A Dual Parameter Synchronous Monitoring System of Brain Edema Based on the Reflection and Transmission Characteristics of Two-Port Test Network

Abstract

As a common secondary disease, edema after traumatic brain injury (TBI) can increase brain volume, resulting in elevated intracranial pressure (ICP), brain shift, and cerebral hernia, and can eventually lead to death. The real-time continuous monitoring of edema may significantly reduce mortality and disability. In this paper, a dual-parameter synchronous monitoring system of edema based on the reflection and transmission characteristics of the two-port test network was established; 15 rabbits were chosen to perform 24-h reflection phase shift (RPS) and transmission phase shift (TPS) simultaneously monitoring the experiments of brain edema. With the development of brain edema, the variation law of the RPS and TPS was investigated. Combined with the power amplitude spectrum and the principle of the two-port test network, the influence of frequency on the detection sensitivity of RPS and TPS was analyzed in detail, and the optimal detection frequency point was found. After that, the classification of three different degrees of edema is performed by the BP algorithm. In the animal experiment, the RPS showed a continuous increasing trend within time, and it presented the variation of (9.35910° ± 1.65702°), (12.60117° ± 2.30218°), and (16.33423° ± 2.11118°) after 6, 12, and 24 h, respectively. Meanwhile, the TPS showed a continuous downward trend with the variation of (-12.62555° ± 0.99441°), (-19.23976° ± 1.27488°), and (-27.26285° ± 2.62291°) after 6, 12, and 24 h, respectively. The RPS was negatively correlated with the TPS. The RPS and the TPS together as a recognition feature can achieve 100% accurate classification of three different brain edema severities. Based on these results, it can be concluded that the system established in this paper can monitor gradual increases in brain edema severity. Furthermore, neither the RPS nor the TPS can be set as the recognition feature alone to achieve the completely accurate classification, which shows the necessity of presenting two parameters in the monitoring process.