Brief overview (greater detail at Research)
Movement is typically taken for granted. Which is a testament to the prowess of our nervous system, which so ably handles tremendously compute-intensive sensory-cognitive-motor integration and seemly effortlessly generates motor control signals without usually even rising to consciousness. We walk, reach, write and speak nearly effortlessly despite making thousands of such movements, very often in novel situations.
There are, however, two major exceptions that remind us of how reliant we are on motor control and why it is one of the major systems and computational neuroscience challenges of our time. Along with one of the greatest biomedical opportunities of our time, as it has recently become possible to translate basic pre-clinical neuroscience and neuroengineering into clinical trials to help people with paralysis with so-called brain-computer interfaces (BCIs), and even on to new industries that aim to help people.
The first reminder is when we are trying to learn, refine and perfect a new and challenging movement, including at the level of an elite athlete (e.g., Fig. 1). It is really hard to get really good and be highly consistent at very spatially and temporally precise movements, so much so society pays people that can master this millions of dollars to do so (i.e., elite athletes).
The second situation where we don't take movement for granted is if we've unfortunately lost the ability to move due to a neurological injury or disease, such as an upper spinal cord injury (e.g., Fig. 2), ALS or brain stem stroke.
Interestingly and importantly, BCIs provide the first means by which it is possible to interact with the world merely by "thinking about it,” which more specifically means attempting to make movements or otherwise reliably modulating neural activity.
To address these challenges and pursue these opportunities we conduct (1) neuroscience research, (2) neuroengineering research and (3) translational research to better understand how the brain controls movement and to then design medical systems to assist people with paralysis. We conduct this research in our Neural Prosthetic Systems Lab (NPSL) which focuses on fundamental computational and systems neuroscience, neuroengineering and electrical engineering. Prof. Krishna Shenoy, PhD is the Director (PI) of the NPSL.
Our research highly aligns with one of the eleven NAE Grand Challenges: "Engineers envision computerized implants that would decode the brain’s intentions... implants that could literally read the thoughts of immobilized patients and signal an external computer, giving people unable to speak or even move a way to communicate with the outside world." National Academy of Engineering Grand Challenges for Engineering, 2008 & 2017. pdf url
Our nonhuman primate neurosurgical work is done in close collaboration with Adjunct Prof. Stephen Ryu, MD (Chair, Department of Neurosurgery, Palo Alto Medical Foundation). Our modeling and computational work is done in close collaboration with Adjunct Prof. David Sussillo, PhD (Director, CTRL-Labs West Coast, a division of Reality Labs / Meta Platforms) and Prof. Maneesh Sahani, PhD (Director of the Gatsby Computational Neuroscience Unit, University College London), which we term Computation Through Dynamics (CTD) and is part of a Simons Foundation Collaboration on the Global Brain (SCGB) program. David Sussillo and Krishna Shenoy are the PIs of the CTD effort at Stanford.
We also conduct most aspects of this research as part of our Neural Prosthetics Translational Lab (NPTL) which focuses on fundamental human neuroscience and translational research with people with paralysis. Prof. Jaimie Henderson, MD and Krishna Shenoy are the co-directors of NPTL.
(1) Neuroscience Research. We investigate the neural basis of movement preparation and generation using a combination of opto-/electro-physiological (e.g., optogenetics, 2p GCaMP imaging, Utah electrode arrays, Neuropixel arrays) techniques in rhesus dorsal premotor cortex (PMd) and primary motor cortex (M1) (Fig. 3), behavioral measurement and training techniques (e.g., eye and arm tracking, EMG muscle measurements, haptic force application / perturbations) and computational and theoretical methods (e.g., dimensionality reduction, dynamical systems, single-trial neural trajectory analysis, recurrent neural networks, deep neural networks). Questions include how neurons in premotor and primary motor cortex plan and guide reaching arm movements.
(2) Neuroengineering Research. We investigate the design of high-performance and highly-robust BCIs. These systems decode neural activity from the brain (see human PMd and M1 sketch above) into control signals for restoring lost motor and communication abilities (e.g., Principles of Neural Science, 6th Edition, Chapter 39). This work includes statistical signal processing, machine learning and real-time system design. Questions include how to design BCIs rivaling the communication rate of spoken language.
(3) Translational Research. We investigate BCI clinical translation (see BCI System sketch above) and highly-related human neuroscience questions (see Natural System sketch above) through the multi-site clinical trail that we are a part of (Timeline on BrainGate, NCT00912041 on clinicaltrials.gov). Questions include how best to bring neuroscience discoveries into clinically-viable BCIs to help people with paralysis in real-world settings.