Data_Sheet_1_Spectral-based thickness profiling of the corpus callosum enhances anomaly detection in fetal alcohol spectrum disorders.docx

IntroductionFetal alcohol spectrum disorders (FASD) range from fetal alcohol syndrome (FAS) to non-syndromic forms (NS-FASD). The neuroanatomical consequences of prenatal alcohol exposure are mainly the reduction in brain size, but also focal abnormalities such as those of the corpus callosum (CC). We previously showed a narrowing of the CC for brain size, using manual measurement and its usefulness to improve diagnostic certainty. Our aim was to automate these measurements of the CC and identify more recurrent abnormalities in FAS subjects, independently of brain size reduction.MethodsWe developed a fast, automated, and normalization-free method based on spectral analysis to generate thicknesses of the CC continuously and at singular points (genu, body, isthmus, and splenium), and its length (LCC). We applied it on midsagittal section of the CC extracted from T1-anatomical brain MRI of 89 subjects with FASD (52 FAS, 37 NS-FASD) and 126 with typically development (6–20 y-o). After adjusting for batch effect, we compared the mean profiles and thicknesses of the singular points across the 3 groups. For each parameter, we established variations with age (growth charts) and brain size in the control group (scaling charts), then identified participants with abnormal measurements (

胎儿酒精谱系障碍（FASD）包括胎儿酒精综合征（FAS）及非综合征型（NS-FASD）形式。孕期酒精暴露的神经解剖学后果主要体现在脑体积的减少，亦伴随如胼胝体（CC）等局部异常。我们先前通过手动测量胼胝体宽度与脑体积之间的关系，并证明了其提高诊断确定性的实用性。我们的目标是自动化胼胝体的测量，并在独立于脑体积减少的情况下，识别FAS受试者中更为常见的异常。方法：我们开发了一种基于光谱分析的高速、自动化且无需归一化的方法，以连续生成胼胝体（ genu, body, isthmus, 和 splenium）各点的厚度及其长度（LCC）。我们将此方法应用于从89名FASD受试者（52名FAS，37名NS-FASD）及126名典型发展个体（6–20岁）的T1解剖脑MRI中提取的胼胝体横切面。在调整批次效应后，我们比较了三组间的平均轮廓和各点的厚度。对于每个参数，我们在对照组中建立了与年龄（生长曲线）和脑体积相关的变化（缩放曲线），并确定了测量异常的参与者。