Electrical impedance tomography (EIT) has been widely used in biomedical research because of its advantages of real-time imaging and nature of being non-invasive and radiation-free. Additionally, it can reconstruct the distribution or changes in electrical properties in the sensing area. Recently, with the significant advancements in the use of deep learning in intelligent medical imaging, EIT image reconstruction based on deep learning has received considerable attention. This study introduces the basic principles of EIT and summarizes the application progress of deep learning in EIT image reconstruction with regards to three aspects: a single network reconstruction, deep learning combined with traditional algorithm reconstruction, and multiple network hybrid reconstruction. In future, optimizing the datasets may be the main challenge in applying deep learning for EIT image reconstruction. Adopting a better network structure, focusing on the joint reconstruction of EIT and traditional algorithms, and using multimodal deep learning-based EIT may be the solution to existing problems. In general, deep learning offers a fresh approach for improving the performance of EIT image reconstruction and could be the foundation for building an intelligent integrated EIT diagnostic system in the future.

Electrical impedance tomography (EIT) is low-cost and noninvasive and has the potential for real-time imaging and bedside monitoring of brain injury. However, brain injury monitoring by EIT imaging suffers from image noise (IN) and resolution problems, causing blurred reconstructions. To address these problems, a least absolute shrinkage and selection operator model is built, and a fast iterative shrinkage-thresholding algorithm with continuation (FISTA-C) is proposed. Results of numerical simulations and head phantom experiments indicate that FISTA-C reduces IN by 63.2%, 47.2%, and 29.9% and 54.4%, 44.7%, and 22.7%, respectively, when compared with the damped least-squares algorithm, the split Bergman, and the FISTA algorithms. When the signal-to-noise ratio of the measurements is 80–50 dB, FISTA-C can reduce IN by 83.3%, 72.3%, and 68.7% on average when compared with the three algorithms, respectively. Both simulation and phantom experiments suggest that FISTA-C produces the best image resolution and can identify the two closest targets. Moreover, FISTA-C is more practical for clinical application because it does not require excessive parameter adjustments. This technology can provide better reconstruction performance and significantly outperforms the traditional algorithms in terms of IN and resolution and is expected to offer a general algorithm for brain injury monitoring imaging via EIT.

No abstract

BackgroundReal-time detection of cerebral blood perfusion can prevent adverse reactions, such as cerebral infarction and neuronal apoptosis. Our previous clinical trial have shown that the infusion of therapeutic fluid can significantly change the impedance distribution in the brain. However, whether this alteration implicates the cerebral blood perfusion remains unclear. To explore the feasibility of monitoring cerebral blood perfusion, the present pilot study established a novel cerebral contrast-enhanced electrical impedance tomography (C-EIT) technique.Materials and methodsRabbits were randomly divided into two groups: the internal carotid artery non-occlusion (ICAN) and internal carotid artery occlusion (ICAO) groups. Both of groups were injected with glucose, an electrical impedance-enhanced contrast agent, through the right internal carotid artery under EIT monitoring. The C-EIT reconstruction images of the rabbits brain were analyzed according to the collected raw data. The paired and independent t-tests were used to analyze the remodeled impedance values of the left and right cerebral hemispheres within and between studied groups, respectively. Moreover, pathological examinations of brain were performed immediately after C-EIT monitoring.ResultsAccording to the reconstructed images, the impedance value of the left cerebral hemisphere in the ICAN group did not change significantly, whereas the impedance value of the right cerebral hemisphere gradually increased, reaching a peak at approximately 10 s followed by gradually decreased. In the ICAO group, the impedance values of both cerebral hemispheres increased gradually and then began to decrease after reaching the peak value. According to the paired t-test, there was a significant difference (P < 0.001) in the remodeling impedance values between the left and right hemispheres in the ICAN group, and there was also a significant difference (P < 0.001) in the ICAO group. According to the independent t-test, there was a significant difference (P < 0.001) of the left hemispheres between the ICAN and ICAO groups.ConclusionThe cerebral C-EIT proposed in this pilot study can reflect cerebral blood perfusion. This method has potential in various applications in the brain in the future, including disease progression monitoring, collateral circulation judgment, tumor-specific detection, and brain function research.