Ask a Caltech Expert: Richard Andersen on Brain–Machine Interfaces

Breakthroughs in neuroscience and technology have made it possible to measure brain signals and use them to control devices such as robots or computers. These technologies are called brain–machine interfaces (BMIs) or brain–computer interfaces (BCIs). BMIs and BCIs have enormous potential to transform the lives of the hundreds of thousands of people who have lost the ability to move or speak as a result of spinal injuries, stroke, or neurological diseases.

In the research group of Richard Andersen, Caltech's James G. Boswell Professor of Neuroscience and director and leadership chair of the T&C Chen Brain-Machine Interface Center, scientists explore the brain signals that encode movement and speech and use that information to develop systems called neural prosthetics that would restore these functions to people who have lost them. Andersen explains:

This is all made possible by being able to record, for the first time, from populations of neurons, which are the building blocks of the brain. Measurements from one neuron provide some information, but to really understand the brain, you have to look at many neurons, and that has been very illuminating.

In one procedure, implants are inserted into regions of the brain that are known to control movement. These implants are smaller than a thumbnail and can hold dozens of electrodes, allowing them to measure hundreds of neurons during each recording session. Participants paralyzed by spinal cord injuries are then asked to think about performing a task, such as grasping a cup on a table, and the resulting brain signals are recorded. The signals can then be converted to electrical control signals to move a robotic arm to grasp the cup.

We have made similar progress with speech. Brain signals measured while a person is thinking about words can be decoded and used to make a computer communicate the same words.

In both instances, the system uses algorithms to interpret the patients' intentions based on neural signals, then converts them into electrical control signals to operate external devices. And, because we are focused on intentions rather than the actual performance of tasks, the studies are helpful for people who cannot physically move or speak.

We are extending our work to new applications. For example, we have a driving demonstration done in collaboration with Ford and Blackrock Neurotech that enables volunteers here in Los Angeles to drive a car located in Michigan.

In my opinion, the volunteers are the heroes in this field. About 35 participants have received implants worldwide at a handful of laboratories, so it's still relatively new technology. Participants enter into a clinical study that lasts about five years, and they work more or less every week. It is a job with a goal, and I believe it significantly improves the quality of life of the subjects. Reciprocally, research participants improve the quality of research in our lab. The students and I get to know their families, and the families get to know us. There are many dimensions when you work with humans in research.

As I think about the future of our work, I think about refining these BMIs and BCIs. Implants should become wireless and last longer with the help of better materials. We are also working with partners to design implants that can collect data from a larger number of neurons, resulting in more accurate decoded signals.

For people who do not suffer from neurological disorders, applications of this technology might include increased memory, the ability to speak multiple languages, and improved navigation. But there would be many ethical considerations, such as who would have access to these systems. For now, the technology is far enough away from being optimized that we have time to think through the ethical considerations—but it should be an active area of discussion.

—Richard Andersen, Caltech's James G. Boswell Professor of Neuroscience, and director and leadership chair of the T&C Chen Brain-Machine Interface Center

Andersen discussed his work as part of the 2023–2024 Earnest C. Watson Lecture Series.

You can submit your own questions to the Caltech Science Exchange.