1. High-frequency spinal cord stimulation and neurophysiological mechanisms:
This project investigates high-frequency spinal cord stimulation (HF-SCS) effects on the primary sensorimotor cortex in sheep, aiming to unravel the mechanisms (using ECoG recordings) of pain modulation. We employ time-frequency analysis, functional connectivity metrics, spectral graph measures, evoked potential analysis, and machine learning algorithms to investigate how HF-SCS affects cortical activity. By understanding these neural mechanisms, we aim to inform therapeutic strategies for pain management. Sharing preprocessed datasets fosters collaboration and further research, bridging animal models and human applications to enhance clinical outcomes for pain treatment. This project aims to provide significant insights into the neural dynamics of pain and potential therapeutic interventions.
2. Unraveling Risk-Reward Processing in Movement Disorders:
Insights from Intracranial Electrophysiology Recently in collaboration with Caltech we unraveled the neural mechanisms (using EEG) underlying risk and reward processing (Man et al 2024, Nature Communicati0ns). Initial results point to the involvement of critical areas like the orbitofrontal cortex, insula, ventral striatum, and temporoparietal junction in processing risk and reward signals. To deepen our understanding, we're employing intracranial recordings within specific brain regions of patients with movement disorders. This approach will allow us to directly record from individual neurons, providing finer-grained insights into the neural dynamics governing decision-making. Ultimately, our research aims to pave the way for more targeted interventions and therapies for disorders associated with impaired decision-making processes.
3. Motor control and decoding the movement kinematics:
Motor control requires complex interplay between brain regions to produce behavior. In this project, we're utilizing intracranial electrode data alongside sensor data from joints study the intricacies of motor control mechanisms. Our goal is to model brain data to fit sensor data, enabling us to decode arm / wrist / finger/ shoulder position and kinematics during a reach task. While we've made promising strides using partial least square regression, achieving approximately 70% accuracy for the shoulder joint, our next step is to explore it using deep learning methods that could potentially yield even better results. With this approach and rich dataset, we anticipate uncovering valuable insights into how the brain coordinates movement.
4. Biophysical properties of neurons and behavior :
Neurons can be classified based on their firing patterns, biophysical and other properties. These classifications are important for understanding the diverse functional roles of neurons in the brain and how they contribute to various behaviors and cognitive processes. A detailed comparison of classified neurons, their biophysical properties, and their implications in patient behavior can be important step toward unraveling the neural circuits and oscillations gone wrong. A comprehensive mapping of such neurons across several patients may drive us in this direction. In long term, they can be employed for single trial decoding for different behaviors (in relation to different stimuli /tasks).