Enhancing Information Maximization With Distance-Aware Contrastive Learning for Source-Free Cross-Domain Few-Shot Learning.

Existing Cross-Domain Few-Shot Learning (CDFSL) methods require access to source domain data to train a model in the pre-training phase. However, due to increasing concerns about data privacy and the desire to reduce data transmission and training costs, it is necessary to develop a CDFSL solution without accessing source data. For this reason, this paper explores a Source-Free CDFSL (SF-CDFSL) problem, in which CDFSL is addressed through the use of existing pretrained models instead of training a model with source data, avoiding accessing source data. However, due to the lack of source data, we face two key challenges: effectively tackling CDFSL with limited labeled target samples, and the impossibility of addressing domain disparities by aligning source and target domain distributions. This paper proposes an Enhanced Information Maximization with Distance-Aware Contrastive Learning (IM-DCL) method to address these challenges. Firstly, we introduce the transductive mechanism for learning the query set. Secondly, information maximization (IM) is explored to map target samples into both individual certainty and global diversity predictions, helping the source model better fit the target data distribution. However, IM fails to learn the decision boundary of the target task. This motivates us to introduce a novel approach called Distance-Aware Contrastive Learning (DCL), in which we consider the entire feature set as both positive and negative sets, akin to Schrödinger's concept of a dual state. Instead of a rigid separation between positive and negative sets, we employ a weighted distance calculation among features to establish a soft classification of the positive and negative sets for the entire feature set. We explore three types of negative weights to enhance the performance of CDFSL. Furthermore, we address issues related to IM by incorporating contrastive constraints between object features and their corresponding positive and negative sets. Evaluations of the 4 datasets in the BSCD-FSL benchmark indicate that the proposed IM-DCL, without accessing the source domain, demonstrates superiority over existing methods, especially in the distant domain task. Additionally, the ablation study and performance analysis confirmed the ability of IM-DCL to handle SF-CDFSL. The code will be made public at https://github.com/xuhuali-mxj/IM-DCL.

Overexpression of EphB4 promotes neurogenesis, but inhibits neuroinflammation in mice with acute ischemic stroke.

Ischemic stroke is one of the most common diseases that has a high rate of mortality, and has become a burden to the healthcare system. Previous research has shown that EPH receptor B4 (EphB4) promotes neural stem cell proliferation and differentiation in vitro. However, little is known regarding its role in the neurogenesis of ischemic stroke in vivo. Thus, the present study aimed to verify whether EphB4 was a key regulator of neurogenesis in ischemic stroke in vivo. Cerebral ischemia was induced in C57BL/6J mice via middle cerebral artery occlusion (MCAO), followed by reperfusion. Immunofluorescence staining was performed to evaluate the effect of EphB4 on the neurogenesis in cerebral cortex. The levels of inflammatory cytokines were determined using an ELISA kit. The expression levels of ABL proto­oncogene 1, non­receptor tyrosine kinase (ABL1)/Cyclin D1 signaling pathway­related proteins were detected via western blotting. The current findings indicated that EphB4 expression was significantly increased in the cerebral cortex of MCAO model mice in comparison with sham­operated mice. Moreover, EphB4 appeared to be expressed in neural stem cells (Nestin+), and persisted as these cells became neuronal progenitors (Sox2+), neuroblasts [doublecortin (DCX)+], and eventually mature neurons [neuronal nuclei (NeuN)+]. Overexpression of EphB4 elevated the number of proliferating (bromodeoxyuridine+, Ki67+) and differentiated cells (Nestin+, Sox2+, DCX+ and NeuN+), indicating the promoting effect of EphB4 on the neurogenesis of ischemic stroke. Furthermore, EphB4 overexpression alleviated the inflammation injury in MCAO model mice. The expression levels of proteins­related to the ABL1/Cyclin D1 signaling pathway were significantly increased by the overexpression of EphB4, which suggested that restoration of EphB4 promoted the activation of the ABL1/Cyclin D1 signaling pathway. In conclusion, this study contributes to the current understanding of the mechanisms of EphB4 in exerting neurorestorative effects and may recommend a potential new strategy for ischemic stroke treatment.

MiR-1b up-regulation inhibits rat neuron proliferation and regeneration yet promotes apoptosis via targeting KLF7.

INTRODUCTION: MicroRNA (miRNA) is known to be involved in nerve injury. Our study aimed to identify the role and mechanism of miR-1b in rat neuron proliferation, regeneration and apoptosis. MATERIAL AND METHODS: Neurons were successfully separated and identified using a microscope and immunofluorescence staining of microtubule-associated protein 2 (MAP-2). The expressions of miR-1b and Krüppel-like factor 7 (KLF7) were detected by quantitative real-time polymerase chain reaction (qRT-PCR). Neuron viability and apoptosis were detected by MTT assay and flow cytometry, respectively. Neuron regeneration states were observed using a microscope and analysed by the ImageJ software. Expressions of C-caspase-3 and cell regeneration-related proteins (nerve growth factor [NGF], ciliary neurotrophic factor [CNTF] and brain-derived neurotrophic factor [BDNF]) were measured by Western blot. Target genes and potential binding sites of KLF7 and miR-1b were predicted by TargetScan 7.2 and confirmed by dual luciferase reporter assay. RESULTS: Neurons were identified as MAP-2-positive. Up-regulation of miR-1b reduced neuron viability and regenerative ability, promoted neuron apoptosis and C-caspase-3 expression, and down-regulated the expressions of cell regeneration-related proteins. KLF7 was the target gene of miR-1b. Overexpressed KLF7 rescued the effects of up-regulation of miR-1b on neuron viability, regeneration and apoptosis. Expressions of NGF, CNTF and BDNF were suppressed yet C-caspase-3 expression was up-regulated by miR-1b mimic, which was partially rescued by overexpressed KLF7. CONCLUSIONS: Up-regulation of miR-1b promoted rat neuron proliferation and regeneration yet inhibited apoptosis via targeting KLF7.