Department of Psychiatry, University of British Columbia, 2255 Wesbrook Mall, Vancouver V6T 2A1, BC, Canada; BC Children's Hospital Research Institute, Vancouver, BC, Canada.
BC Children's Hospital Research Institute, Vancouver, BC, Canada.
Neuroimage. 2021 Feb 1;226:117537. doi: 10.1016/j.neuroimage.2020.117537. Epub 2020 Nov 11.
Patterns of functional connectivity are unique at the individual level, enabling test-retest matching algorithms to identify a subject from among a group using only their functional connectome. Recent findings show that accuracies of these algorithms in children increase with age. Relatedly, the persistence of functional connectivity (FC) patterns across tasks and rest also increases with age. This study investigated the hypothesis that within-subject stability and between-subject similarity of the whole-brain pediatric connectome are developmentally relevant outcomes. Using data from 210 help-seeking children and adolescents, ages 6-21 years (Healthy Brain Network Biobank), we computed whole-brain FC matrices for each participant during two different movies (MovieDM and MovieTP) and two runs of task-free rest (all from a single scan session) and fed these matrices to a test-retest matching algorithm. We replicated the finding that matching accuracies for children and youth (ages 6-21 years) are low (18-44%), and that cross-state and cross-movie accuracies were the lowest. Results also showed that parcellation resolution and the number of volumes used in each matrix affect fingerprinting accuracies. Next, we calculated three measures of whole-connectome stability for each subject: cross-rest (Rest1-Rest2), cross-state (MovieDM-Rest1), and cross-movie (MovieDM-MovieTP), and three measures of within-state between-subject connectome similarity for Rest1, MovieDM, and MovieTP. We show that stability and similarity were correlated, but that these measures were not related to age. A principal component analysis of these measures yielded two components that we used to test for brain-behavior correlations with IQ, general psychopathology, and social skills measures (n = 119). The first component was significantly correlated with the social skills measure (r=-0.26, p = 0.005). Post hoc correlations showed that the social skills measure correlated with both cross-rest stability (r=-0.29, p = 0.001) and with connectome similarity during MovieDM (r=-0.28, p = 0.002). These findings suggest that the stability and similarity of the whole-brain connectome relate to the development of social skills. We infer that the maturation of the functional connectome simultaneously achieves patterns of FC that are distinct at the individual subject level, that are shared across individuals, and that are persistent across states and across runs-features which presumably combine to optimize neural processing during development. Future longitudinal work could reveal the developmental trajectories of stability and similarity of the connectome.
功能连接模式在个体水平上是独特的，这使得测试-再测试匹配算法能够仅使用功能连接体来识别组中的个体。最近的研究结果表明，这些算法在儿童中的准确性随着年龄的增长而提高。相关地，功能连接（FC）模式在任务和休息之间的持久性也随着年龄的增长而增加。本研究假设，整个大脑连接体的个体内稳定性和个体间相似性是与发展相关的结果。使用来自 210 名寻求帮助的儿童和青少年的数据（健康大脑网络生物库），我们为每个参与者在两部不同的电影（电影 DM 和电影 TP）和两次无任务休息（均来自单次扫描会话）期间计算了全脑 FC 矩阵，并将这些矩阵输入到测试-再测试匹配算法中。我们复制了这样一个发现，即儿童和青少年（6-21 岁）的匹配准确性较低（18-44%），并且跨状态和跨电影的准确性最低。结果还表明，分区分辨率和每个矩阵中使用的卷数会影响指纹识别的准确性。接下来，我们为每个受试者计算了三个整体连接体稳定性度量：跨休息（Rest1-Rest2）、跨状态（MovieDM-Rest1）和跨电影（MovieDM-MovieTP），以及三个在 Rest1、MovieDM 和 MovieTP 状态下的个体间连接体相似性度量。我们表明，稳定性和相似性是相关的，但这些度量与年龄无关。对这些度量的主成分分析得出了两个分量，我们使用这些分量来测试与智商、一般精神病理学和社交技能测量（n=119）的脑-行为相关性。第一个分量与社交技能测量显著相关（r=-0.26，p=0.005）。事后相关性表明，社交技能测量与跨休息稳定性（r=-0.29，p=0.001）和 MovieDM 期间的连接体相似性（r=-0.28，p=0.002）相关。这些发现表明，整个大脑连接体的稳定性和相似性与社交技能的发展有关。我们推断，功能连接体的成熟同时实现了个体水平上独特的 FC 模式、个体间共享的模式和状态间及运行间持久的模式，这些特征可能结合起来在发育过程中优化神经处理。未来的纵向研究可以揭示连接体稳定性和相似性的发展轨迹。
