BRAIN EAGER: Robust longitudinal characterization of brain oscillations in the first 3 years of life

人类的大脑在生命的前3年经历了快速而深刻的变化，伴随着新的认知技能的出现，包括记忆和识别面孔和物体，获得词汇，以及将注意力集中在手头的任务上等等。虽然如此丰富的功能需要多个大脑区域的协调，但人们对不同大脑区域的脑电活动变化与行为变化之间的关系知之甚少。这种知识差距部分是由于主流脑成像方法功能性磁共振成像（fMRI）所施加的方法限制。fMRI不仅要求婴幼儿在观察过程中保持静止或睡眠状态，而且不能以高时间分辨率（毫秒）直接测量神经元活动的快速变化。随着最近技术和计算的进步，克服这些技术障碍，直接测量行为婴儿的脑电活动已经成为可能。在美国国家科学基金会的支持下，波士顿儿童医院/哈佛医学院认知神经科学实验室的Stamoulis博士和他的同事们将有难得的机会，利用一些先进的计算方法，系统地描述从3至36个月大的婴幼儿中反复获得的纵向hdEEG数据中获得的神经信号的发育相关变化。这项研究将提供关于大脑在早期发育过程中如何变化的基本信息。这个项目的发现也有望帮助家庭了解早期经历是如何塑造大脑和促进认知功能的，并将启发开发新的课程和教学材料，以教育学生、研究人员和临床医生认知发展的行为和神经机制之间的关系。该项目是一项雄心勃勃的尝试，旨在通过分析高密度脑电图（hdEEG）数据来描述发育中的人类大脑的变化，这些数据收集自同一婴儿的头三岁，使用源定位和不同脑功能区域内部和之间的神经振荡频率分析。研究将集中在表征早期发育过程中来自不同空间位置的多个时间点的脑电信号的振荡波形。这些波形在不同频段的功率，如theta、alpha、beta和gamma功率，已知在早期发育的不同时间点出现，并与外部刺激、信息处理需求和行为的变化有关。然而，在这个年龄范围内，主要振荡频率、功率和大脑区域空间分布的年龄相关变化尚未得到系统的表征。本研究将对200名正常发育婴儿在3、6、9、12、18、24和36个月的纵向高密度脑电图数据进行分析，并分别在相同类型的任务和无任务条件下进行。新的信号源分析方法将应用于hdEEGs，提取和定位优势信号源，并将信号源分解为单个振荡分量，并在不同年龄进行比较。静息和功能网络以及识别源之间的定向连接也将被系统地量化并在不同年龄进行比较。该项目有望为利用脑电图研究人类大脑发育提供一种新的基于源的语言，并揭示神经信号如何在时间、频率和大脑空间中变化，从而使婴儿能够与世界交流并获得新技能。

The human brain undergoes rapid and profound changes during the first 3 years of life, which accompany the emergence of new cognitive skills, including remembering and recognizing faces and objects, acquiring vocabularies, and focusing attention on the task at hand, among others. While such a rich repertoire of functions requires the coordination of multiple brain regions, little is known about how changes in the brain's electrical activity across different brain regions correlate with changes in behavior. This knowledge gap is in part due to methodological limitations imposed by the predominant brain imaging method, functional magnetic resonance imaging (fMRI). fMRI not only requires infants or toddlers to remain still or asleep during the observation, but also cannot directly measure rapid changes in neuronal activity at a high temporal resolutions (millisecond). With recent technological and computational advances, it has become possible to overcome these technical barriers and obtain direct measurements of brain electrical activity in behaving infants. With the support of the National Science foundation, Dr Stamoulis and colleagues at the Laboratory of Cognitive Neuroscience at Boston Children's Hospital/Harvard Medical School, will have the rare opportunity to systematically characterize development-related changes in neural signals derived from longitudinal hdEEG data acquired in infants and toddlers repeatedly across 3 to 36 months of age using a number of advanced computational approaches. This study will provide fundamental information regarding how the brain changes across early development. Findings from this project are also expected to help educate families on how early experiences shape the brain and facilitate cognitive functions, and will inspire the development of new courses and instructional materials to educate students, researchers and clinicians on the relationships between behavioral and neural mechanisms of cognitive development.The project is an ambitious attempt at characterizing changes in the developing human brain by analysing high-density electroencephalography (hdEEG) data collected from the same infants across the first three years of life using source localization and frequency analysis of neural oscillations within and between different functional brain regions. The investigation will focus on characterizing oscillatory waveforms of brain electrical signals originating from different spatial locations across multiple time points during early development. The power of these waveforms in different frequency bands, e.g. theta, alpha, beta, and gamma power, are known to emerge at different time points during early development and to be associated with variations in external stimuli, information processing demands, and behaviors. However, age-related changes in the dominant oscillation frequency, power and spatial distribution among brain regions have not been systematically characterized during this age range. Longitudinal high-density EEG data from about 200 typically developing infants at 3, 6, 9, 12, 18, 24, and 36 months of age will be analyzed under the same type of tasks and no-task conditions. Novel source analysis methods will be applied to hdEEGs, to extract and localize dominant sources and to decompose source signals into individual oscillation components and compare them across ages. Resting and functional networks and directional connectivities between identified sources will also be systematically quantified and compared across ages. This project is expected to provide a new source-based language for investigating human brain development using EEG and to reveal how neural signals change in time, frequency and brain spaces to enable infants to communicate with the world and to acquire new skills.