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

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

Office: B12 Scaife Hall

Email: darcyg@andrew.cmu.edu

Google Scholar: Darcy M. Griffin

▲ Back to the top

Research video

▲ Back to the top

Projects

Wearable HD-EMG Sleeve for Enhanced Neuromuscular Decoding

High-density electromyography (HD-EMG) offers a powerful method to monitor and decode upper-limb muscle activity by capturing the electrical signals generated during movement. Current HD-EMG systems are limited by their lack of portability and adaptability, often making them unsuitable for common clinical or home use. These systems tend to focus on forearm muscles for gesture classification while neglecting comprehensive coverage of other arm muscles essential for more nuanced motor control. Our research focuses on leveraging HD-EMG technology and flexible electronics to develop a wearable sleeve capable of capturing high-resolution muscle activity data across the entire upper limb. This system optimizes the selectivity of electrodes over major muscle groups in the arm to ensure maximal signal capture for discrete movements while maintaining adaptability and a sense of comfort. Our goal is to make this technology widely available for healthy individuals as well as individuals with upper limb disabilities. In doing so, we can expand the capabilities of non-invasive solutions for rehabilitation and neuroprosthetic control.

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.

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.

BG-thalamocortical pathways and CB-thalamocortical pathways

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.

schematic of spinal-cord stimulation for restoration or motor function

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.

t-junction filtering and gate control in dorsal horn (diagram)

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.

illustration of leg moving, brain activity, device control

Brain computer interface for digital device control

Brain computer interfaces (BCI) offer new possibilities for individuals with paralysis to control
external devices through thought alone.The Stentrode is a minimally invasive BCI implanted via
the cerebral vasculature to record neural signals directly from the motor cortex.

In this project, we focus on evaluating the signal characteristics obtained from the Stentrode,
analyzing the fidelity, stability, and longevity of the neural recordings. By assessing these
properties, we aim to better understand the signals and enhance decoding algorithms that
translate neural activity into actionable commands for assistive technologies.

Our research explores the possibilities of the Stentrode in restoring motor function and enabling
communication for patients with severe neurological conditions. This includes testing the
device's capabilities in real-world scenarios, optimizing signal processing techniques, and
collaborating with clinical partners to advance this technology towards widespread clinical
application.

▲ Back to the top

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

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

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

Lauren Parola’s LinkedIn

Hung-Te (Howard) Wu

Masters

Research interests: Pain, EEG, Machine Learning, Neuroprosthetics
Email: hungtew@andrew.cmu.edu

Mahavir Prasad

Masters

Research interests: Artificial Intelligence, Human-Machine Integration, Bidirectional Neural Interfaces, Brain-Computer Interfaces, Motor Control, Neuro-health, Assistive Devices
Email: mahavirp@andrew.cmu.edu

Mahavir Prasad’s LinkedIn Profile

Defne Kudug

Masters

Research interests: Neuroprosthetics, Bioelectronics, Prosthetic Control, Therapeutics Engineering, Neurorehabilitation
Email: dkudug@andrew.cmu.edu

LinkedIn Profile

Dheemant Jallepalli

Masters

Research interests: Brain-Computer Interfaces, Brain-Machine Interfaces, Neuroprosthetics, Motor Control, Sensorimotor interaction
Email: djallepa@andrew.cmu.edu

LinkedIn Profile

Zachary Kushnir

Masters

Research interests: Robotics, Mechatronics, Machine Learning, Robot-Human Interaction, Medical Devices, Prosthetics, Motor Control, Neuro-health, Assistive Devices
Email: zkushnir@andrew.cmu.edu

Kushnir Website

Adam Barmash Rubinchik

Masters

Co-Advisor: Tzahi Cohen-Karni
Research interests: Neural engineering, bioelectronics, neurostimulation, electrochemical systems, neurorehabilitation, devices
Email: abarmash@andrew.cmu.edu

Linkedin profile

Sheldon-Anthony Amofah

Masters

Research interests: Neural engineering, Brain Computer Interfaces, neurorehabilitation, neuroprosthetics, motor control, sensorimotor interaction
Email: samofah@andrew.cmu.edu

Linkedin profile

Emma Heinle

Research Associate

Research interests: Systems Biology, Machine Learning, Zoology
Email: eheinle@andrew.cmu.edu

Jenn Shanahan

Lab Manager

Email: jenns@andrew.cmu.edu

▲ Back to the top

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)

▲ Back to the top

Media Mentions

Pitt CTSI

Grover and Weber awarded research funding to study female pain

ECE’s Pulkit Grover and MechE’s Doug Weber won $50,000 to research female pain

CMU Engineering

Carnegie Mellon lands ARPA-H award for implantable bioelectric medicine project

A CMU-led project team secured an award of up to $42M from ARPA-H to accelerate the development of implantable bioelectronic devices that deliver patient-specific therapy and monitor disease status.

Scientific American

Weber weighs in on first Neuralink clinical trial

MechE’s Doug Weber was asked by Scientific American about Neuralink’s first human trial, and he commented on the difficulty of measuring neuron activity, the possibility of setbacks, and future directions for the next clinical trial.

Axios

Weber comments on a new paralysis treatment

MechE’s Doug Weber spoke with Axios on a new treatment for arm paralysis, which is expected to be approved by the FDA.

See More 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.

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.

▲ Back to the top