Schraegle William A, Tillman Rachael, Ailion Alyssa, Babajani-Feremi Abbas, Titus Jeffrey B, DeLeon Rosario C, Clarke Dave, Hermann Bruce P
Department of Neurology, Dell Medical School, The University of Texas at Austin, Austin, Texas, USA.
Comprehensive Pediatric Epilepsy Center, Dell Children's Medical Center, Austin, Texas, USA.
Epilepsia. 2022 May;63(5):1177-1188. doi: 10.1111/epi.17193. Epub 2022 Feb 17.
A broad spectrum of emotional-behavioral problems have been reported in pediatric temporal lobe epilepsy (TLE), but with considerable variability in their presence and nature of expression, which hampers precise identification and treatment. The present study aimed to empirically identify latent patterns or behavioral phenotypes and their correlates.
Data included parental ratings of emotional-behavioral status on the Behavior Assessment System for Children, 2nd Edition (BASC-2) of 81 children (mean age = 11.79, standard deviation [SD] = 3.93) with TLE. The nine clinical subscales were subjected to unsupervised machine learning to identify behavioral subgroups. To explore concurrent validity and the underlying composition of the identified clusters, we examined demographic factors, seizure characteristics, psychosocial factors, neuropsychological performance, psychiatric status, and health-related quality of life (HRQoL).
Three behavioral phenotypes were identified, which included no behavioral concerns (Cluster 1, 43% of sample), externalizing problems (Cluster 2, 41% of sample), and internalizing problems (Cluster 3, 16% of sample). Behavioral phenotypes were characterized by important differences across clinical seizure variables, psychosocial/familial factors, everyday executive functioning, and HRQoL. Cluster 2 was associated with younger child age, lower maternal education, and higher rate of single-parent households. Cluster 3 was associated with older age at epilepsy onset and higher rates of hippocampal sclerosis and parental psychiatric history. Both Cluster 2 and 3 demonstrated elevated family stress. Concurrent validity was demonstrated through the association of psychiatric (i.e., rate of Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) disorders and psychotropic medication) and parent-rated HRQoL variables.
Youth with TLE present with three distinct behavioral phenotypes that correspond with important clinical and sociodemographic markers. The current findings demonstrate the variability of behavioral presentations in youth with TLE and provide a preliminary framework for screening and targeting intervention to enhance support for youth with TLE and their families.
已有报道称小儿颞叶癫痫（TLE）存在广泛的情绪行为问题，但这些问题的出现及其表现性质存在相当大的差异，这妨碍了精确的识别和治疗。本研究旨在通过实证确定潜在模式或行为表型及其相关因素。
数据包括81名TLE患儿（平均年龄 = 11.79，标准差[SD] = 3.93）的父母对其情绪行为状况在《儿童行为评估系统第二版》（BASC - 2）上的评分。对九个临床分量表进行无监督机器学习以识别行为亚组。为了探究所识别集群的同时效度及其潜在构成，我们考察了人口统计学因素、癫痫发作特征、心理社会因素、神经心理表现、精神状态以及健康相关生活质量（HRQoL）。
识别出三种行为表型，包括无行为问题（集群1，占样本的43%）、外化问题（集群2，占样本的41%）和内化问题（集群3，占样本的16%）。行为表型的特征是在临床癫痫变量、心理社会/家庭因素、日常执行功能和HRQoL方面存在重要差异。集群2与儿童年龄较小、母亲教育程度较低以及单亲家庭比例较高相关。集群3与癫痫发作起始年龄较大、海马硬化发生率较高以及父母精神病史相关。集群2和集群3均表现出家庭压力升高。通过精神疾病（即《精神疾病诊断与统计手册第五版》（DSM - 5）疾病发生率和精神药物）与父母评定的HRQoL变量之间的关联证明了同时效度。
患有TLE的青少年呈现出三种不同的行为表型，这些表型与重要的临床和社会人口统计学标志物相对应。当前研究结果表明了TLE青少年行为表现的变异性，并为筛查和针对性干预提供了一个初步框架，以加强对TLE青少年及其家庭的支持。
