BrainStorm: Highly Extensible Software for Advanced Electrophysiology and MEG/EEG Imaging

点击翻译按钮获取中文摘要

PROJECT SUMMARY / ABSTRACT Brainstorm is a Matlab/Java multi-platform (Linux, MacOS, Windows) software environment for analysis and visualization of electrophysiological (e-phys) data. While originally focused on noninvasive EEG and MEG, Brainstorm now also includes tools for analysis and advanced visualization of a broader range of data modalities: intracranial EEG including ECOG, SEEG macroelectrode recordings, MUA and LFP microelectrode data in humans and in-vitro/in-vivo preparations in animal models, and noninvasive NIRS data. The software is widely used and plays an increasingly important role in clinical and cognitive neuroscience research as reflected in the following statistics over the past decade: 1,700 published articles reporting analyses performed with Brainstorm, 30,000 user accounts, 34,000 posts on the online user forum, and 2,000 students, postdocs and faculty around the globe have attended in-person or online Brainstorm training events. Over the past four years we have built on the extensive capabilities of Brainstorm by defining and implementing a BIDS-compatible database structure and adding tools for cloud/advanced computing over distributed data resources. We have added improved interfaces with other EEG/MEG software, leading spike-sorting e-phys software, and hardware (including the ability to import from multiple commercial file data formats). We have also added interoperability with MRI- analysis software and deployed tools for FEM-based head modeling. Under this renewal we will address the following: Aim 1: Enabling Tools for Naturalistic and Single-Trial Neurophysiology Studies: There is increasing scientific interest in eliciting, recording, and analyzing neural responses to non-repetitive and naturalistic stimuli. Data of this type present particular challenges in terms of documenting rigorously varied, uncontrolled sensory events and behavioral responses with machine and human readable annotations, and relating such complex sensory presentations and behavior to measured brain activity. We will therefore develop a suite of tools for preparing and analyzing multimodal e-phys data in the context of naturalistic neuroscience. Aim2: Tools for Invasive Recordings: We will build on developments in the current project period to provide tools for preprocessing and analysis of invasive recordings and FEM-based modeling of intracranial current fields including labelling, localization and co-registration of complex iEEG electrode montages from X-ray CT to MRI, automated identification and tagging of interictal features in iEEG, and characterization, localization and validation of neuronal sources from iEEG. Aim 3: Software Enhancement, Integration, Maintenance, Training, and Documentation: We will pursue a three-pronged development track: (i) a hub-based framework in which third-party tools can be linked, run and with results visualized entirely from Brainstorm as a hub; (ii) build on our existing provenance tools so that users can save analysis histories to allow full reproducibility; and (iii) provide links and interoperability with statistical packages to promote rigor. We will continue to develop Brainstorm with multiplatform compatibility, web-based documentation, user forums, and hands-on training.