Previous Research

*** This is an old website from when I was a PhD Student in Dynamical Neuroscience at UCSB. Please see https://sites.brown.edu/luke-rosedahl/ for my current postdoc research at Brown. ***

My Research Story

We all have a story for how we ended up in research. This is mine.

It was 3am on the last day of fall semester sophomore year, and my research project was due that day. As I hunched over my computer, alone in the apartment dining room at the University of Minnesota, I deeply regretted choosing to add additional features at the last minute. It wasn't my grade I was worried about; I had already submitted the required paper a week ago and had received my grade. No, the reason I was drinking coffee at 3am was that I knew I was close.

I made one final adjustment and pushed run. As the program uploaded the custom brain atlas warping code to the supercomputer, I sat back and rubbed my eyes. In five minutes I would know if it worked, but until then I watched the gentle falling snow swirl by the streetlight outside our window. After what seemed like an eternity, I got the alert that signified the program had finished. As I looked down and saw the perfectly warped brain atlas overlaid on the MRI, I felt a thrill of excitement run down my spine. At that moment, I knew I was hooked.

At the time I didn't know how much I would grow to love research. I could never have imagined myself spending a summer doing research in Pittsburgh, being an invited speaker at a conference, going to graduate school, or pursuing a career in academia. All I knew is that I wanted to do more.

Project 1: Bettering Electrode and Brain Region Visualization during Neurosurgical Planning.
Excited to get more involved with research, I took on my first real project: editing an existing program Monkey Cicerone to provide better visualization of electrode lead trajectories. Working with Matthew Johnson in the Neuromodulation Research and Technology Lab, I achieved this by adding the ability to visualize up to three co-registered MRIs. As different weighted MRIs highlight different structures, this allows the user to better visualize brain targets in relation to the lead trajectories.

Beyond the technical TCL coding experience, developing Monkey Cicerone gave me experience running my own project, setting goals, creating timelines, coming up with new development ideas, prioritizing tasks, incorporating feedback from others, and helped me become accustomed to lab dynamics.

The most important impact of the project, however, came from presenting it at the Institute for Engineering in Medicine Retreat. Going in, I had a very narrow view of the field of neuroengineering. I thought that there was little beyond deep brain stimulation, lead design, and neurosurgery. At the conference, I saw the true breadth of research going on and became motivated to further explore the field of neuroengineering.

Project 2: Planning, Developing, and Speaking on a 3D Printable Microdrive. Inspired to branch out in my research, I started my second project: developing a 3D printable microdrive system. The goal was to develop a completely open-source, easily constructed system providing automatic control of microelectrode depth and neuronal tracking with real-time visualization. I built a closed-cylinder drive supporting 10 electrodes with automated actuation, visualization, and easy tracking algorithm implementation. The final prototype met all the goals of the project and will continue to be iterated upon in my absence at the University of Minnesota.

This prototype led to a poster presentation at the Minnesota Neuromodulation Symposium, as well as an invited talk at the National Biomedical Engineering Society Annual Meeting in Tampa, Florida. Presenting my research to a room of 60-80 people honed my oral presentation skills and receiving development suggestions from individuals with different backgrounds taught me the importance of group scientific feedback.

Overall, this project solidified my interest in research as an undergraduate and motivated me to apply for a summer research opportunity at the University of Pittsburgh.

Project 3: Using Machine Learning Methods to Classify Local Field Potentials. Certain that I enjoyed research but uncertain of my future plans, I spent a summer in the Schwartz lab at the U of Pittsburgh. Out of the many options there, I found myself drawn to the task of altering their existing spike-based idle detection method during NHP robotic arm tasks. The goal was to find a way to remove the daily calibrations and make the system more stable over time. Starting with their existing code, I developed a system in which time/frequency data were extracted from Local Field Potentials using Gabor Wavelets and used to train a Linear Discriminant Analysis model.

While LFPs ended up showing much lower performance than spike data, the project taught me the importance of persevering with research and accepting a negative result as a result. More importantly, during this project I found that I loved academia, machine learning, and analyzing data, which set the course for my career in computational neuroscience.

Project 4: Working as a Team to Research and Write a Paper on Phase-Amplitude Coupling in Parkinson's Patients. Coming out of the summer in Pittsburgh wanting to do more data analysis, I started my final U of MN project: studying Parkinson's disease biomarkers. Parkinson's patients exhibit pathological phase-amplitude coupling (PAC) in motor cortex and basal ganglia. However, the variation of PAC across patients and brain regions remains unclear. The goal of the project was to investigate variation using magnetoencephalography (MEG) and isolate PAC differences between PD and healthy individuals.

Working in a diverse team (comprised of myself, a grad student, a med student, and an undergrad), I designed analysis methods and compared their performance in evaluating PAC. While the final results and paper are being drafted, PAC was found to be significantly greater for several key regions. This result further supports the potential for PAC being used as a diagnostics tool in Parkinson's disease and brings the field closer to early-onset diagnoses and better treatments.

This project taught me how to collaborate with others on a complex, multi-faceted project. I learned how to communicate our current and long-term goals and visions, work on parallel project aspects, integrate our contributions into a cohesive unit, and draft papers as a team.

Together, these projects taught me a lot about research and motivated me to pursue graduate school, which led to my current Current Research as a PhD Student in Dynamical Neuroscience at UCSB.

Live Free!