MSMCNet: Differential context drives accurate localization and edge smoothing of lesions for medical image segmentation.

Medical image segmentation plays a crucial role in clinical assistance for diagnosis. The UNet-based network architecture has achieved tremendous success in the field of medical image segmentation. However, most methods commonly employ element-wise addition or channel merging to fuse features, resulting in smaller differentiation of feature information and excessive redundancy. Consequently, this leads to issues such as inaccurate lesion localization and blurred boundaries in segmentation. To alleviate these problems, the Multi-scale Subtraction and Multi-key Context Conversion Networks (MSMCNet) are proposed for medical image segmentation. Through the construction of differentiated contextual representations, MSMCNet emphasizes vital information and achieves precise medical image segmentation by accurately localizing lesions and enhancing boundary perception. Specifically, the construction of differentiated contextual representations is accomplished through the proposed Multi-scale Non-crossover Subtraction (MSNS) module and Multi-key Context Conversion Module (MCCM). The MSNS module utilizes the context of MCCM coding and redistribute the value of feature map pixels. Extensive experiments were conducted on widely used public datasets, including the ISIC-2018 dataset, COVID-19-CT-Seg dataset, Kvasir dataset, as well as a privately constructed traumatic brain injury dataset. The experimental results demonstrated that our proposed MSMCNet outperforms state-of-the-art medical image segmentation methods across different evaluation metrics.

Deep learning-based lung image registration: A review.

Lung image registration can effectively describe the relative motion of lung tissues, thereby helping to solve series problems in clinical applications. Since the lungs are soft and fairly passive organs, they are influenced by respiration and heartbeat, resulting in discontinuity of lung motion and large deformation of anatomic features. This poses great challenges for accurate registration of lung image and its applications. The recent application of deep learning (DL) methods in the field of medical image registration has brought promising results. However, a versatile registration framework has not yet emerged due to diverse challenges of registration for different regions of interest (ROI). DL-based image registration methods used for other ROI cannot achieve satisfactory results in lungs. In addition, there are few review articles available on DL-based lung image registration. In this review, the development of conventional methods for lung image registration is briefly described and a more comprehensive survey of DL-based methods for lung image registration is illustrated. The DL-based methods are classified according to different supervision types, including fully-supervised, weakly-supervised and unsupervised. The contributions of researchers in addressing various challenges are described, as well as the limitations of these approaches. This review also presents a comprehensive statistical analysis of the cited papers in terms of evaluation metrics and loss functions. In addition, publicly available datasets for lung image registration are also summarized. Finally, the remaining challenges and potential trends in DL-based lung image registration are discussed.

Construction of prediction model for prognosis of uterine corpus endometrial carcinoma based on pyroptosis gene.

PURPOSE: To assess the expression of genes that are relevant to pyroptosis and the relationship between these genes and prognosis in uterine corpus endometrial carcinoma (UCEC). METHODS: The research identifies 16 pyroptosis regulators with different expressions in normal endometrium and UCEC. In accordance with the differentially expressed genes (DEGs), the various kinds of UCEC are classified into two sub-types. With the help of the Cancer Genome Atlas (TCGA), the prognostic value of all pyroptosis-related genes for survival was assessed, and a multigene model has constructed accordingly. Ten genes were modeled by applying the minimum criteria for determining risk score selection (LASSO) Cox regression method. Meanwhile, by referring to the TCGA atlas, UCEC patients were divided into the high- and low-risk subgroups. The effects of the gene with significant differences on the proliferation of two cancer cells were also verified. RESULTS: The survival rate of UCEC cases with higher risk was higher than that with lower risk (P < 0.001). Through the median risk score of TCGA atlas, UCEC cases were ranked as patients with higher risk and patients with lower risk. The low risk has a significant relationship with the prolongation of overall survival (OS) (p = 0.001) in the low-risk subgroup. Moreover, the KEGG and gene ontology (GO) enrichment models indicated that among the patients in the high-risk subgroup, their immune-related genes were concentrated but with decreased immune status. CONCLUSION: The apoptosis-related genes are crucial for the immunity of tumors and may forecast the prognosis of UCEC.

Group Membership Modulates Fairness Consideration Among Deaf College Students-An Event-Related Potential Study.

Group interaction is an essential way of social interaction and plays an important role in our social development. It has been found that when individuals participate in group interactions, the group identity of the interaction partner affects the mental processing and behavioral decision-making of subjects. However, little is known about how deaf college students, who are labeled distinctly different from normal hearing college students, will react when facing proposers from different groups in the ultimatum game (UG) and its time course. In this study, we recruited 29 deaf college students who played the UG in which they received extremely unfair, moderately unfair, or fair offers from either outgroup members (normal hearing college students) or ingroup members (deaf college students), while their brain potentials were recorded. The behavioral results showed that group membership did not impact the acceptance rate of deaf college students. But, event-related potential (ERP) analysis demonstrated an enhanced feedback-related negativity (FRN) elicited by ingroup members compared to outgroup members. Importantly, we found that under fairness conditions, deaf college students induced more positive P2 and P3 facing ingroup members compared to outgroup members. Our results demonstrated that group membership may modulate the performance of deaf college students in the UG and the existence of ingroup bias among deaf college students. This provides some evidence for the fairness characteristics of special populations, so that to improve the educational integration of colleges and universities.

Prolonged post-stimulation response induced by 980-nm infrared neural stimulation in the rat primary motor cortex.

The post-stimulation response of neural activities plays an important role to evaluate the effectiveness and safety of neural modulation techniques. Previous studies have established the capability of infrared neural modulation (INM) on neural firing regulation in the central nervous system (CNS); however, the dynamic neural activity after the laser offset has not been well characterized yet. We applied 980-nm infrared diode laser light to irradiate the primary motor cortex of rats, and tungsten electrode was inserted to record the single-unit activity of neurons at the depth of 800-1000 µm (layer V of primary motor cortex). The neural activities were assessed through the change of neural firing rate and firing pattern pre- and post-stimulation with various radiant exposures. The results showed that the 980-nm laser could modulate the firing properties of neurons in the deep layer of the cortex. More neurons with post-stimulation response (78% vs. 83%) were observed at higher stimulation intensity (0.803 J/cm2 vs. 1.071 J/cm2, respectively). The change of firing rate also increased with radiant exposures increasing, and the response lasted up to 4.5 s at 1.071 J/cm2, which was significantly longer than the theoretical thermal relaxation time. Moreover, the increasing Fano factors indicated the irregularity firing pattern of post-stimulation response. Our results confirmed that neural activity maintained a prolonged post-stimulation response after INM, which may provide necessary measurable data for optimization of INM applications in CNS.

Inhibition of cortical neural networks using infrared laser.

The aim of the present study is to optimize parameters for inhibiting neuronal activity safely and investigating thermal inhibition of rat cortex neural networks in vitro by continuous infrared (IR) laser. Rat cortex neurons were cultured on multi-electrode arrays until neural networks were formed with spontaneous neural activity. Neurons were then irradiated to inhibit the activity of the networks using different powers of 1550 nm IR laser light. A finite element heating model, calibrated by the open glass pipette method, was used to calculate temperature increases at different laser irradiation intensities. A damage signal ratio (DSR) was evaluated to avoid excessive heating that may damage cells. The DSR predicted that cortex neurons should be safe at temperatures up to 49.6°C for 30 seconds, but experiments suggested that cortex neurons should not be exposed to temperatures over 46°C for 30 seconds. Neural response experiments showed that the inhibition of neural activity is temperature dependent. The normal neural activity could be inhibited safely with an inhibition degree up to 80% and induced epileptiform activity could be suppressed. These results show that continuous IR laser radiations provide a possible way to safely inhibit the neural network activity.