Faster Typing With the Brain, Including Error Correction
Neuroscientists at Stanford University have developed a technique that continuously corrects brain readings, resulting in a more precise way to type using a thought-controlled cursor. The researchers found that their technique is nearly as good as one-finger typing. Promising results have been obtained with monkeys, and a pilot clinical trial for human use is underway.
The research is published in Nature Communications with the title “Single-trial dynamics of motor cortex and their applications to brain-machine interfaces.” The research paper is freely available online.
A Fundamentally New Approach to Brain-Machines interfaces
Current brain-controlled prostheses such as thought-controlled keypads try to estimate motor commands that involve millions of neurons, but can sample only a few hundred neurons. So errors in the sample, for example false readings from neurons that fire too fast or too slow, reduce the precision and speed of the brain readout process. The technique developed by the Standord neuroscientists analyzes the neuron sample and makes corrective adjustments to the estimate of the brain’s electrical pattern in realtime.
The Stanford researchers performed hundreds of experiments to analyze and model the brain dynamics that underlies hand and finger motion on a keypad, for example the neural activity that corresponds to hitting a key. The brain dynamics models were used to tweak the signals measured from the sampled neurons to adapt them to the best matching models and make the thought-controlled prosthetic more precise.
“These brain dynamics are analogous to rules that characterize the interactions of the millions of neurons that control motions,” said first author Jonathan Kao, a doctoral student in electrical engineering. “They enable us to use a tiny sample more precisely.”
“Our neural observations are both low-resolution (on the order of hundreds of electrodes) and noisy (with the arrival of action potentials being Poisson-like),” note the scientists in the research paper. “However, a recent body of literature hypothesizes that analogous dynamical laws, describing how the activity of population of neurons evolves through time, exist in motor cortex. These dynamics characterize how the neural population activity modulates itself over time (for example, through recurrent connectivity) so that the neural population activity at time k is informative of the population activity at time k+1.”
This dynamical estimation should result in more accurate neural state trajectories than could be inferred by merely smoothing the observations without knowledge of neural dynamics.
Thought controlled keypads allow patients with paralysis or ALS (Lou Gehrig’s disease) to run an electronic wheelchair and use a computer or tablet.
“Brain-controlled prostheses will lead to a substantial improvement in quality of life,” said team leader Krishna Shenoy, a Stanford professor of electrical engineering. “The speed and accuracy demonstrated in this prosthesis results from years of basic neuroscience research and from combining these scientific discoveries with the principled design of mathematical control algorithms.”
The U.S. Food and Drug Administration recently gave Shenoy’s team the green light to conduct a pilot clinical trial of its thought-controlled cursor on people with spinal cord injuries. “This is a fundamentally new approach that can be further refined and optimized to give brain-controlled prostheses greater performance and therefore greater clinical viability,” Shenoy said.
The error correction technique developed by the Stanford researchers is likely to permit improving the speed of Brain-Machines interfaces (BMIs) and find applications to consumer BMI systems – for example brain-controlled games – as well.
Images from Stanford University and Wikimedia Commons.