Modulation of Epileptic Circuit
Neuropsychiatric disorders like Obsessive Compulsive Disorder (OCD) and depression disrupt the lives of millions of people in the world. State-of-the-art methods for treatment are mainly based on behavior therapy and medications, which may not be widely effective for various types of disorders and can be even risky to patients. The Rice team is performing modeling and simulations for both the low intensity focused ultrasound (LIFU) and millimeter wave (mmWave) arms of the non-invasive DBS project. For LIFU, work will involve detailed computational modeling of how the stimulation transducer pulses traverse biological media (e.g., skull, brain) together with how pulses from multiple transducers interact to produce a focused modulation of the targeted anatomic structure in a safe, controlled manner. Parallel modeling with similar aims will be conducted for the mmWave approach. Work will also involve validating the modeling for LIFU and mmWave approaches in the form of hardware implementation using phantom and animal models.
Collaborators: Wayne Goodman M.D., Raymond Cho M.D., and Sameer Sheth M.D.
Current Student Researchers: Fatima Ahsan and Elise Gibney
Funded By:
Using EmDMD to Predict Seizures
The underlying spatio-temporal mechanism that leads to the formation of seizures in the brain has been a critical topic of studies for decades. Different techniques have been proposed to extract the dynamics of epileptic recordings that are involved in seizure formation. In this project, we introduce a data-driven tool called the Dynamic Mode Decomposition (DMD) with time-delay embedding (EmDMD) to extract the underlying spatio-temporal dynamics of seizure formation. This technique will enable us to represent recordings of a seizure on a graph and focus on the dynamical changes of these graphs, capturing key features of epileptic recordings in our attempt to predict and prevent seizures. Using EmDMD, we can observe changes in the extracted features of the graphs sometime before the start of the seizure. These changes enable us to distinguish the pre-seizure state from the regular activity of an epileptic brain to predict seizures a few seconds before their start in order to activate stimulation to prevent the onset of a seizure. Thus, we use the extracted information to develop a seizure prediction in epilepsy treatment as a closed-loop control system of the embedded system.
Collaborators: Lan Luan, Chong Xie and Sameer Sheth M.D.
Current Student Researcher: Dorsa E.P. Moghaddam
Funded By:
