Investigating and Augmenting Cholinergic Signaling for Enhanced Neuroprosthetic Performance

About This Grant

PROJECT SUMMARY/ABSTRACT Neuroprosthetic interfaces provide paralyzed patients increased agency and independence. These intracortical implants interface with cortex at the cellular level, readout neural activity in real-time, and transform this activity into the actuation of computer cursors, robotic prostheses, and speech generators. Recent advances in the software underlying these devices have led to highly accurate brain-computer interfaces (BCIs) for improving health outcomes in patients with severe spinal cord injuries, neurodegenerative diseases, and stroke. However, there are still many fronts that require improvements. Namely, these devices have a learning curve on the order of days to weeks very similar to that of learning a new skill. In addition, even when accurate performance is reached, the natural drift of neural recordings necessitates system recalibrations, which requires the patient to briefly re-learn the new algorithmic parameters. Here, we propose that these learning periods can be reduced by augmenting the neural signaling pathways the brain uses to learn new BCI skills. A structure deep in the brain called the basal forebrain is known to play a strong role in learning new tasks, such as new motor movements. This region extends axons from cholinergic neurons up to cortical regions, where they are extremely active during the learning phase of new tasks, enhancing the local plasticity of cortical circuits. Recently, studies have shown that manipulating these axons can influence learning and task performance. For example, stimulating the vagus nerve can increase cholinergic activity and when timed properly can lead to faster learning rates and more accurate performance on a variety of sensorimotor tasks. In this work, we propose that the cholinergic neurons in the basal forebrain also play a major role in the learning and control of neuroprosthetic interfaces, particularly when these devices are implanted in motor cortex. To investigate this hypothesis, we will use a state-of-the-art combination of multiphoton imaging and transparent electrode arrays to record cholinergic axon activity near the implanted BCI as rodents learn to perform neuroprosthetic skills. In addition, we propose that vagus nerve stimulation can be used to excite cholinergic activity and reinforce neuroprosthetic learning, reducing the training time needed for animals to achieve expert-level control. Importantly, vagus nerve stimulation is an approved human treatment. Therefore, not only will our work reveal a fundamental circuit mechanism critical for BCI control, but our research opens the door to new applications of vagus nerve stimulation for enhancing neuroprosthetic performance in humans.