期刊
EPILEPSY RESEARCH
卷 133, 期 -, 页码 28-32出版社
ELSEVIER
DOI: 10.1016/j.eplepsyres.2017.03.007
关键词
Machine learning; Neuroimaging; Seizures
资金
- FACES Foundation (Finding A Cure for Epilepsy and Seizures)
- NYU Langone Medical Center
- Epilepsy Study Consortium (ESCI)
- UCB Pharma
- Finding A Cure for Epilepsy and Seizures
- Pfizer
- Lundbeck
- Andrews Foundation
- Friends of Faces
Objective: We used whole brain Tl-weighted MRI to estimate the age of individuals with medically refractory focal epilepsy, and compared with individuals with newly diagnosed focal epilepsy and healthy controls. The difference between neuroanatomical age and chronological age was compared between the three groups. Methods: Neuroanatomical age was estimated using a machine learning-based method that was trained using structural MRI scans from a large independent healthy control sample (N= 2001). The prediction model was then used to estimate age from MRI scans obtained from newly diagnosed focal epilepsy patients (N = 42), medically refractory focal epilepsy patients (N= 94) and healthy controls (N = 74). Results: Individuals with medically refractory epilepsy had a difference between predicted brain age and chronological age that was on average 4.5 years older than healthy controls (p = 4.6 x 10(-5)). No significant differences were observed in newly diagnosed focal epilepsy. Earlier age of onset was associated with an increased brain age difference in the medically refractory group (p = 0.034). Significance: Medically refractory focal epilepsy is associated with structural brain changes that resemble premature brain aging. (C) 2017 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据