https://scholar.dgist.ac.kr/handle/20.500.11750/1194 Wed, 05 Aug 2026 18:34:08 GMT 2026-08-05T18:34:08Z https://scholar.dgist.ac.kr/handle/20.500.11750/60591 Title: Self-Refining Segment Anything Model for Nuclei Segmentation as Contrastive Learning Approach to Label-Efficient Pathological Imaging Author(s): Nam, Siwoo; Park, Sang Hyun Abstract: Background/Objectives: Precise nuclei instance segmentation is a prerequisite for reliable digital pathology, yet the scarcity of pixel-level annotations remains a significant bottleneck for deep learning models. Methods: We propose a self-evolving framework for robust nuclei segmentation that uses only sparse point annotations, extending the Segment Anything Model (SAM). To overcome the limitations of static pseudo-labels, our method introduces a self-evolving labeling strategy via Exponential Moving Average (EMA), which adaptively refines learning targets. We also integrate instance-aware contrastive learning using point prompts as spatial anchors and implement a consensus-based filtering mechanism between prompt-guided and prompt-free decoders. Results: Extensive evaluations on CPM17, MoNuSeg, and the challenging CoNSeP datasets demonstrate that our framework achieves state-of-the-art performance across various backbones, including ViT-B and ViT-H. Conclusions: By enabling a seamless transition from general-purpose foundation models to specialized histopathology experts, this self-refining approach delivers a highly efficient, accurate solution for automated diagnostic workflows in clinical settings. Tue, 31 Mar 2026 15:00:00 GMT https://scholar.dgist.ac.kr/handle/20.500.11750/60591 2026-03-31T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60587 Title: A laser-induced graphene neural probe for multiplexed multimodal readout via site-selective functionalization Author(s): Lee, Seungjun; Kim, Giheon; Lee, A-Hyeon; Kang, Dae-Si; Kim, Jeong Hun; Koo, Ja Wook; Lee, Suk Won; Chou, Namsun; Shin, Hyogeun Abstract: Simultaneous monitoring of multiple neurochemicals in the brain is critical for elucidating complex neuronal processes, yet it remains technically challenging because of limited spatial selectivity, device complexity, and signal cross-interference. Existing multiplexed neurochemical probes often depend on physically separated sensors or intricate microfluidic architectures, which constrain spatial resolution, scalability, and practical integration for in vivo applications. Here, we present a flexible neural probe based on laser-induced graphene (LIG) that enables multiplexed, multimodal readout within a compact, scalable platform. Using a site-selective, sequential enzyme functionalization strategy, we created independent sensing interfaces for glucose, lactate, and dopamine on closely spaced electrodes while minimizing cross-contamination and interference between adjacent channels. The fabricated probe exhibited selective, concentration-dependent electrochemical responses in vitro and successfully detected stimulus-dependent dopamine release in neuronal cell models. In vivo experiments further demonstrated stable sensing performance after implantation, enabling real-time, multiplexed monitoring of neurochemical dynamics in the mouse medial prefrontal cortex. In addition, the probe supported simultaneous recording of neurochemical signals and electrophysiological activity, highlighting its potential for multimodal neural interfacing in vivo. Collectively, this work establishes a practical and scalable approach to high-density multiplexed neurochemical sensing and provides a versatile platform for investigating dynamic neurochemical signaling and its relationship to brain function. Sat, 31 Oct 2026 15:00:00 GMT https://scholar.dgist.ac.kr/handle/20.500.11750/60587 2026-10-31T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60586 Title: MetaRT: structure-embedded stacked ensemble machine learning prediction of retention time of short peptides Author(s): Mahmood, Raisul Awal; Mahin, Md Ibrahim Shikder; Alam, Rafiqul; Jeong, Seungjun; Kang, Hyungu; Cho, Daeheum; Kim, Sunghwan Abstract: Predicting peptide retention time (RT) remains a significant challenge, particularly when training data is limited. In this study, we present MetaRT, a stacked-ensemble machine learning framework designed to predict the RTs from small dataset of hydrophobic peptides. Peptides composed of hydrophobic amino acids-phenylalanine (F), isoleucine (I), methionine (M), and tryptophan (W) were synthesized, and their experimental RTs were measured from the mixture entities. MetaRT utilizes a graph convolutional network (GCN) to extract structural features from the peptide sequences. The MetaRT model architecture employed multiple base learners, integrating the outputs through a meta-learner optimized via hyperparameter tuning and 3-fold cross-validation. Besides, the performance of MetaRT was compared to three ensemble methods - weight averaging, bagging, and boosting. The results demonstrated that structure-based MetaRT outperformed both base learners and the ensemble models, achieving a lower root mean square error (RMSE) of 0.08 and a maximum RT deviation of approximately 1.4 min. Compared to the prediction performance on molecular descriptors inclusion, the structure-guided model consistently performed well in terms of RMSE. Notably, MetaRT accurately predicted the RTs of sequence isomers by leveraging the structural features, with deviations ranging from 0.2 to 1.3 min. In contrast, descriptor-based model showed increased prediction error for the isomeric sequences. For peptides with lower hydrophobicity that were not included in the training data, the structure-based predictions led to the maximum deviation of 4.9 min from the experimental RTs. The entire predicted RTs were subsequently validated by linear regression analyses with the corresponding experimental values. These findings highlight the potential of MetaRT as a structure-based predictive tool for improving RT prediction accuracy, especially in data-limited scenarios. Future work will focus on enhancing the robustness of MetaRT by incorporating a wider variety of peptide classes to further refine its predictive capabilities. Tue, 30 Jun 2026 15:00:00 GMT https://scholar.dgist.ac.kr/handle/20.500.11750/60586 2026-06-30T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60556 Title: PM2.5 impairs gliovascular coupling via endothelial AHR-mitochondrial signaling in mice Author(s): Kim, Kyu-sung; Kim, Dong-im; Hwang, Sungsu; Park, Inyeong; Jeon, Min-tae; Kim, Yujung; Son, Suhyeon; Lee, Jaehyeok; Park, Kyemyung; Lee, Kyuhong; Kim, Dogeun Abstract: Particulate matter (PM2.5) is a pervasive air pollutant increasingly linked to neurovascular dysfunction, but the cellular mechanisms remain unclear. We identify the aryl hydrocarbon receptor (AHR) as a key endothelial sensor of PM2.5 that initiates mitochondrial stress and Parkin-dependent mitophagy. Across complementary inhalation and intratracheal instillation models, integrated with spatial transcriptomics, high-resolution imaging, and in vitro assays, endothelial mitochondrial injury and oxidative stress constricted cerebral vessels and reduced perfusion. These vascular insults propagated to astrocytes, where calmodulin-dependent mislocalization of aquaporin-4 (AQP4) disrupted perivascular water homeostasis and glymphatic exchange. System-level consequences included dendritic degeneration, microglial activation, and hypoxic stress, with the hippocampus showing heightened vulnerability. Spatial transcriptomics resolved region-and cell type-specific injury and synaptic remodeling that bulk RNA sequencing failed to detect, while endothelial readouts evidenced canonical AHR engagement. Collectively, the data establish endothelial mitophagy as a metabolic checkpoint linking environmental particulate exposure to gliovascular dysfunction and impaired brain clearance, and nominate AHR signaling as a potential therapeutic target to preserve brain homeostasis under chronic air pollution. These mechanistic links provide a framework for interpreting epidemiological associations between PM2.5 exposure and neurodegenerative disease risk. Sat, 31 Jan 2026 15:00:00 GMT https://scholar.dgist.ac.kr/handle/20.500.11750/60556 2026-01-31T15:00:00Z