|
The 800 ft2 Falk Neurorehabilitation Engineering Research Laboratory is the flagship laboratory of the Falk Center, and is one of five labs within the first floor of the Schroeder Complex North building, collectively covering 3,000 ft2, whose primary activity targets neurorehabilitation research.
This laboratory, coordinated by Drs. Brian Schmit and Jack Winters, is unique in that it is available to any group on campus engaged in neurorehabilitation research that can take advantage of the available equipment and capabilities. These include:
- motion analysis equipment (Flock of Birds),
- several force platforms (AMTI),
- 6-axis force transducer (ATI),
- several EMG systems,
- software to integrate motion, force and EMG data collection (Motionmonitor)
- muscle stimulation system,
- a range of computer interface technologies such as joysticks, wheels, the TheraJoy (see photo), a cyberglove and customized devices.
- UniTherapy software for computer-assisted assessment and therapy in neurorehabilitation (see below)
 
Five ongoing research and development projects are highlighted here:
TheraJoy Technology:
This project was motivated by input from therapists and insights from the Falk Workshop, where a need was identified for technologies for arm rehabilitation that included some of the successful features of both rehabilitation robots and of mass-market input devices. This included the need for computer-assisted motivating rehabilitation (CAMR)-type technologies that provide arm movements over a larger range of motion. The initial, Phase I TheraJoy consisted of a mass-marketed force-reflecting joystick that was extended to telescope up to 1 meter and thus enable a larger range of arm movements in a horizontal plane with a slight arc. It also had, in addition to the small force-reflecting joystick motors, larger motors that were connected in series with mechanical springs to actuate the device, capable through computer control of providing moderate assistive and resistive forces by changing the set point of the joystick. A pilot study involving six subjects with stroke showed the viability of the technology, with the results of a user questionnaire showing clearly positive feedback about the potential for this technology.
The Phase II TheraJoy was motivated by user feedback from the first study plus input from therapists that suggested that a device with movement in the vertical plane would be very useful for stroke rehabilitation. The new design, refined based largely on insights from a three-dimensional dynamic model of the new mechanism, can be adjusted for either horizontal or vertical movements and also includes new options for hand contact. It is currently under evaluation by graduate student Laura Johnson and colleagues for a population of stroke subjects. A team of students is concurrently involved in a Phase III TheraJoy that is replacing the linear actuators (currently located under the top base plate) with artificial muscles (braided pneumatic actuators).
UniTherapy:
One of the key barriers to telerehabilitation for neurorehabilitation applications is the lack of software tools that are customized to the needs and technologies of the rehabilitation process. UniTherapy, developed in Microsoft’s Visual Studio .Net environment using C# and the DirectX.9 software development kit (SDK) library of multimedia gaming tools, intends to address this need. This technology, part of the doctoral project of Xin "Tyre" Feng, supports mass-marketed force-reflecting joysticks, wheels (including TheraDrive, see below), mouse pointers and other input devices to provide interactive upper limb assessment and therapy. The structure is shown int eh diagram below. UniTherapy includes support for both Patient Interfaces (PI) and Local Practitioner Interfaces (LPI)/TelePractitioner Interfaces (TI), both with wireline (Ethernet) and wireless (IEEE 802.11 or Bluetooth) network/Internet connections, as well as multiple platforms. UniTherapy also includes a number of assistive and accessibility features that are aimed at making CAMR techniques more viable for home-based upper limb telerehabilitation where the home user may have other co-morbidities or functional impairments in addition to stroke, and likely has a limited range of motion (e.g., voice control interface, a user-abililty space based on initially mapping between input device and cursor location). Modes of rehabilitator assistance include:
- Passive training mode: The patient/client uses one force-reflecting device with the impaired arm. A predefined therapy program is prescribed for the device, thus providing automated therapy (under supervision by practitioner).
- Interactive control mode: The telepractitioner can participate in therapy by cooperative assistance or resistance with the subject during goal-directed computer-assisted therapy. A subset of this would be a training supervision mode, where the telepractitioner applies forces to the patient/client device.
- Bi-manual mode: The patient/client uses two force-reflecting devices simultaneously (e.g., his/her non-impaired arm can assist and “feel” the impaired arm)..
Fine-tuning these modes remains an ongoing research and development activity. A Tracking Protocol Toolbox is available, enabling the protocol designer (e.g., researcher, therapist) to design predictable or unpredictable tracking patterns with a mix of various spacial tracking patterns (square, circle, sinusoidal, random) and temporal sequences. Additionally, a System Identification Toolbox is available for force input perturbations (e.g., impulse, step, ramp, sinusoid, pseudo-random noise responses). These tools can be used to help develop models for the human operator which then can be used to tune input device impedance in ways that map well to user abilities. Performance data can be displayed in real-time, either locally or at the remote Telepractitioner Terminal. A Protocol Manager Toolbox can be used to design suites of tasks to be administed sequentially.
TheraDrive technology:
The aim of this work is to create a low-cost, commercially-viable, home-based rehabilitation system that can capitalize on CAMR concepts of steering wheel game therapy and skill training with functional training related to real activity to induce user-dependent CNS plasticity. The TheraDrive concept combines the motivational elements of re-learning steering tasks with playing driving video games using commercial force-feedback steering wheels to create an upper arm stroke therapy environment that is usable at home or in the clinic. It uses both commercial software and UniTherapy, and builds on a more sophisticated but specialized Driver's Seat system that was part of Dr. Michelle Johnson's doctoral research at Stanford University. For more information see Dr. Johnson's new research page.
Cortical Reorganization after Spinal Cord Injury and Stroke:
There is considerable interest within our neurorehabilitation group to taraget research activities that take advantage of excellence within our BIEN department and colleagues at MCW in the area of functional neuroimaging. Of special interest is the science of relating brain activity to actual neuromotor movement activity, and brain plasticity to rehabilitative training protocols. One example project, involving Matthew Cole, a doctoral student housed in the Falk Lab, examines the mapping of damage in Spinal Cord Injury (SCI), and considerations of cortical reorganization. This project studies sensory and motor mapping for small cyclical ankle movements, and uses a non-ferrous apparatus with pneumatic actuator to measure the movement component while brain activity is being measured using fMRI. The actual experiments are conducted at MCW. For some experiments the ankle is physically actuated to move, often in synergy with the subject thinking about moving at the same frequency (1 Hz). A related project involving Mathew Cole and Yayha Bahlool is developing diffusion tensor imaging techniques to quantify the spinal tract loss following spinal cord injury. In another project, Aaron Siminski, a doctoral student in the nearby Neuro-Evaluation Lab directed by Dr. Scheidt, has developed a non-ferrous, single-joint robotic manipulandum for use in studying human motor behavior and control while concurrently imaging the brain with fMRI techniques. His focus is on feedback stabilization of wrist posture, and he has collected preliminary data exploring the role of sensory information during posture control in patients recovering from hemiparetic stroke.
Intelligent Telerehab Assistants for Neurorehab - MedPredict:
For Intelligent Telerehabilitation Assistants (ITA's) to be effective as intelligent user assistants or rehab prognosis predictors, the ITA needs to have context-awareness, i.e. a continually updated estimate of the state of the biosystem of the client that is based on any new information plus a-priori expectations of healing processes. To address this need, Yu "Fish" Wang, a doctoral student, has developed an event-driven dynamic recurrent neurofuzzy framework to predict prognosis in neurorehabilitation based on available evidence. In this model framework, system behavior is a function of both spontaneous recovery mechanisms and a sequence of therapeutic interventions, implemented in the format of fuzzy rules and classic non-linear models of dynamic bio-processes which are based on the available evidence and the scientific literature. The overall system is composed of four layers: input, rule-state, output and outcome. Many linguistic variables (e.g., age, severity of stroke, medication, exercise) may have an effect on the outcome of rehabilitation. These variables are treated as inputs to this system, of four forms: facts, contexts, therapeutic interventions, and pharmochologic interventions. The states are normally estimates of the patient's degree of impairment and/or physiological status. Classic mathematical models are employed prior to the fuzzy reasoning when appropriate, such as a pharmacokinetic model that predicts concentrations and effects. To date, effect-state relations have been developed for 50 medications classically used for key areas of neurorehabilitation, based on collaboration with postdoctoral trainee Dr. Nicole Sirota and physical medicine faculty from MCW. Outputs of MedPredict are predicted measures of performance and/or capabilities, such as rater scores for commonly used instruments in neurorehabilitation. Predicted model outcomes are overall performance measures of patient that are to be maximized or minimized, similar to optimization “performance criteria. ” One aim is for the user to be able to explore different intervention plans and have the fuzzy expert system provide a “prognosis” as predicted outcomes, with each outcome presented as a curve that is a function of time (e.g., prognosis with BOTOX and exercise interventions), but also to suggest the best treatment strategy under the supervision of the user by comparing the outcomes of different intervention plans.
There are many other projects related to neurorehabilitation that are taking place in the laboratories surrounding this "flagship" laboratory. For a brief review, see the list of graduate student research projects.
|
|
