Dear Colleagues, We are organizing a virtual workshop titled "Neuroscience Gateway enabling large scale modeling, AI/ML, data processing, and software dissemination on supercomputers". Please see the list of speakers, title of talks, abstracts, date and time below. If you are interested in attending please register via the registration link. Thank you. Amit Majumdar and Subhashini Sivagnanam San Diego Supercomputer Center University of California San Diego ----------------------------------------------------------------------------------------------------------------- Date: 13 November (Thursday) 2025 Time: 9:00 – 12:00 USA Pacific Time Workshop Abstract: The Neuroscience Gateway (NSG) serves the neuroscience community by providing researchers and students easy, open and free access to a large number of neuroscience software and tools on supercomputing resources, academic cloud computing resources, and associated storage resources, which are located at various national academic supercomputer centers in the US and funded by the National Science Foundation. NSG is also a dissemination platform and enables dissemination of neuroscience modeling and data processing software to the community. This virtual workshop is intended for computational, cognitive and experimental neuroscientists – researchers and students - who use large scale modeling, AI/ML and data processing for their research. Computational neuroscientists who are building and using computational models of neurons and network of neurons, and cognitive neuroscientists who are processing EEG, fMRI and other types of data, as well as neuroscientist who are using AI/ML for their research, will find this workshop of interest. NSG is funded by the National Science Foundation and the National Institutes of Health. Registration deadline: November 11, 2025 Register for this session at: https://na.eventscloud.com/ereg/newreg.php?eventid=862126<https://na.eventscloud.com/ereg/newreg.php?eventid=862126&>&<https://na.eventscloud.com/ereg/newreg.php?eventid=862126&> Agenda (times in USA Pacific Time): 9:00 – 9:30: Human Neocortical Neurosolver (HNN): An open-source software for cellular and circuit-level interpretation of human MEG/EEG, Dylan Daniels, Brown University Abstract: HNN is a user-friendly neural modeling software designed to provide a cell and microcircuit-level interpretation of macroscale magneto- and electroencephalography (M/EEG) signals (hnn.brown.edu, Neymotin et al 2020). The foundation of HNN is a biophysically-detailed neocortical model, representing a patch of neocortex receiving thalamic and corticocortical drive. The HNN model was designed to simulate the time course of primary current dipoles and enables direct comparison, in nAm units, to source-localized M/EEG data, along with layer-specific cellular activity. HNN workflows are constructed around simulating commonly measured ERPs and low-frequency oscillations. In this workshop, we will review the scientific foundations underlying the HNN model. Thereafter, will we walk through how to get started with HNN using NSG, and we will briefly demo various features of the HNN GUI. 9:35 – 10:05: AI-driven Brain Digital Twins: Large-scale biophysical models of neuronal circuits to study brain function and disease, Salvador Dura Bernal, SUNY Downstate Health Sciences University Abstract: Understanding brain function and disease requires studying interactions across multiple scales, from molecular and cellular mechanisms to circuit-wide dynamics and behavior. Biophysically detailed brain circuit models provide a powerful tool to integrate diverse experimental data across these scales, allowing researchers to simulate and analyze brain activity in a mechanistic and predictive manner. Leveraging AI and high-performance computing, we have developed large-scale, biologically realistic models of various thalamocortical circuits, including motor, somatosensory, and auditory regions, each comprising ~15,000 neurons and ~30 million synapses. These models incorporate experimentally derived neuronal morphologies, electrophysiological properties, and connectivity, enabling them to reproduce cell-type and layer-specific electrical activity patterns observed in vivo across scales, from membrane voltages and action potentials to local field potentials and electroencephalogram signals. Our models have been instrumental in elucidating the cellular and circuit mechanisms underlying neurological disease and psychiatric disorders such as schizophrenia, epilepsy, dystonia, Alzheimer's and Parkinson’s disease. They also offer insights into therapeutic targets to restore healthy neural dynamics through novel pharmacological and neurostimulation treatments, such as transcranial magnetic stimulation (TMS). These mechanistic models provide the foundation for constructing AI-driven Brain Digital Twins—virtual replicas of an individual’s brain state that integrate multimodal patient data to simulate brain function, predict disease progression, and optimize personalized treatment strategies. By bridging mechanistic modeling of brain circuits and AI, Brain Digital Twins have the potential to transform diagnosis, prevention and treatment of neurological and psychiatric conditions, addressing key challenges in brain research and clinical care. 10:10 – 10:40: NeuroelectroMagnetic Data Archive and high-performance computing for human EEG/MEG/iEEG data: the NEMAR-NSG platform, Arnaud Delorme; Seyed Yahya Shirazi; Dung Truong; Choonhan Youn; Subhashini Sivagnanam (UC San Diego); Russell A. Poldrack (Stanford University); Amitava Majumdar; Scott Makeig (UC San Diego) Abstract: The NeuroElectroMagnetic Data Archive and Tools Resource integrates open data, standardized annotation, and high-performance computing to accelerate reproducible analysis of human EEG, MEG, and iEEG. Built as a gateway to OpenNeuro, NEMAR indexes BIDS-formatted datasets with HED event annotations, provides search, quality assessment, and interactive visualizations, and brokers compute to SDSC resources through the Neuroscience Gateway for scalable, containerized workflows including EEGLAB-based pipelines. By coupling archive storage to NSG-mediated execution on systems such as Expanse, NEMAR minimizes data movement and enables end-to-end, provenance-preserving analyses at scale, supporting FAIR principles and cross-study meta-analysis. Together, NEMAR and NSG lower barriers for large-cohort neurophysiology, promote standardized event semantics, and provide a transparent path from dataset discovery to replicable results. 10:45 – 11:15: Neuron and NSG – parallel software meets parallel hardware, Ted Carnevale, Yale University Abstract: This presentation surveys how the Neuroscience Gateway has enabled productive use of the NEURON simulator to address problems in the domain of computational neuroscience that require high-performance and high-throughput computing. 11:20 – 11:50: HiAER-Spike: Running large spiking neural networks on an FPGA neuromorphic computing platform through NSG, Kenneth Yoshimoto, Diana Vins UC San Diego Abstract: HiAER-Spike is a hardware (FPGA)/software (Python) platform for reconfigurable neuromorphic computing. It has the potential to execute spiking neural networks up to 160 million neurons and 40 billion synapses. A Python interface allows users to specify neurons, connectivity, and input and direct timesteps. This tool is available through NSG. This talk will present integration of NSG with the hardware and show data demonstrating running a network on the FPGA. -----------------------------------------------------------------------------------------------