The Department of Computational Medicine and Bioinformatics and the Department of Biomedical Engineering are pleased to welcome Anne Draelos, Ph.D., as Assistant Professor. Dr. Draelos is a neuroinformatician who develops models and algorithms to study the relationship between neuronal activity and evoked behaviors.
What is neuroinformatics?
In the 1990’s, bioinformatics developed quickly as a field of research to support the huge amount of biological data collected from DNA and genetics studies. Similarly, over the last 5-10 years, neurological research has experienced a major scientific boost with the development of new technologies that can record from 10,000s to millions of neurosignals, at the same time and for long periods of time. This technological advance results in an immense amount of 3-D data to be processed and analyzed, which is the expertise of neuroinformatics.
“This new technology is fantastic, like never before, we can see much more of what’s going on in the brain, and how it translates into behaviors.”
—Anne Draelos, Ph.D.
Dr. Draelos became interested in neuroinformatics after extensively studying mathematics and computer science, and receiving a Ph.D. in physics. As a graduate student at Duke University, she was an experimental physicist working on quantum devices, investigating low temperature quantum transport in nano materials.
As a postdoc, she joined John Pearson’s Lab in the Department of Neurobiology at Duke University, following other physicists who use their strong mathematical background and apply physics methodology to address biology problems. There, she studied machine learning and statistical techniques to perform real-time analysis of neuronal data. She became broadly interested in the design and construction of new measurements and methods that analyze successive images and model real-time situation. This experience as a postdoc drew her toward the academic career path, wanting to explore her own questions while also enjoying mentoring students.
At the University of Michigan (U-M), Dr. Draelos is further investigating causal relationships between the activity of millions of neurons and actual behaviors. These are extremely complicated 3-D relationships that also change over time. Using new algorithms and data metrics, she is developing data-driven models that can be used to test hypotheses. With such models, one can change the hypotheses as the experimentation is happening—as opposed to a more traditional approach where one tests A-B causality, one experiment at a time. Using statistical machine learning tools (Bayesian inference and optimisation), Dr. Draelos studies the real-time dynamics variations between neuronal activity and behavior at the individual level, rather than working on average. The models are abstractions of what happens, and look for patterns in the data across time. This novel approach balances informed decisions, human biases, and randomly looking, leading to unexpected discoveries. For example, when studying the behavior of a mouse, scientists often focus explicitly on the eyes, the nose or the whiskers. These types of unsupervised artificial intelligence models can instead point to, for example, subtler muscular activity in the face muscles of a mouse.
Most of Dr. Draelos’s work is computational and her U-M lab will have two main sides. One will be highly collaborative, applying machine learning algorithms on computational hardware that can be put into a collaborator’s lab and interface with other electronics. These algorithms can be used to run adaptive and real-time experiments in collaboration with other scientists. The other side of her lab will be more experimental and aim at unraveling naturalistic behavior. Neuroscience often uses zebrafish as a useful model organism because, when young, these are transparent and their entire brain is visible. Additionally, zebrafish are relatively complex organisms that do have movement, visual responses, social and hunting behaviors. They make it easy to study unsupervised behaviors and responses to different stimuli.
Dr. Draelos loves exploratory research and understanding complex networks. She seeks to answer fundamental challenging questions: “My research is more driven by finding the right hypothesis to test rather than by translational applications,” she said.
“Teasing apart highly complex systems is very interesting to me, and that’s what drew me to neuroscience. And I think that there will be interesting unresolved questions for the rest of my career!”
—Anne Draelos, Ph.D.
Draelos was drawn to U-M by its people who are genuinely open to collaboration, and she considers that her work will be most impactful when working with other experimentalists. “People here seem genuinely interested in what they are doing, sharing their excellent expertise. There are also great students and abundant resources, all of which are necessary to do excellent work,” she said.
Outside the lab, she enjoys hanging out with her two dogs, sailing in the summer, and she looks forward to resuming ballroom dancing with her husband. From her experience with martial arts, she said “it is like a three-dimensional chess game with your body.”
Dr. Draelos started her U-M joint appointment as an Assistant Professor in January 2023, and is launching her own lab, seeking collaborators and students. She can be reached at [email protected].