Recovery of function after severe spinal cord injury (SCI) has not yet been achieved, producing undue burden on patients, their families and society as a whole. The long-term goal of this work is to develop interventions for those subjected to severe neurological injury or disease that restore function. The goal of this project is to develop novel machine learning approaches that are able to control functional electrical stimulation (FES) into the spinal cord to restore function. Electrical stimulation will be combined with physical therapy to work to optimize recovery. The investigator hypothesizes that combinations of traditional physical therapy, known to improve functional recovery after severe SCI, and the proposed FES will significantly improve outcome over either alone. It is important to understand this interaction because, in the case of FES, maintaining muscle mass, bone density and overall cardiovascular health through physical therapy is necessary. This is especially true for the case of those with paraplegia (loss of use of both legs) who desire to walk using their own legs. The Intellectual Merit of this work is that there are currently on-going studies in humans that use brain-derived signals to augment physical therapy after severe neurological injury or disease. The studies planned are expected to provide insight into the mechanisms that support improved outcome in order to provide guidance into how best to use these brain-derived signals. To date, these signals have been almost exclusively applied to restoration of upper limb function. The planned work is significant because it will expand these capabilities to lower limb functions. Restoring lower limb function after SCI is distinctly different from restoring upper limb function because much of the lower limb movements are rhythmic locomotion and brain control of lower limbs is more often associated with adjustments to these rhythmic movements (e.g. avoiding obstacles) than that of upper limbs. Moreover, this work will engage a new generation of girls and young women in neuroscience and engineering through workshops, undergraduate research opportunities and graduate thesis work. This will be accomplished by performing one-on-one mentorships with members of the Society for Women Engineers to connect female engineering students with research projects of interest to them.
The goal of this project is to develop Brain Machine Interface (BMI) controlled functional electrical stimulation (BMI-FES) that can account for the impact of physical therapy on the encoding of motor control in the brain using a rat model of complete mid-thoracic spinal transection. The driving hypothesis is that combinations of traditional physical therapy, known to improve functional recovery after severe spinal cord injury (SCI), and BMI-FES will significantly improve outcome over either alone. To test this hypothesis, the research plan is organized under three objectives. The FIRST Objective is to identify the impact of physical therapy on learning BMI. The BMI task for the SCI animal model is to learn to control a platform that is being tilted to the right or the left, which involves postural control and bilaterally engages the cortex. Initial recordings will be used to parameterize the decoder based on a peristimulus time histogram (PSTH)-based classifier developed by the PI. BMI Performance (functional recovery and limb kinetics) will be compared without and with physical therapy (motorized bike and treadmill locomotion) over a 12 week time frame. The expected outcome is that BMI with therapy will improve BMI Performance because therapy will create a new cortical circuit that can encode for movements of the tilt platform better than the cortical circuit that develops after SCI in the absence of therapy. The SECOND Objective is to develop a dynamical systems model of BMI-FES. Using systems identification analysis, a dynamic model of motor cortex will be developed to identify the state-space within which a linear model is sufficient. Data recorded from OBJ 1 will be used to develop a first pass of the model system. In the second pass, neural data will be recorded while a broad range of stimulations are delivered via the epidural stimulator (EES) that excite the neural system. The recorded responses across electrodes will be used to extract information about the structure of the system under stimulation, which allows complete specification of a linear time-invariant (LTI) system. The expected outcome is that a single linear model can account for all of the tilts and relevant EES stimulations tested within a given recording session. Model parameters are expected to change over time, but as plasticity in response to therapy and BMI performance in the task stabilize, that a single model will emerge. The THIRD Objective is to design a closed-loop control system for BMI-FES that accommodates plasticity from three sources: 1) physical therapy, 2) learning BMI control and 3) somatosensory feedback from the response of the rat's sensorimotor system to FES. Experiments will determine if BMI controlled FES can improve functional recovery of the animal in the tilt task, on the treadmill and in the open space, and will assess the impact of therapy on this functional recovery. The expected outcome is that controlled FES will improve functional recovery more than no stimulation or tonic stimulation and that BMI-FES will be more effective in animals that receive therapy compared to those that do not. This is important because it would suggest that the dynamical system framework is an effective method to generate stimulation patterns and that plasticity induced by therapy is synergistic with plasticity induced by learning the BMI control.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
