor behaviors in developing C. elegans. PLOS Computational Biology, 20(3), e1011906.