Jeon Soohyun, Kang Jae-Eon, Hwang Jundong, Calhoun Vince D, Lee Jong-Hwan
Department of Brain and Cognitive Engineering, Korea University, Seoul, South Korea.
Center for Translational Research in Neuroimaging and Data Science (TReNDS): Georgia State University, Georgia Institute of Technology, and Emory University, Atlanta, USA; Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, USA.
Biol Psychiatry Cogn Neurosci Neuroimaging. 2025 Jun 11. doi: 10.1016/j.bpsc.2025.06.002.
Impulsivity in highly heritable attention deficit hyperactivity disorder (ADHD) has been studied using neural activity via fMRI or genetic data, but rarely with multivariate methods linking both. We investigated coupled neural activity and gene expression signatures, using parallel independent component analysis (pICA) and Adolescent Brain Cognitive Development data.
Children with ADHD (n = 394; 63% males) and healthy controls (n = 1,000; 47% males) of European ancestry were included. The subjects were randomly divided into 80% discovery and 20% replication datasets with demographic stratification. We analyzed neural activity and gene expressions from the discovery datasets using pICA and extracted paired independent components (pICs). The loading coefficients of the pICs were utilized to predict behavioral and cognitive data for stop signal task (SST) in replication datasets.
We identified three pICs estimated from gene expression in the cortex, cerebellum, and nucleus accumbens. Significant neural activity was mainly localized to the orbital/inferior/middle frontal gyri, rectal gyrus, precuneus, inferior temporal gyrus, inferior parietal lobule, and cerebellum. Significant gene components were associated with immunoglobulin, taste receptor, and immunity-related terms and were overlapped with ADHD-related genes. The extracted fMRI-/Gene-ICs were significantly correlated with mean reaction time, stop signal reaction time of SST, and behavioral inhibition with a large boost in sensitivity when both the paired fMRI-/Gene-ICs and their interaction were used in a multimodal regression analysis.
We reported biologically plausible pairs of neural activity and gene sets using pICA, which were significantly associated with ADHD impulsivity-related behavioral and cognitive data.
在高度遗传性注意力缺陷多动障碍（ADHD）中，冲动性已通过功能磁共振成像（fMRI）的神经活动或基因数据进行研究，但很少使用将两者联系起来的多变量方法。我们使用并行独立成分分析（pICA）和青少年大脑认知发展数据，研究了耦合神经活动和基因表达特征。
纳入欧洲血统的ADHD儿童（n = 394；63%为男性）和健康对照（n = 1000；47%为男性）。受试者通过人口统计学分层随机分为80%的发现数据集和20%的复制数据集。我们使用pICA分析发现数据集中的神经活动和基因表达，并提取配对独立成分（pICs）。利用pICs的负荷系数预测复制数据集中停止信号任务（SST）的行为和认知数据。
我们从皮质、小脑和伏隔核的基因表达中确定了三个pICs。显著的神经活动主要定位于眶额/额下回/额中回、直肠回、楔前叶、颞下回、顶下小叶和小脑。显著的基因成分与免疫球蛋白、味觉受体和免疫相关术语相关，并与ADHD相关基因重叠。当在多模态回归分析中同时使用配对的fMRI-/基因-ICs及其相互作用时，提取的fMRI-/基因-ICs与平均反应时间、SST的停止信号反应时间以及行为抑制显著相关，敏感性大幅提高。
我们使用pICA报告了神经活动和基因组的生物学合理配对，它们与ADHD冲动性相关的行为和认知数据显著相关。
