Elom Amematsro is a fourth-year graduate student in the Center for Theoretical Neuroscience at Columbia University, where he is completing his Ph.D. research under the supervision of Larry Abbott and Mark Churchland. Prior to joining the Theory Center, Elom received his bachelor’s degree in physics and mathematics from the University of Utah. Elom’s graduate work is focused on understanding how biological and artificial systems solve multiple tasks using fixed circuitry.
Principal Investigator: Liam Paninski
Fellow: Nana Dufie Akowuah
Project: Developing neural ‘decoders’ (mappings from neural activity to motor output) that are accurate across many types of movements is a central challenge in neural engineering and motor control. Overcoming this challenge is of great importance to brain-computer interfaces (BCIs), which could help restore movement to those with spinal cord injuries, stroke and more. For BCIs to be widely applicable, it is essential that they can “generalize” to new movements that the BCI decoder was not explicitly trained on. For example, if researchers train a decoder using slow movements, they would still want it to work when making fast movements. In this project, the researchers will work to develop machine learning decoders that are better able to generalize to new conditions. To do so, they will fit decoding models to collaborators’ data that includes simultaneous neural and muscle recordings. The ability to code in Python will be beneficial for this project.