Decoding the Neural Representations of Movement Minimally Invasively via Functional Ultrasound

California Institute Of Technology, Pasadena CA

Investigators

Abstract

PROJECT SUMMARY/ABSTRACT Neurological conditions such as spinal cord injury and ALS severely impair patient function, leaving many patients paralyzed. Brain-machine interfaces (BMIs) offer patients the ability to restore their autonomy by translating brain signals into actionable outputs. However, current neurorecording techniques commonly require invasive surgical implantation and are constrained by technical limitations, sacrificing spatiotemporal resolution for field of view and vice versa. These constraints hinder our understanding of motor variable encoding in the primary motor cortex (M1) and posterior parietal cortex (PPC). Functional ultrasound (fUS), a novel neuroimaging technique, offers a solution by minimally invasively providing high spatial resolution (<300 um), sensitivity (~100 ms), and large plane of view (several cm) for recording neural activity at a mesoscopic scale from outside the dura. This proposal aims to leverage fUS to clarify the representations of motor variables such as movement effector and trajectory in M1 and PPC and develop a minimally invasive motor BMI that can restore functionality to patients with motor disabilities. Specific Aim 1 seeks to identify the mesoscopic neural representations of motor execution and planning across different effectors in M1 and PPC respectively, key areas commonly targeted in motor BMIs. To do so, we will record fUS data from human participants with sonolucent cranial implants as they perform instructed delay motor tasks using various movement effectors such as fingers, arm, leg, and face. This will generate a mesoscopic mapping of effector representation across both M1 and PPC, bridging microscopic and macroscopic motor encoding. Specific Aim 2 builds on these findings and investigates the neural encoding of movement trajectory in M1 and PPC to develop a real-time fUS-based motor BMI capable of predicting movement direction for cursor control. We will first use an instructed delay task using 8 different directions of hand and arm movement to discern the basic encoding of movement trajectory in M1 and PPC and then decode trajectory on a continuous scale, using a target tracking task for real-time BMI cursor control. Specific Aim 3 will examine the relationship between neurovascular signal and single-neuron activity by comparing fUS signal to prior recorded single-unit activity from Utah electrode arrays. This will help validate our fUS findings in Aim 1 and Aim 2, further establishing fUS as a robust neuroimaging technique that can be used in place of electrophysiology, and elucidate the transfer functions between single-neuron activity and neurovascular changes in cerebral blood volume. Altogether, this proposal will provide a better understanding of motor variable encoding in the cortex, providing crucial information that can be used for more effective BMI design. This will validate the abilities of fUS both as a neurorecording technique and as a minimally invasive BMI alternative. These findings have the potential to expand the accessibility of BMIs and enable further development of fUS for more widespread clinical applications such as chronic neuropsychiatric monitoring and closed-loop neuromodulation.