For his latest research on motor skills, visual learning, and their effects on human physiology, School of Biological Sciences associate professor Lewis Wheaton and his team went all the way back to the Paleolithic Era to study a very retro skill: stone toolmaking.

“One of the cool things about this particular study,” Wheaton says, “is this opportunity to look at a completely novel motor task, something most people have no idea how to do, and that’s making a stone tool.”

The new research, published today in Communications Biology, attempts to fill in the gaps when it comes to the science of how we learn complex motor skills — and what may be required to relearn them. 

Wheaton says there are studies researching the behavioral changes that are involved with learning complex skills. But research is still thin on how people adapt their visuomotor skills (how vision and movements combine) to carry out a complex task. Wheaton’s current study sought to quantify and evaluate the changes and relationship in action perception processes – how we understand actions, then select, organize, and interpret what needs to be done for a particular task. 

“The overall motivation was to determine if we could see any kind of emerging relationship between the perceptual system and the motor system, as somebody is really trying to learn to do this skill,” Wheaton says. Those are important processes to understand, he adds, not just for how people attain complex motor skills learning, but what would be needed for motor relearning, as in a rehabilitation setting.

Wheaton conducted the research with graduate students Kristel Yu Tiamco Bayani and Nikhilesh Natraj, plus three researchers from Emory University’s Department of Anthropology.

Tracking the eyes to learn about learning 

The test subjects in the study watched videos of paleolithic stone toolmaking for more than 90 hours of training. The subjects’ visual gaze patterns and motor performance were checked at three different training time points: the first time they watched the video, at 50 hours of training, and at approximately 90 hours. Everybody was able to make a stone tool (with varying degrees of success) at 90 hours, but some picked up the skills at 50 hours.

Wheaton says there was a lot of information to pay attention to in the videos. “There’s a lot of physics in (making stone tools). You’re hitting a rock which is made up of all different kinds of material. There could be a fissure or fault lines, and if you hit it the wrong way it could crumble. When you’re doing it at first, you don’t know that.”

As the video training went on, the participants started to pick up cues about how to strike the rock, along with other aspects of toolmaking. “At first you’re watching from curiosity, then you’re watching with intent.”

That was the exciting part for Wheaton and his team: Being able to see the different phases of learning during the training — which they actually could see by monitoring gaze tracking, or where the subjects’ eyes landed on the video screen as they watched (see photo.)

“Part of the study was to understand the variability where they are visually focused as they get better at the task,” he says.

That’s how Wheaton’s team found there are certain parts of the skills learning that connect better to gaze, but others that connect better to the physical act of making a stone tool. “As you’re going through time, your motor abilities are changing, and at some point that allows you to watch somebody else perform the same task differently, suggesting you’re able to follow the action better, and pull more information from the video in a much clearer way.”

The study not only found a connection between gaze and motor skills learning, but that the connection evolved as the learning went on. The next step in this research, Wheaton says, should include brain imaging “heat maps” to determine where learning takes place with this process. 

That could also help Wheaton’s team apply these lessons for rehabilitation purposes.

“That’s the link between that and some of the other work we’ve done in a rehab context,” he says. “If you’re watching somebody perform a task, if you’re undergoing rehab, there are different ways you’re watching the task. You’re not always watching it the same way. Maybe it depends on how good you are, or how you’re impaired, but all those variables play a role into what you’re visually pulling out” of the rehab training.

 

DOI: doi.org/10.1038/s42003-021-02768-w

A new research study from researchers in the School of Biological Sciences and School of Physics focuses on the evolution of reproductive specialization – how early single cells first got together to create more complex multicellular organisms. In particular, scientists leading the study sought to better understand how those early cells decided which ones would focus on reproduction, and which ones would get busy building parts of a larger organism.

The work, published this month in the journal eLife, references “division of labor,” “trade,” “productivity” and “return on investment,” (ROI) to describe those cellular activities. If that sounds like a paper destined for a business magazine instead of a peer-reviewed journal on biological sciences research, there’s a good reason. 

As the study, led by assistant professor Peter Yunker and associate professor Will Ratcliff, notes in the abstract, “A large body of work from evolutionary biology, economics, and ecology has shown that specialization is beneficial when further division of labor produces an accelerating increase in absolute productivity.” In other words, the prevailing theories state that specialization pays off only when it increases total productivity – whether it’s multicellular organism or widgets streaming out of a factory. 

What Yunker, from the School of Physics and the Parker H. Petit Institute for Bioengineering and Bioscience, and Ratcliff, from the School of Biological Sciences and co-director of the Interdisciplinary Ph.D. in Quantitative Biosciences (QBioS) have found is that the conditions for the evolution of specialized cells were actually much broader than previously thought. Absolute productivity be darned, the cells seem to say; specialization appeared to be a winning strategy, even under conditions that should favor cellular self-sufficiency. 

Why? It has to do with the topology of the network of cells within the organism – what Ratcliff calls a branchy structure. That topology determines that the division of labor can be favored, even if productivity suffers. 

“Topological constraints in early multicellularity favor reproductive division of labor” is the title of the team’s paper. Yunker and Ratcliff collaborated with several other Georgia Tech faculty and graduate students on the research: Joshua S. Weitz, Patton Distinguished Professor in the School of Biological Sciences and co-director of QBioS; School of Physics graduate students David Yanni and Shane Jacobeen; and School of Biological Sciences graduate student Pedro Marquez-Zacarias. All are members of Georgia Tech’s Center for Microbial Dynamics and Infection.

Multicellular multitasking

As cells get more complex, they begin to specialize. Some cells are dedicated to reproduction, while others are devoted to other general tasks such as making and maintaining the organism’s body. “In this paper, what we’re trying to figure out is, when is it a good idea to specialize and have that pay off, and when it is a good idea for your cells to remain generalists?” Ratcliff says. “Under what conditions does evolution favor specialization, and in what conditions do simple multicellular organisms keep every cell a generalist?”

For centuries, scientists have known that specialization is very important for multicellularity. “Once we had microscopes, we were off to the races learning about specialization,” Ratcliff says. 

The thinking for the last few decades has been that more specialized cells evolve when specialization results in increasingly higher productivity. “That will push things to complete specialization because there’s more to be gained by specializing than not specializing.” 

Yet what if those cells are not interacting randomly with a lot of other cells, but only with a few cells over and over again? “This is actually the case for a little branchy structure that contains mom and all her kids. The only cells you are attached to are the ones that gave rise to you, and the ones that arise from you,” he says. Those “branchy structures” offer the topological constraints mentioned in the title of the research study. 

Branch banking of cellular products

Yunker explains that those tree-branchy structures can be thought of as similar to fractals, in which math functions are repeated again and again and are depicted as jagged borders stretching into infinity. 

“Mandelbrot sets and the broader study of fractals have been an inspiration for a lot of this,” Yunker says. “After the concepts behind fractals were identified, people eventually started to see them everywhere. Instead of some unique esoteric thing, it was pervasive. In a similar vein, the structures that we find make evolving division of labor easier, these sparse filaments and branched topologies, are common in nature,” including so-called snowflake yeast and some forms of algae.

Yunker agrees that it may seem counter-intuitive, but as you restrict cellular interactions, like swapping of products that can enhance reproduction or specialization, that specialization actually becomes easier according to his team’s mathematical models. 

Cells that produce the same products won’t interact or 'trade' with each other, since that would be a waste of energy and efficiency. “A redundancy comes into play here,” Yunker says. “If you have a lot of similar cells trading, that increased productivity doesn’t do you a lot of good. Whereas if you have dissimilar or opposites trading, even with lower productivity, they’re able to direct those resources in a more efficient manner.”

What can economists and cancer researchers learn from these cells?

Since economics has already figured into the study of how multicellular organisms evolved, with all of that labor and trade and ROI, could that discipline have something to learn from Yunker and Ratcliff’s new theory — could the lessons mean a more efficient way to make all kinds of products?

“Could this apply in economics? Could it apply elsewhere?” Yunker echoes. “This is something we would love to pursue going forward.”

Ratcliff notes the multidisciplinary approach his biophysics and biosciences team took to approaching the study, which also involved mathematical models developed by Weitz. “We were really motivated by understanding both how life got to be complex, and the rules for why it did,” he says. “This paper follows into the ‘why’ category. Fundamental mathematics tells you about the rules evolution plays by, and there are a lot of downstream applications, like cancer research, agriculture, and infectious disease. You never really can predict how someone will leverage basic insight.”