Secondary analysis of resting state MEG data using the Human Neocortical Neurosolver software tool for cellular and circuit-level interpretation

使用 Human Neocortical Neurosolver 软件工具对静息态 MEG 数据进行二次分析，以进行细胞和电路级解释

• 批准号: 10505661
• 负责人: STEPHANIE Ruggiano JONES
• 金额: $ 117.36万
• 依托单位: BROWN UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-08-01 至 2025-07-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10505661
• 关键词: Accounting; Address; Adult; Age; Aging; Area; BRAIN initiative; Behavioral; Beta Rhythm; Biological Markers; Biophysics; Brain; Brain imaging; Cells; Characteristics; Communities; Data; Data Set; Development; Diagnostic; Diagnostic Procedure; Electroencephalography; Event; Follow-Up Studies; Foundations; Gender; Goals; Human; Individual; Information Theory; Link; Longevity; Magnetoencephalography; Maps; Measures; Methods; Motor; Mus; Neurosciences; Output; Probability; Reproducibility; Research; Rest; Sensorimotor functions; Shapes; Signal Transduction; Software Design; Software Tools; Testing; Time; Translations; Validation; Variant; age group; brain abnormalities; cell type; connectome; data access; deep neural network; experience; gamma-Aminobutyric Acid; improved; information processing; large scale data; meter; models and simulation; neocortical; neural model; neurodevelopment; neuromechanism; novel; novel diagnostics; open data; open source; relating to nervous system; repository; response; secondary analysis; simulation; source localization; targeted treatment; temporal measurement; theories; tool; treatment strategy

Project Summary The neuroscience community is experiencing a revolution in its ability to share and analyze vast amounts of human brain imaging data, with support from the BRAIN Initiative and other substantial data-sharing efforts. One domain in which there has been significant open access progress is Magnetoencephalography (MEG), where data is available from hundreds of subjects during resting states and various behavioral conditions. While MEG (and EEG) provide biomarkers of healthy and abnormal brain dynamics with fine temporal resolution, these macroscopic scale signals have lacked interpretability at the underlying cellular and circuit level. This difficulty limits translation of M/EEG into mechanistic theories of information processing, or into new diagnostic methods and treatments that target e.g., specific cell types. To address this need, with support from the BRAIN initiative, we developed an open-source neural modeling software designed for circuit level interpretation of M/EEG data, the Human Neocortical Neurosolver (HNN), which is now freely available (https://hnn.brown.edu). The utility of this new tool can be best demonstrated by application to large-scale data, where theories on the neural mechanisms underlying reproducible MEG signals, such as resting state oscillations, and changes in these signals across subjects can be developed. We propose to re-analyze open-access MEG data with a focus on identifying stereotypical time-domain waveforms during resting state oscillations and variability across subjects (Aim 1), and to apply the HNN software tool to develop biophysically-constrained hypothesis on the underlying cellular and circuit generators of these waveforms and their variability (Aim 2). The application here focusses on quantifying and interpreting changes in sensorimotor resting state oscillations across developmental trajectories in adults (18-88yrs). This example case will provide the foundation for the ultimate goal of this project, which is to develop a framework in which the wealth of open-source M/EEG data can be harnessed to define stereotypical waveform shapes in MEG/EEG signals and quantifiable shape differences across groups. These waveforms can then be imported into HNN for biophysically constrained predictions on circuit mechanisms that generate individual subject data and group differences. This framework has the potential to transform M/EEG from being purely diagnostic to providing targeted treatment strategies to improve brain function.

项目概要 神经科学界正在经历一场关于共享和分析大量信息的能力的革命。 人脑成像数据，在 BRAIN Initiative 和其他大量数据共享工作的支持下。一 开放获取取得重大进展的领域是脑磁图（MEG），其中 可以从数百名受试者在静息状态和各种行为条件下获得数据。而MEG （和脑电图）提供健康和异常大脑动态的生物标志物，具有良好的时间分辨率，这些 宏观尺度信号在底层细胞和电路层面缺乏可解释性。这个难度 限制了 M/EEG 转化为信息处理的机械理论或新的诊断方法 以及针对特定细胞类型的治疗。为了满足这一需求，在 BRAIN 计划的支持下， 我们开发了一款开源神经建模软件，专为 M/EEG 数据的电路级解释而设计， 人类新皮质神经解算器 (HNN)，现已免费提供 (https://hnn.brown.edu)。的效用 这种新工具可以通过应用于大规模数据得到最好的证明，其中神经网络理论 可重复 MEG 信号的机制，例如静息态振荡，以及这些信号的变化 可以开发跨学科的信号。我们建议重新分析开放获取的 MEG 数据，重点关注 识别静息态振荡期间的刻板时域波形和受试者之间的变异性 （目标 1），并应用 HNN 软件工具来开发基于生物物理约束的假设 这些波形及其变异性的细胞和电路发生器（目标 2）。这里的应用重点是 量化和解释发育轨迹中感觉运动静息态振荡的变化 成人（18-88 岁）。该示例案例将为该项目的最终目标提供基础，即 开发一个框架，可以利用丰富的开源 M/EEG 数据来定义刻板印象 MEG/EEG 信号中的波形形状以及各组之间可量化的形状差异。这些波形可以 然后导入 HNN，对产生的电路机制进行生物物理约束的预测 个体受试者数据和群体差异。该框架有可能将 M/EEG 从 纯粹的诊断以提供有针对性的治疗策略来改善大脑功能。