SimpleNMR: An interactive graph network approach to aid constitutional isomer verification using standard 1D and 2D NMR experiments

Despite progress in computer automated solutions, constitutional isomer verification by NMR using one- and two-dimensional data sets is still, in the main, a manual, user-intensive activity that is challenging for a number of reasons. These include the problem of simultaneously keeping track of the information from a number of separate NMR experiments and the difficulty of another researcher subsequently verifying the assignments made without having to independently repeat the whole analysis. This paper describes a graphical interactive approach that overcomes some of these problems. By using concepts used to visualise graph networks, we have been able to represent the NMR data in a manner that highlights directly the link between the different NMR experiments and the molecule of interest. Furthermore, by making the graph networks interactive, a user can easily validate and correct the assignment and understand the decisions made in arriving at the solution. We have developed a usable proof-of-concept computer program, 'simpleNMR', written in Python to illustrate the ideas and approach.

Temporal transformer-spatial graph convolutional network: an intelligent classification model for anti N-methyl-D-aspartate receptor encephalitis based on electroencephalogram signal

Encephalitis is a disease typically caused by viral infections or autoimmunity. The most common type of autoimmune encephalitis is anti-N-methyl-D-aspartate receptor (NMDAR) antibody-mediated, known as anti-NMDA receptor encephalitis, which is a rare disease. Specific EEG patterns, including "extreme delta brush" (EDB), have been reported in patients with anti-NMDA receptor encephalitis. The aim of this study was to develop an intelligent diagnostic model for encephalitis based on EEG signals. A total of 131 Participants were selected based on reasonable inclusion criteria and divided into three groups: health control (35 participants), viral encephalitis (58 participants), and anti NMDAR receptor encephalitis (55 participants). Due to the low prevalence of anti-NMDAR receptor encephalitis, it took several years to collect participants' EEG signals while they were in an awake state. EEG signals were collected and analyzed following the international 10-20 system layout. We proposed a model called Temporal Transformer-Spatial Graph Convolutional Network (TT-SGCN), which consists of a Preprocess Module, a Temporal Transformer Module (TTM), and a Spatial Graph Convolutional Module (SGCM). The raw EEG signal was preprocessed according to traditional procedures, including filtering, averaging, and Independent Component Analysis (ICA) method. The EEG signal was then segmented and transformed using short-time Fourier transform (STFT) to produce concatenated power density (CPD) maps, which served as inputs for the proposed model. TTM extracted the time-frequency features of each channel, and SGCM fused these features using graph convolutional methods based on the location of electrodes. The model was evaluated in two experiments: classification of the three groups and pairwise classification among the three groups. The model was trained using two stages and achieved the performance, with an accuracy of 82.23%, recall of 80.75%, precision of 82.51%, and F1 score of 81.23% in the classification of the three groups. The proposed model has the potential to become an intelligent auxiliary diagnostic tool for encephalitis.

Distance-Weighted Graph Neural Networks on FPGAs for Real-Time Particle Reconstruction in High Energy Physics

Graph neural networks have been shown to achieve excellent performance for several crucial tasks in particle physics, such as charged particle tracking, jet tagging, and clustering. An important domain for the application of these networks is the FGPA-based first layer of real-time data filtering at the CERN Large Hadron Collider, which has strict latency and resource constraints. We discuss how to design distance-weighted graph networks that can be executed with a latency of less than one μs on an FPGA. To do so, we consider a representative task associated to particle reconstruction and identification in a next-generation calorimeter operating at a particle collider. We use a graph network architecture developed for such purposes, and apply additional simplifications to match the computing constraints of Level-1 trigger systems, including weight quantization. Using the hls4ml library, we convert the compressed models into firmware to be implemented on an FPGA. Performance of the synthesized models is presented both in terms of inference accuracy and resource usage.

Retrospective analysis of hemispheric structural network change as a function of location and size of glioma

Gliomas are neoplasms that arise from glial cell origin and represent the largest fraction of primary malignant brain tumours (77%). These highly infiltrative malignant cell clusters modify brain structure and function through expansion, invasion and intratumoral modification. Depending on the growth rate of the tumour, location and degree of expansion, functional reorganization may not lead to overt changes in behaviour despite significant cerebral adaptation. Studies in simulated lesion models and in patients with stroke reveal both local and distal functional disturbances, using measures of anatomical brain networks. Investigations over the last two decades have sought to use diffusion tensor imaging tractography data in the context of intracranial tumours to improve surgical planning, intraoperative functional localization, and post-operative interpretation of functional change. In this study, we used diffusion tensor imaging tractography to assess the impact of tumour location on the white matter structural network. To better understand how various lobe localized gliomas impact the topology underlying efficiency of information transfer between brain regions, we identified the major alterations in brain network connectivity patterns between the ipsilesional versus contralesional hemispheres in patients with gliomas localized to the frontal, parietal or temporal lobe. Results were indicative of altered network efficiency and the role of specific brain regions unique to different lobe localized gliomas. This work draws attention to connections and brain regions which have shared structural susceptibility in frontal, parietal and temporal lobe glioma cases. This study also provides a preliminary anatomical basis for understanding which affected white matter pathways may contribute to preoperative patient symptomology.