Yonsei University · 医学
Professor Byung-Hoon Kim's research lab specializes in computational neuroscience and neuroimaging analytics, focusing on leveraging machine learning and graph neural networks to decode brain network dynamics from functional MRI data. The lab investigates functional connectivity patterns in psychiatric and neurological disorders, with particular emphasis on identifying neuroimaging biomarkers for conditions such as bipolar disorder, social anxiety, and glioma radiogenomics. A key research direction involves developing explainable AI frameworks—especially Graph Isomorphism Networks (GIN)—to enhance the interpretability of brain connectivity models in a neuroscience-informed manner.
Figures are computed from collected data and may differ slightly.
Graph neural networks (GNN) rely on graph operations that include neural network training for various graph related tasks. Recently, several attempts have been made to apply the GNNs to functional magnetic resonance image (fMRI) data. Despite recent progresses, a common limitation is its difficulty to explain the classification results in a neuroscientifically explainable way. Here, we develop a framework for analyzing the fMRI data using the Graph Isomorphism Network (GIN), which was recently p
Abstract Background The eukaryotic translation initiation factor 3 (eIF3) has multiple roles during the initiation of translation of cytoplasmic mRNAs. How individual subunits of eIF3 contribute to the translation of specific mRNAs remains poorly understood, however. This is true in particular for those subunits that are not conserved in budding yeast, such as eIF3h. Results Working with stable reporter transgenes in Arabidopsis thaliana mutants, it was demonstrated that the h subunit of eIF3 co
The island-like distribution of the oxygen functional groups of graphene oxide was identified by deposition of TiO<sub>2</sub> on the graphene oxide surface using low-temperature atomic layer deposition.
We have obtained high pressure H(2) isotherms with respect to the interlayer distance of multilayered graphene oxide (GO) modulated by thermal annealing. The maximum storage capacity is 4.8 (0.5) wt% at 77 K (298 K) and at 9.0 MPa pressure. We found the optimum GO interlayer distance for maximum H(2) uptake at 6.5 Å, similar to the predicted distances from first-principles calculations for graphite materials. Our results reveal that multilayered GO can be a practical material of choice to allow
Functional connectivity (FC) between regions of the brain can be assessed by the degree of temporal correlation measured with functional neuroimaging modalities. Based on the fact that these connectivities build a network, graph-based approaches for analyzing the brain connectome have provided insights into the functions of the human brain. The development of graph neural networks (GNNs) capable of learning representation from graph structured data has led to increased interest in learning the g
Studies of the interaction between hydrogen and graphene have been increasingly required due to the indispensable modulation of the electronic structure of graphene for device applications and the possibility of using graphene as a hydrogen storage material. Here, we report on the behaviour of molecular hydrogen on graphene using the gate voltage-dependent resistance of single-, bi-, and multi-layer graphene sheets as a function of H₂ gas pressure up to 24 bar from 300 K to 345 K. Upon H₂ exposu
Thermally reducible pyroprotein-based electronic textiles (e-textiles) are fabricated using graphene oxide and a pyroprotein such as cocoon silk and spider web without any chemical agents. The electrical conductivity of the e-textile is 11.63 S cm<sup>-1</sup> , which is maintained even in bending, washing, and temperature variation.
COX-2 overexpression in PC is correlated with a decrease in apoptosis and an increase in angiogenesis. Chronic inflammation in BPH causes an overexpression of COX-2, which induces the increased expression of Bcl-2 and VEGF. It is likely that chronic inflammation plays a role in the intermediate step of carcinogenesis in the prostate.
Adequate treatment of respiratory comorbidities and prevention of cardiovascular complications might be the critical factors in reducing postoperative mortality in elderly patients undergoing hip fracture surgery.
O6-methylguanine-DNA methyl transferase (MGMT) methylation prediction models were developed using only small datasets without proper external validation and achieved good diagnostic performance, which seems to indicate a promising future for radiogenomics. However, the diagnostic performance was not reproducible for numerous research teams when using a larger dataset in the RSNA-MICCAI Brain Tumor Radiogenomic Classification 2021 challenge. To our knowledge, there has been no study regarding the
Open papers in the app to read, cite, and organize with AI.
