NeuroMechatronics Lab

The NeuroMechatronics lab is home to an inclusive and multidisciplinary community of neuroscientists and engineers. We work collaboratively to understand the neurophysiology of sensory and motor systems in the body. We strive to invent new strategies and devices that will allow us to transform the treatment of neurological disorders. Our research spans the full spectrum from fundamental to applied research and we leverage an expansive network of academic and industry partners to tackle important problems aggressively and at scale. Throughout it all, we focus on the neuroethical underpinnings and wider societal implications of our neurotechnology.

Faculty

Doug Weber

Akhtar and Bhutta Professor, Carnegie Mellon University

Doug Weber is broadly interested in understanding the role of sensory feedback in supporting and regulating a wide range of perceptual, motor, cognitive, and autonomic functions. His research combines fundamental neuroscience and engineering research to understand physiological mechanisms underlying sensory perception, feedback control of movement, and neuroplasticity in sensorimotor systems. Knowledge gained from these studies is being applied to invent new technologies and therapies for enhancing sensory and motor functions after stroke, spinal cord injury, or limb loss. These principles are also being applied to develop wearable devices for enhancing sensory, motor, and cognitive functions in healthy humans. He is committed to transitioning outputs of his academic research into practical technologies that support real-world applications, and he works actively with industrial partners to bridge the gap from bench to market.

Publications

Linkedin

Office

B12 Scaife Hall

Email

dougweber@cmu.edu

Google Scholar

Doug Weber

Darcy M. Griffin

Research Faculty, Carnegie Mellon University

Darcy Griffin is interested in how the brain controls skilled voluntary movements. Her focus is on signaling through the corticomotoneuronal system. She uses computational and behavioral tools to investigate how input from the motor nuclei of the thalamus influence processing in the motor cortex and the descending motor command.

Publications

Linkedin

Office

B12 Scaife Hall

Email

darcyg@andrew.cmu.edu

Google Scholar

Darcy M. Griffin

Research video

Projects

Normalizing timing of rhythms across internal networks (NTRAIN) of circadian clocks

Desynchronization of the peripheral and central “clock” systems of the body, such as occurs during jet lag, has numerous consequences that impair behavior, including cognition. Together with an interdisciplinary group of researchers from Northwestern University, Rice University, University of Minnesota, and Blackrock Microsystems, we are developing and testing the NTRAIN system, an implantable, wirelessly controlled bioelectronic interface to the body’s peripheral “clock.” We will deploy and test the implant using a minimal basis set of biophysical signals to estimate circadian rhythm phase. Mismatches in peripheral and central circadian rhythm phase causes deficits in cognitive performance; therefore, our objective is to validate the phase detection algorithm in a behavioral model using a combination of neurophysiological gold-standards as well as new behavioral assays of working memory. Ultimately, NTRAIN will interface to an external hub capable of acting as a closed-loop controller that accurately predicts circadian phase mismatch and restores the peripheral and central clock synchronization by issuing commands to produce the required peptides depending on the detected state of the peripheral “clock.”

Image credit: Jonathan Rivnay, ADAPTER: NTRAIN team.

Bioelectronics for tissue regeneration

When more than 20% of a muscle is damaged, as is common for soldiers wounded in recent overseas conflicts, the tissue can’t regenerate, and a stiff scar forms in place of the missing muscle, which often leads to significant disability. To address this issue, A multi-institution research team led by the University of Pittsburgh secured a $22 million grant from the Defense Advanced Research Projects Agency (DARPA) to develop a device combining artificial intelligence, bioelectronics and regenerative medicine to regrow muscle tissue, especially after combat injuries. Researchers at Carnegie Mellon, Northwestern, Rice, University of Vermont, University of Wisconsin and Walter Reed National Military Medical Center are also part of this four-year initiative (DARPA BETR).

The REPAIR Patch will be developed and tested in an animal model of volumetric muscle loss (VML). Our research goal is to evaluate muscle and nerve function during recovery from injury. We also assess nerve regeneration processes during chronic nerve stimulation, specifically sciatic nerve. For nerve stimulation and measurements of muscle activities (EMG), implantable devices will be used to schedule stimulating and recording which are repeated at multiple timepoints throughout a 24-hour schedule with battery life up to 12 months.

Schematic shows the bioengineered patch that will allow for enhanced and accelerated wound healing.

Related news story.

Injectrode

The “Injectrode” is an injectable electrode comprised of a polymer that is injected onto the stimulation site using a syringe and cures in place. The Injectrode is flexible compared to traditional metal electrodes, with a Young’s modulus more closely matching that of tissue, and conforms to the neural tissue in the implant site. Electrical stimulation using the Injectrode has been demonstrated in both the brachial plexus and vagus nerve in animal models, demonstrating the potential for wide use in neuromodulation.

We are currently investigating the use of the Injectrode for dorsal root ganglion (DRG) stimulation. DRG stimulation is a common neuromodulation method to treat intractable pain. We are assessing the recruitment properties of the Injectrode compared to a clinical-like electrode in order to further demonstrate the clinically-relevant applications of the Injectrode.

This research is in collaboration with investigators from several institutions including Carnegie Mellon University, University of Pittsburgh, Case Western Reserve University, University of Wisconsin-Madison, as well as the company Neuronoff.

Novel techniques for noninvasive stimulation and closed-loop control of sensory percepts

Together with an interdisciplinary group of researchers from Carnegie Mellon University and Air Force Research Labs, we are developing and validating our StimSculpt method. This method relies upon spatiotemporally patterned delivery of stimulating currents to provide cell-type specific, localized neurostimulation that will be used to produce somatosensory perception. These read/write capabilities will be incorporated into new technologies and devices for enhancing sensory and motor rehabilitation after stroke, spinal cord injury, or limb loss.

Image credit: Pulkit Grover, Sharp Focus: N3 team

Please contact Dr. Douglas Weber to inquire.

Neuromodulation for substance abuse disorders

Cocaine is a powerful stimulant that carries a high risk for addiction. Cocaine use amongst adolescents and adults (ages 12 and older) has plateaued over the last few years, but overdose deaths involving cocaine use has risen significantly since 2016. In 2017, only about 19% of the estimated 20.7 million people needing treatment for substance use disorders actually received treatment, and even with intervention strategies, relapse rates for substance use continues to hover around 40-60%. There is currently no FDA-approved pharmacological treatment or medical device approved for the treatment of cocaine use disorders.

Previous work in the Torregrossa lab has shown that cue-associated cocaine administration strengthens the connectivity between the medial geniculate nucleus (MGN) and the lateral amygdala (LA) and that these connections can be suppressed with an optogenetic manipulation. This discovery has revealed a potential target for treating drug addiction using neurotechnology that breaks the chain of addiction-related signaling in the brain. In order to translate these discoveries into clinically-relevant solutions for humans, we are bringing a neural engineering focus to this work by extending investigations of cocaine-induced potentiation of the MGN-LA circuit in order to (1) identify neural signals associated with drug craving behaviors and (2) alter these neural signals via electrical stimulation in order to curb the drug cravings that could lead to relapse.

Upper limb HDEMG decoding with deep learning

High density electromyography (HDEMG) is often used to “read” the intention of muscles with high precision by way of detecting the small electrical activity produced during muscular activity. We are currently using this technology to read the intention of a forearm in an attempt to predict what the hand will do as a result of said muscle activity. This is commonly referred to as “decoding” the activity. Many kinds of algorithms in the past have been utilized to decode HDEMG activity, but recently most research has converged on utilizing deep learning as it has shown to be robust in its applications, fast in a live environment, and highly accurate.

We are using a parallel multi-tiered deep learning decoding algorithm in an effort to decode HDEMG activity with an extremely high accuracy and allowing for robust transfer learning capabilities. Our hope with this project is to train the algorithm on an easily accessible healthy population, before utilizing transfer learning to apply what the deep learning algorithm knows to individuals with various upper limb disabilities such as amputation or paralysis. In doing this, we seek to improve the accuracy and capabilities of non-invasive solutions to control schemes to neuroprosthetics.

Thalamic inputs to the motor cortex

We are currently investigating how the motor thalamus mediates the influence of cerebellum and basal ganglia on cortical activity and the descending motor command. Ultimately, we seek to generate a cohesive and intuitive model by which we can begin to understand the rules which govern learning, performing, and maintaining the ability to produce skilled voluntary movements. To this end, we focus on the brain’s control of the arm and hand. From changing a lightbulb to shredding a guitar solo, skilled movements of the hand showcase our dexterity and unique ability to move the digits independently. Only primates with a corticomotoneuronal (CM) system display this level of movement precision. Much is known about how the CM system contributes to the complex patterns of muscle activity required for performing highly skilled movements. We are interested in how the thalamic projections influence the excitability, modulation, plasticity and action potential generation of CM cells and their target muscles.

Wearables for multi-modal sensing of hand-grasp

An adequate and comprehensive model is essential in determining dexterous and sophisticated hand-grasp actions during the control of motor units such as robot end-effector, assistive exoskeleton, or powered prosthesis. For instance, when operating a remote clinical surgery with the robotic arm, local subtle control of mechanical torque, momentum, and strength is necessary to avoid surgery accidents. From the natural biomechanical system, hand motion is correlated with muscle activity and joint kinematics. Therefore, the overall control of motion torque, strength, and momentum can be better explained with muscle activity synergy and joint kinematics rather than each of them alone.

In this study, a wearable multi-model sensing platform was built combining multiple hand motion detection sensors. This platform is used to monitor and model the relationship between hand grasping kinematic, grasping force, and hand-control muscle EMG level. Our goal with this platform is to detect and classify hand-grasp actions based on various participant’s performance in different hand-grasp tasks. Such classification can be further used to (1) refine the motor units’ subtle force control, and (2) monitor daily hand-grasp actions that reflect patient’s hand function level.

Spinal-cord stimulation for restoration of motor function after stroke

Stroke is the leading cause of permanent disability in the United States. There are more than seven million stroke survivors living in the US and two-third of them are affected by motor deficits. Physical therapy can improve motor function in people with post-stroke hemiparesis but outcomes are highly variable with moderate to severe paresis, leaving these patients with permanent motor deficits in arm and hand.

In stroke, while the cortico-spinal tract is interrupted, the spinal circuits, capable of controlling arm and hand muscles, are still intact. Previous work has shown to restore voluntary movements immediately, in the lower limbs, by the delivery of Spinal Cord Stimulation to these spared circuits. In this study, we aim at testing the efficacy of a system delivering Epidural Electrical Stimulation, targeting the cervical spinal cord, to enable people with post-stroke hemiparesis to produce functional arm and hand movements, improve muscle weakness and loss in dexterity.

Investigating the mechanisms of DRG stimulation in blocking nociceptive signals

In recent years, dorsal root ganglion (DRG) stimulation has emerged as a treatment option for individuals suffering from chronic pain. While DRG stimulation has been shown to be clinically effective, the mechanisms of action leading to this reduction in pain are still unknown. Two potential hypotheses include (1) an enhanced low-pass filtering effect at the T-junction of pseudounipolar nociceptive fibers and (2) gate control involving inhibition in the dorsal horn. These mechanisms rely on the activation of different afferent fiber types by DRGS, small-diameter C-fibers (1) and large-diameter Aβ-fibers (2). In order to determine which of these mechanisms plays a primary role in generating analgesia, we are studying which afferent fibers types are activated by therapeutic modes of DRG stimulation. In this preclinical study, we are measuring neural activity in the peripheral nerves, dorsal roots, and dorsal horn during stimulation to determine where in the sensory pathway nociceptive signals are suppressed during DRG stimulation. Our results may help engineers and clinicians to optimize stimulation parameters to provide more effective clinical outcomes for patients with chronic pain.

Epidural spinal cord stimulation for restoring sensation in lower-limb amputees

Recent advances in design and actuation have led to important improvements in prosthetic limbs. However, these devices lack a means for providing direct sensory feedback, requiring users to infer information about limb state from pressure on the residual limb. Lack of sensation limits ability to control the prosthesis and leads to slow gait and increased risk of falling. There is also evidence that lack of sensory feedback contributes to phantom limb pain (PLP), and that electrical stimulation at the dorsal root ganglia (DRG) can reduce PLP. The primary objective of this study is to use commercially available, FDA-cleared spinal cord stimulator (SCS) leads to test the effects of electrical stimulation of the DRG and dorsal rootlets (DR) as a means of restoring naturalistic sensation (e.g. pressure, movement), reducing PLP, and improving balance and gait function in transtibial amputees. We use stimulation to (1) produce sensations of pressure and joint movement, (2) reduce PLP, (3) evoke patterns of muscle activity that mimic automatic responses that occur normally during standing and walking, and (4) improve postural stability when standing and walking with a sensorized prosthesis. We characterize these effects over 90 days, including a 7-day take-home trial which incorporates a belt-worn stimulator distributed by Ripple, LLC. Experiments are­ conducted with SCS leads implanted percutaneously in the lateral lumbar epidural space.

Research team

Max Murphy

Post-Doctorate

Research interests

Cortex, Bayesian Filter, Neurorehabilitation

Email

mdmurphy@andrew.cmu.edu

Max Murphy’s website

Alpaslan Ersöz

Post-Doctorate

Research interests

Design and validation of electronic systems to act as neurostimulators

Email

aersoz@adrew.cmu.edu

Ernesto Bedoy

Post-Doctorate

Co-Advisor

George Wittenberg

Research interests

Sensorimotor impairment, neurorehabilitation, neuromodulation

Email

ebedoy@andrew.cmu.edu

Ernesto Bedoy’s LinkedIn page

James "Drew" Beauchamp

Post-Doctorate

Research interests

Pain, motor units, neurorehabilitation

Email

jabeauchamp@cmu.edu

LinkedIn

Omar El-Refy

Doctorate

Co-Advisor

Hartmut Geyer

Research interests

"Motor control and learning, Bipedal locomotion"

Email

refy@cmu.edu

Omar El-Refy’s LinkedIn profile

Nikhil Verma

Doctorate

Research interests

Neuroprosthetics, Human-Computer Interaction, Machine Learning, Brain-Computer Interfaces

Email

nikhilv@andrew.cmu.edu

Nikhil Verma’s website

Ariel Feldman

Doctorate

Research interests

Brain-Computer Interfaces, Neuromodulation, Information Representation, Machine Learning, Learning & Memory

Email

arielfeldman@cmu.edu

Ariel Feldman’s website

Jonathan Shulgach

Doctorate

Research interests

Neuroprosthetics, Telerobotics, Machine Learning, Brain-Computer Interfaces, Sensorimotor Integration

Email

jshulgac@andrew.cmu.edu

Jonathan Shulgach’s website

Nikole Chetty

Doctorate

Research interests

Brain-Computer Interfaces, Neuroprosthetics, Machine Learning, Computational Neuroscience, Signal Processing

Email

nchetty@andrew.cmu.edu

Nikole Chetty’s LinkedIn page

Dailyn Despradel

Doctorate

Co-Advisor

Nancy Pollard

Research interests

Virtual, Augmented, and Mixed Reality, Neuroprosthetics, Prosthetic training, Haptic Perspection, Motor Skill Learning

Email

ddesprad@andrew.cmu.edu

Dailyn Despradel’s LinkedIn page

Ruitong (Larry) Jiang

Doctorate

Research interests

Circadian Rhythm, Learning & Memory, EEG

Email

ruitongj@andrew.cmu.edu

Prakarsh Yadav

Doctorate

Research interests

Spike sorting, Machine Learning, Neuroprosthetics

Email

pyadav@andrew.cmu.edu

Prakarsh Yadav’s web site

Rawan Fakhreddine

Doctorate

Research interests

Brain-Computer Interfaces, Neurorehabilitation, Plastic Window, Motor Skill Learning

Email

ryf@andrew.cmu.edu

Luigi Borda

Doctorate

Research interests

Spinal cord stimulation, AI, EEG, neuroprosthetic, sensory feedback restoration

Email

lborda@andrew.cmu.edu

Luigi Borda’s LinkedIn page

Jehan Yang

Doctorate

Co-Advisor

Zackory Erickson

Research interests

Robotics, EMG, Machine Learning, Neuroprosthetics

Email

jehany@andrew.cmu.edu

Jehan Yang's website

Kriti Kacker

Doctorate

Research interests

Brain-computer interfaces, Neuromodulation, Machine Learning, Signal Processing

Email

kkacker@andrew.cmu.edu

Kriti Kacker's LinkedIn profile

Lauren Parola

Doctorate

Research interests

Biomechanics, EMG, wearable sensing with commercial devices

Email

lparola@andrew.cmu.edu

LinkedIn

Jenn Shanahan

Lab Manager

Email

jenns@andrew.cmu.edu

Alumni

• Mehrdad Javidi (PostDoc)
• Monica Liu (Doctorate)
• Ashley Dalrymple (PostDoc)
• Thomas Hyatt (Research Associate)
• Rifeng Jin (Reseasrch Associate)
• Jessica Kleinbart (Doctorate)
• Ruiqi Zhang (Masters)
• Jordyn Ting (Doctorate)
• Devapratim Sarma (Postdoc)
• Forrest Z Shooster (Doctorate)
• Julian Low (Masters)
• Seo Jeong "Sharon" Park (Masters)
• Ruiqi Zhang (Masters)
• Jackie Hull (Research Associate)
• Sharon John (Research Associate)

Course taught

Course	Course Name	Location	Units	Semester Offered
24-354	Special Topics: Gadgetry: Sensors, Actuators, and Processors	Pittsburgh	9	Fall

Media Mentions

CMU Engineering

CMU, Meta seek to make computer-based tasks accessible with wristband technology

In collaboration with Meta, Doug Weber’s lab is exploring how sEMG signals can enable people with spinal cord injuries to interact with computer and mixed reality systems.

CMU Engineering

Boosting oxygenation for transplanted cell-based therapies

To advance cell-based therapies, researchers have identified a novel device that makes on-site oxygen for biological cells transplanted inside the body.

CMU Engineering

ARPA-H fast tracks development of new cancer implant tech

ARPA-H has awarded $45 million to a multi-institutional team of researchers to rapidly develop sense-and-respond implant technology that could slash U.S. cancer-related deaths.

Wired

Weber emphasizes importance of sensory feedback for prosthetic limbs

MechE’s Doug Weber discusses the future of bionic limbs and their ability to receive sensory feedback.

See More Mentions

The New York Times

Weber discusses tech that restores movement for stroke patients

MechE’s Douglas Weber was mentioned in The New York Times for the research he and other researchers are working on which looks at restoring mobility in stroke patients.

CMU Engineering

New hope for people living with paralysis after stroke

Spinal cord stimulation technology developed by Douglas Weber in collaboration with the University of Pittsburgh offers new hope for people living with impairments that would otherwise be considered permanent.

Mechanical Engineering

Engineering neurotechnology for paralysis after stroke

CMU and Pitt collaborators will develop and test an implantable system to electrically stimulates the spinal cord of stroke survivors to reduce arm and hand motor impairment.

Psychology Today

Weber featured on first human brain implant

MechE’s Doug Weber is among the team monitoring the first human implant of a brain-computer interface (BCI). The BCI was implanted at Mount Sinai Health System in New York City. The goal of the trial is to evaluate safety and efficacy in helping patients with ALS.

Bloomberg

Weber’s NIH trial covered

MechE’s Douglas Weber was referenced in Bloomberg after an NIH-funded trial that he is leading with David Putrino of Mount Sinai placed a stentrode implant in its first patient.

CMU Engineering

Sensing signals in paralyzed muscles

Doug Weber and an international team of researchers detected electrical signals in paralyzed muscles, which could be used to control robotic assistive devices.

Pittsburgh Post-Gazette

Weber study on brain implant featured

MechE’s Doug Weber recently had his research study on a brain implant that will allow paralyzed people to use a computer with their thoughts was featured in The Pittsburgh Post-Gazette.

Mechanical Engineering

Bridging campuses, increasing opportunities in STEM

CMU and Pitt will offer a joint MS-to-PhD program focused on AI, robotics, and neural engineering to increase the participation of underrepresented students in science and engineering.