Mark Nelson's research interests took him from high energy physics to computational neuroscience to studying one of nature's most unusual creatures. It's a journey that's more direct than might appear at first glance.

Nelson earned a Ph.D. in physics from Cal-Berkeley and did post-doctoral work at Caltech in high energy physics. During his time there, Caltech began a new program called Computation and Neural Systems that sought to bring together physicists, biologists, and engineers to address issues of how the brain works. Nelson was interested in computational aspects of neuroscience, so he joined a neurobiology lab, which happened to be studying weakly electric fish because of their large cerebellum.

"I was in this lab and got intrigued by the electric fish, particularly as a physicist, at the idea that these fish emit electric fields and find their way around the environment using electricity," Nelson said. "It is an intriguing strategy that's different from most animals."

The fascination with weakly electric fish is something that has never left Nelson, who took his work to the University of Illinois as a professor in the Department of Molecular and Integrative Physiology and as a full-time faculty member in Beckman's NeuroTech group.

Today, Nelson's Electrosensory Signal Processing Lab is one of the premier facilities in the country for studying weakly electric fish in its mission to "understand neural mechanisms and computational principles that animals use to actively acquire sensory information in complex, dynamic environments." His research has used computational models to provide new insights into the fish's electrosensory world and pioneered the use of naturalistic contexts for studying the fish's sensory input patterns in ways that were more realistic than what had been done before.

Even though his research is now focused on biology and neuroscience, Nelson said physics is an excellent place for any scientist to begin a career.

"A lot of people make the switch earlier in their careers," Nelson said. "But coming from a physics background you tend to see everything as physics: chemistry is physics; biology is physics, so it's sort of easier to broaden your horizons."

For Nelson, his research interests have taken a natural path of scaling down in order to scale up.

"In the educational process there is this tendency to take the reductionist approach," Nelson said. "You start to analyze biology, then you find out about cells and that the cells are made up of chemicals. You start to think about the biochemistry and realize that the interactions are mediated by physical forces between the atoms. So you pull it down to look to what are the fundamental building blocks of nature. That path led me to particle physics."

The opposite path led Nelson to larger organisms.

"But you can change your mindset from the reductionist analysis of taking things apart into smaller and smaller pieces into turning it around and start thinking about synthesis," Nelson added. "How do you build systems, how did biology get created in the first place and what is its evolutionary history. It takes you back on a reverse pathway from the initial biochemistry to the first cells that were formed to the evolution of nervous systems and brains."

Nelson's interest in nervous systems is focused on weakly electric fish mainly due to their sensory abilities, which are based on generating and perceiving electric fields that allow them to "see" in the dark.

"The general problem is the issue of how animals and humans acquire sensory information about their environment and how they process that information," Nelson said. "I'm particularly interested in the very early stages of processing the data as it is coming into the nervous system, when the nervous system is flooded with information typically from millions of input channels.

"One of the first things that has to be done is the animal has to sort that data and pull out the behaviorally useful information, the information it needs for guiding its behavior in whatever environment it lives in and throwing out a lot of the extraneous background information. The fish are a particularly good model system for understanding how that initial stage of sorting the data works. But the principles generalize broadly across lots of different animals, including humans."

Nelson said that weakly electric fish, which are found in freshwater rivers in Africa and South America, have the largest cerebellum in relation to their brain size of any animal, even more so than humans.

"We're trying to understand why that might be conceptually and have some hypotheses about the role the cerebellum might play in actually controlling the sensory data the organism is receiving," Nelson said. "A lot of people think about motor actions (only) as performing some task like hitting a tennis ball or running. But a lot of what organisms need to do in order to get information about their environment involves motor functioning like turning your head or turning your eyes, or reaching out with your hands to touch something. So a lot of the motor actions that animals perform actually have a sensory acquisition component."

Nelson's lab primarily studies two species of weakly electric fish from South America (the nocturnal black and brown ghost knifefish), but also has used the African species, including a type called Gymnarchus in its research. One large Gymnarchus, nicknamed Gymmy, lived for more than a decade in the lab's aquarium, becoming its unofficial mascot and a popular part of the lab's Beckman's Open House exhibit until his death last year.

Weakly electric fish use a special electric organ that runs almost the length of their body to generate an electrical current that flows out through the water and back into the fish. Nelson said that since the current can't flow through objects such as a rock, it creates "an electrical shadow and the sensors on the fish's skin act like little volt meters that sense the change in voltage or current flowing across the skin. If they swim by a prey object it creates a little electrical bright spot that sweeps across the sensor array."

How the nervous system processes these dynamically changing patterns is what intrigues Nelson and other researchers and what could lead to a better understanding of how sensory input works in the brain.

"Certain animals like the electric fish illuminate their world with electricity and bats illuminate their world with sonar," Nelson said. "For these active sensing organisms, I'm interested in understanding the strategy they use to control that emitted energy because that influences the information that they get back. For example, how do they control the spatial spread of that energy, do they emit it in a narrow cone, which is what dolphins do, or do they emit it all the way around their body which is what electric fish do, to give them a different view of the world. Different species come up with different solutions. I'm interested in comparing different strategies across animals."

Nelson seems to have found a research line that is not only scientifically valuable and interdisciplinary in nature, but also interesting and fun.

"It's a tremendous amount of fun just trying to get your head around this problem and this amazing world that we can't see and we can't hear, but the fish live it every day," he said. "Just trying to see the world through a new sense really opens up your thinking about the problems that animals face and how sensory systems work.

"We just really enjoy the ability to be here at Beckman and merge biology and physics and computation. We don't even think about them now as separate disciplines; they are all combined as we go about trying to understand how the systems work."

The work could have applications in creating sensors for robotics and computer vision. Nelson is currently involved in a project with Beckman faculty members Narendra Ahuja and Al Feng looking at how different animals are able to pick out individual signals from a cluster of similar signals, with the overall goal of understanding general principles for localizing signals, both auditory and visual, in complex environments. Their work could lead, for example, to software that is able to pick out individuals in a crowd or detect specific sounds in a noisy environment.

"In the end all sensory systems, whether visual or auditory or electric, have a physical basis for extracting information from the environment," Nelson said. "It's natural to try and capture this process and describe it in some sort of mathematical model and that's where computation comes in."

Nelson said almost all of the papers coming out of his lab have a neuroscience component, with the computational modeling work undergirding the neuroscience aspects and adding to their understanding of weakly electric fish.

"Some of our papers will be purely experimental while some will be almost purely modeling, but all will be related to these early stages of sensory processing," he said.

Nelson said that his students come from diverse scientific backgrounds, including computer science, neuroscience, computational biology and, yes, even physics. When he tells fellow physicists about his research, they too become interested in the topic.

"Part of it is the attraction that the system holds for physicists once you tell them about the electric field and electric sensing," he said. "A lot of physicists get intrigued."