New Advances in Neural Interfaces Research at Johns Hopkins

New research on neural interface technologies has shown:

• How quickly electrically stimulated “artificial touch” is perceived by the brain, relative to mechanical vibrations delivered to the hand
• How the representations of forearm muscles in the brain change over time in a person with an injured spinal cord
• How blending autonomy and human control in a brain-machine interface improves the performance of everyday activities using robotic arms
• How altering the pattern of electrical stimulation of the eyes can improve visual perception in people with impaired vision
• How performance of the APL-developed Modular Prosthetic Limb improved during a nine-week at-home trial with real-time data logging

Two Johns Hopkins University divisions — the Applied Physics Laboratory (APL) in Laurel, Maryland, and the School of Medicine (SOM) in Baltimore — are working at the forefront of neural interface technologies for functional restoration, rehabilitation and augmentation for people affected by spinal cord injury, upper limb loss or blindness. Five recent peer-reviewed publications provide a snapshot of advances in this domain, building on work that originated under the Defense Advanced Research Projects Agency’s Revolutionizing Prosthetics program.

“Our research has been evolving into new and exciting directions focused on state-of-the-art invasive and noninvasive neural interfaces and bidirectional controls, such as the ability to decode information from the brain and deliver information back to the brain,” said Francesco Tenore, a senior neuroengineer with APL’s Intelligent Systems Center who leads several projects in brain-machine interfaces (BMIs) and applied neuroscience. “In addition, we are working toward moving these technologies into our study participants’ homes to help increase their independence, which presents challenges and opportunities well beyond our participants’ in-lab experiences.”

These publications, he added, expand our understanding of current and future capabilities underlying invasive neural interfacing technologies.

Artificial Touch and Mechanical Vibrations

In “Perceived timing of cutaneous vibration and intracortical microstimulation of human somatosensory cortex” published in June in Brain Stimulation, researchers investigated how quickly “artificial touch,” or touch elicited by electrical stimulation of human somatosensory cortex hand areas, was perceived relative to mechanical vibrations delivered to the hand.

“Interestingly, even though cortical stimulation bypasses much of the touch processing pathways, we found that artificial touch lagged behind the detection of vibration. However, these timing differences are small, particularly when both stimuli have the same intensity, implying that feedback from artificial touch is rapid enough to be effective in neuroprosthetic applications,” said Breanne Christie, a biomedical engineer in APL’s Research and Exploratory Development Department (REDD) and the lead author on the paper.

Understanding the cortical representations of movements and their stability can shed light on improved brain-computer interface (BCI) approaches to decode these representations without frequent recalibration. In “Characteristics and stability of sensorimotor activity driven by isolated‑muscle group activation in a human with tetraplegia,” published in June in Scientific Reports, researchers characterized the spatial organization (somatotopy) and stability of bilateral sensory and motor maps of forearm muscles in a person with incomplete cervical-level spinal-cord injury. This allowed the researchers to observe how the representations of forearm muscles in the brain changed over time.

“This was an important undertaking because the better we understand changes in cortical representations over time, the more robustly and effectively we can design our brain-machine interfaces to decrease the training time required before each use of the interface,” Tenore explained. “Our goal is to connect a user to the system and immediately be ready to decode neural activity to conduct activities of daily living and much more, but daily training is still a necessary activity for effective use of the BMI.”

Blending Autonomy and Human Control of Robotic Arms

In “Shared Control of Bimanual Robotic Limbs With a Brain-Machine Interface for Self-Feeding,” published in June in Frontiers in Neurorobotics, the team investigated an activity of daily living: cutting food and eating, but accomplished using two robotic arms.

“We showed what is possible when we add intelligence to the robotic prosthetic system,” explained lead author David Handelman, a senior roboticist at APL. “The collaborative approach to control blends BCI signals with robot autonomy to enable a user to customize complex tasks, such as which food on a plate to eat and how big a piece to cut. The ultimate goal is adjustable autonomy that leverages whatever BMI signals are available to their maximum effectiveness, enabling the user to control the few robot actions that most directly impact the performance of a task while the robot takes care of the rest.”

Beyond invasive cortical implants for restoring functionality and sensory perception for individuals with spinal cord injuries, APL researchers are also developing techniques for restoring vision through neural interfaces in the eye.

Improving Visual Perception of Shapes With Electrical Stimulation

In “Sequential epiretinal stimulation improves discrimination in simple shape discrimination tasks only,” published in June in the Journal of Neural Engineering, researchers investigated the utility of a novel paradigm for relaying information about the appearance of an object to people affected by retinitis pigmentosa, a genetic disorder of the eyes that causes loss of vision. Electrical stimulation of the retina can elicit flashes of light called phosphenes, which can be used as rudimentary visual feedback, explained the paper’s lead author, APL’s Christie. Electrical stimulation can be delivered through electrodes that are surgically implanted on the retinal surface of the eye, Christie said.

Christie explained that functional sight requires stimulation through multiple electrodes, but phosphenes tend to merge in a way that quickly becomes uninterpretable for people using these visual prostheses. Instead of stimulating the electrodes simultaneously, APL researchers found that sequentially stimulating electrodes improved the research participants’ abilities to discriminate between simple shapes, such as differently sized squares, but did not help them discriminate between more complex shapes, such as letters of the alphabet.

Improved Understanding of Modular Prosthetic Limb From a Nine-Week At-Home Trial

In addition to demonstrating novel progress for invasive neural interfaces, researchers are also building on APL’s work with people with upper-limb loss using the Modular Prosthetic Limb (MPL) to better understand how an advanced prosthesis is used in the home during daily living, demonstrating the translational impact of this rehabilitation research.

In “Monitoring at-home prosthesis control improvements through real-time data logging,” published in the Journal of Engineering in May, a participant was able to take the MPL home and use it for nine weeks. This recent work follows an MPL take-home study in which an amputee was able to use the dexterous prosthesis to learn, among other things, how to play the piano.

“Because of the extensive sensors and data logging on the MPL, we were able to measure improvements in prosthesis control and usage over time without the need for in-clinic visits or evaluations,” explained lead author Luke Osborn, a neuroengineering researcher at APL. “This is important because we are starting to discover new ways to quantify function and rehabilitation progress that are more specific to each individual user. Understanding how a prosthesis is being used outside of the lab or clinic provides critical insights on how it is used and how it functions. This information is helpful for creating more effective user outcome metrics and developing devices that minimize chances of a patient abandoning a prosthesis.”

Ongoing Research in Restoring Limb Function to Amputees

APL and SOM are recruiting study participants with a high-level spinal cord injury (a condition known as quadriplegia or tetraplegia) to continue investigating the use of these devices to provide functional restoration.

“We also continue to explore novel technologies and interfaces, such as noninvasive neural interfaces and flexible electrodes that measure muscle activity and sensory feedback approaches, for the purpose of restoring and improving sensorimotor function for upper-limb amputees,” Tenore said. “Ideally, in some cases, augmentation may even be possible if we can show that through these neural interfaces, our participants may be able to do things that no other human can.”

These recent developments are critical steps in not only demonstrating scientific breakthroughs in neural interfaces but also collectively contributing to the advancement of improving motor and sensory functionality.