Exploring Cerebrospinal Fluid Dynamics in Perivascular Spaces via Numerical Simulations and Convolutional Long Short-Term Memory Models

The glymphatic system, crucial for brain waste clearance, involves cerebrospinal fluid (CSF) flow through spaces surrounding blood vessels, known as perivascular spaces (PVS). The dysfunction of the glymphatic system is linked to neurodegenerative diseases such as Chronic traumatic encephalopathy and Alzheimer's disease. Despite its importance, the fluid dynamics of this recently discovered system has remained relatively understudied. In this study, we used numerical simulations and a Convolutional Long Short-Term Memory (ConvLSTM) model to study the fluid dynamics of CSF inside PVS, and the effect of different factors including blood pulsation and PVS size on CSF flow. Unsteady deforming mesh simulations are used to generate a comprehensive dataset and consequently a ConvLSTM model is used to predict CSF velocity fields, marking its first application in this context. The results show that ConvLSTM effectively captures spatiotemporal patterns in CSF flow, providing insights into the mechanical behavior of the system. This approach advances predictive modeling in neurodegenerative disease research, offering potential for improved diagnostic and therapeutic strategies.