Brain wave reading robot may help stroke patients
21 August 2012
New research by the University of Houston, Rice University and TIRR Memorial Hermann aims to help stroke victims recover the ability to 'think then do' to the fullest extent possible.
The US National Institutes of Health (NIH) has funded a multi-disciplinary team to develop and validate a non-invasive brain-machine interface (BMI) to a robotic orthotic device that is expected to innovate upper-limb rehabilitation. The new neurotechnology will interpret brain waves that let a stroke patient willingly operate an exoskeleton that wraps around the arm from the fingertips to the elbow.
The University of Houston (UH) is developing the electroencephalograph (EEG) based neural interface and Rice the exoskeleton. The combined device will be validated by UTHealth physicians at TIRR Memorial Hermann, with as many as 40 volunteer patients in the final two years of the four-year, $1.17 million programme.
Repetitive motion has proven effective at retraining motor nerve pathways damaged by a stroke, but patients must be motivated to do the work, said principal investigator Marcia O'Malley, an associate professor of mechanical engineering and materials science and director of Rice's Mechatronics and Haptic Interfaces Lab.
"With a lot of robotics, if you want to engage the patient, the robot has to know what the patient is doing," O'Malley said. "If the patient tries to move, the robot has to anticipate that and help. But without sophisticated sensing, the patient has to physically move – or initiate some movement."
The team led by José Luis Contreras-Vidal, director of UH’s Laboratory for non-invasive BMI systems and a professor of electrical and computer engineering, was the first to successfully reconstruct 3D hand and walking movements from brain signals recorded in a non-invasive way using an EEG brain cap. The technology allows users to control, with their thoughts, robotic legs and below-elbow amputees to control neuroprosthetic limbs. The new project will be one of the first to design a BMI system for stroke survivors.
Initially, EEG devices will translate brain waves from healthy subjects into control outputs to operate the MAHI EXO II robot, and then from stroke survivors who have some ability to initiate movements, to prompt the robot into action. That will allow the team to refine the EEG-robot interface before moving to a clinical population of stroke patients with no residual upper limb function.
When set into motion, the intelligent exoskeleton will use thoughts to trigger repetitive motions and retrain the brain's motor networks. An earlier version of the MAHI-EXO II developed by O'Malley, already in validation trials to rehabilitate spinal cord injury patients at the UTHealth Motor Recovery Lab at TIRR Memorial Hermann, incorporates sophisticated feedback that allows the patient to work as hard as possible while gently assisting – and sometimes resisting – movement to build strength and accuracy.
"The capability to harness a user’s intent through the EEG neural interface to control robots makes it possible to fully engage the patient during rehabilitation," Contreras-Vidal said. "Putting the patient directly in the 'loop' is expected to accelerate motor learning and improve motor performance. The EEG technology will also provide valuable real-time assessments of plasticity in brain networks due to the robot intervention – critical information for reverse engineering of the the brain."
The three institutions bring unique perspectives to the project, O'Malley said. Rice's robotic devices and UH's neural interfaces will make it possible for TIRR Memorial Hermann, led by Gerard Francisco, director of the UTHealth Motor Recovery Lab, to facilitate translational research to fast-track engineering findings into clinical practice.
For more information, click here.