Research

Assisted robotics devices
Development of a Robotic Assistive Device with Neuromuscular Insight

Find the minimal amount of assistance that will restore healthy gait in stroke patients with Stiff-Knee gait (SKG) after stroke.

Feedback
Feedback control of the BOLD signal

This project aims to push neurofeedback one step forward by creating a control system that differentially regulates the subject's brain activity to a desired value.


A Predictive Kinematic Biomarker of Recovery for Acute Patients with SCI and Stroke
Monitoring gait kinematics during therapy of acute SCI and stroke patients and formulate better predictors of recovery

Assisted device
Towards Neurally-Guided Physical Therapy Following Stroke
Using real-time fMRI neurofeedback to enable people to differentially self-regulate their own brain activity.

A Predictive Kinematic Biomarker of Recovery for Acute Patients with SCI and Stroke

Project Overview:
Monitoring gait kinematics during therapy of acute spinal cord injury (SCI) and stroke patients and formulate better predictors of recovery

Project Description:
The first few months after neurological insult such as spinal cord injury (SCI) or stroke, known as the acute stage, are the most critical to recovery. We propose that acquiring data during therapy, measuring both energy (effort) and performance, will more accurately predict clinical outcomes compared to initial level of impairment. Our goal is to examine how much and how well patients perform gait therapy in the acute phase by unobtrusively measuring 3D gait kinematics of patients throughout the entire therapy session over a total of 15 sessions. The results of this study will produce a far more comprehensive biomarker during acute therapy than previously attempted, as it monitors the entire therapy experience.

Development of a Robotic Assistive Device with Neuromuscular Insight

Robotically assisted device

Project overview:
Find the minimal amount of assistance that will restore healthy gait in stroke patients with Stiff-Knee gait (SKG) after stroke.

Project description:
Our initial study assisting knee flexion torque during preswing predictably improved knee flexion, but surprisingly increased hip abduction. Since biomechanics could not identify this phenomenon, we concluded that there must be an abnormal discoordination. The aim of this project is to identify the nature of this discoordination, include it in a novel neuromuscular model, and then simulate the effects of other kinds of assistance to predict what the optimal assistance should be.

Previous studies show different classes of lower-limb discoordination (i.e. reflex and voluntary) following stroke, but they have never been identified during gait due to lack of perturbation studies such as ours. The first stage of this project is to verify the existence of these mechanisms during gait using the collected data during perturbations from our existing dataset and execute dynamic simulations. The next stage is to develop a patient-specific neuromuscular simulation workflow for SKG assistance which incorporates neuromuscular impairments (discoordination, reflexes etc.) specific to SKG and than obtain optimal assistance using a virtual assistive device with forward dynamic simulation. Finally we will design and test a device capable of providing the type of assistance provided by simulations.

Feedback control of the BOLD signal

Feedback

Project overview:
This project aims to push neurofeedback one step forward by creating a control system that differentially regulates the subject's brain activity to a desired value.

Project description:
In a first approach, this project aims to investigate the primary motor cortex (M1) and create a multi-parametric model of the M1 activation in regards to force and complexity of finger movements. This first experiment will give important insights into the differential sensorimotor response of the brain to multiple input parameter combinations. In a second approach, this project attempts to exploit the discovered multi-parametric model by creating a non-invasive neurofeedback control algorithm that adjusts its inputs to the subject according to their online measured M1 activity. Such a virtual therapist could improve current therapeutic methods as it is autonomously and in real-time adapting finger movement instructions of the subject to achieve a desired M1 activation. This project will take neurofeedback one step forward as it externalizes the controller and avoids unpredictable or difficult reproducible outcomes of self-regulation made with the subject's own thoughts.

Towards Neurally-Guided Physical Therapy Following Stroke

Neutrally guided

Project Overview:
Using real-time fMRI neurofeedback to enable people to differentially self-regulate their own brain activity

Project Description:
Current physical therapy following stroke aims to indirectly induce neuroplasticity in motor regions of the brain using physical manipulation of the limbs. This practical clinical approach has the potential to induce maladaptive neural changes that may inhibit long-term recovery. We aim to enhance these physical therapy techniques by taking into account neural activity during therapy using functional magnetic resonance imaging (fMRI). Our goal is to develop a non-invasive technique for healthy subjects to endogenously generate patterns of brain activity associated with increased motor performance, as a proof-of-concept for neurally-guided stroke therapy. Our technique uses real-time fMRI to provide subjects with moment-by-moment feedback of how closely their patterns of brain activity match desired levels. This study will provide a basis to guide stroke patients to differentially self-regulate their brain activity using physical movement, in order to assist in recovery.