Multiplexed optical fiber sensors for dynamic brain monitoring

Abstract

Dynamic brain monitoring can mitigate traumatic brain injury (TBI) deterioration and enable precise treatment. Despite many studies on brain monitoring systems for neuroscience applications, current technologies are limited in providing a continuous and real-time readout of multiple brain biomarkers simultaneously because of limited sensor performance and signal interference. Regression modeling, being a subfield of machine-learning (ML) algorithm, offers great advantages in signal enhancement and prediction. However, studies on ML-integrated optical fiber sensors for precise brain monitoring have been rarely reported. Here, a multiplexed optical fiber sensor regulated by regression algorithms is developed for the dynamic monitoring of brain pH, temperature, dissolved oxygen, and glucose levels. The proposed sensor has demonstrated excellent sensing abilities and can perform dynamic monitoring of TBI stages in ex vivo brain models. The results indicate the capability of the multiplexed optical fiber sensor for continuous brain physiology reflection, suggesting a promising prospect for clinical applications.