Dynamic cerebral blood flow assessment based on electromagnetic coupling sensing and image feature analysis

Abstract

Dynamic assessment of cerebral blood flow (CBF) is crucial for guiding personalized management and treatment strategies, and improving the prognosis of stroke. However, a safe, reliable, and effective method for dynamic CBF evaluation is currently lacking in clinical practice. In this study, we developed a CBF monitoring system utilizing electromagnetic coupling sensing (ECS). This system detects variations in brain conductivity and dielectric constant by identifying the resonant frequency (RF) in an equivalent circuit containing both magnetic induction and electrical coupling. We evaluated the performance of the system using a self-made physical model of blood vessel pulsation to test pulsatile CBF. Additionally, we recruited 29 healthy volunteers to monitor cerebral oxygen (CO), cerebral blood flow velocity (CBFV) data and RF data before and after caffeine consumption. We analyzed RF and CBFV trends during immediate responses to abnormal intracranial blood supply, induced by changes in vascular stiffness, and compared them with CO data. Furthermore, we explored a method of dynamically assessing the overall level of CBF by leveraging image feature analysis. Experimental testing substantiates that this system provides a detection range and depth enhanced by three to four times compared to conventional electromagnetic detection techniques, thereby comprehensively covering the principal intracranial blood supply areas. And the system effectively captures CBF responses under different intravascular pressure stimulations. In healthy volunteers, as cerebral vascular stiffness increases and CO decreases due to caffeine intake, the RF pulsation amplitude diminishes progressively. Upon extraction and selection of image features, widely used machine learning algorithms exhibit commendable performance in classifying overall CBF levels. These results highlight that our proposed methodology, predicated on ECS and image feature analysis, enables the capture of immediate responses of abnormal intracranial blood supply triggered by alterations in vascular stiffness. Moreover, it provides an accurate diagnosis of the overall CBF level under varying physiological conditions.