Bio

I am an Assistant Professor in the Allen School of Computer Science and Engineering at the University of Washington. My research group focuses on the intersection of neuroscience, neuroengineering, machine learning, and data science. We develop computational models and algorithms for understanding how single-trial neural population activity drives our abilities to generate movements, make decisions, and learn from experience.

Previously, I was a Postdoctoral Fellow in the Department of Electrical Engineering at Stanford University, where I was jointly advised by Professors Krishna Shenoy (EE, BioE & Neurobiology; Stanford & HHMI), Bill Newsome (Stanford Neurobiology), and David Sussillo (Stanford EE & Facebook Reality Labs). My postdoctoral work was focused on developing deep learning techniques for understanding population-level neural computations underlying perceptual decision making in the brain. This work was recognized by a K99/R00 Pathway to Independence Award from the National Institutes of Health.

I completed my PhD at Carnegie Mellon University, where I was jointly advised by Professors Byron Yu and Steve Chase. There, I developed brain-computer interfaces as a scientific paradigm for investigating the neural bases of learning and feedback motor control. My dissertation, titled "Interpreting neural population activity during feedback motor control," was awarded the A.G. Milnes Best Thesis Award by the Department of Electrical & Computer Engineering.

Contact: mgolub@cs.washington.edu

Research interests


Machine learning for neural data analysis

Probabilistic latent variable models, deep neural networks, and other optimization-based frameworks. How can we use tools like these to extract meaning from recordings of large neural populations? What new statistical techniques are needed to test hypotheses about neural computation?

Decision making and motor control

How do populations of neurons generate decisions based on sensory evidence, and how do these decision-making processes flexibly interlace with the action selection and execution processes that they inform?

Population-level changes in neural activity during learning

How do populations of neurons change their joint patterns of activity during learning? What are the limitations of these changes, what are the timescales of those limitations, and how can those limitations be overcome to facilitate faster learning to higher levels of proficiency?

Brain-machine Interfaces (BMIs)

By translating intracortical recordings into signals for driving prosthetic devices, BMIs offer restored movement and communication for those with spinal cord injuries, neurodegenerative diseases or limb amputations. These systems can also serve as a powerful testbed for basic neuroscientific discovery.