Scholarlink

Cortical dysplasia is a developmental disorder that affects the brain’s outermost layer, frequently leading to epilepsy. Among its various subtypes, focal cortical dysplasia (FCD) is marked by abnormalities in cortical structure or cellular composition, often resulting in drug-resistant seizures. FCD is one of the most prevalent causes of epilepsy, arising when the brain’s top layer does not develop properly. These malformations may occur due to irregularities in cortical architecture or cytological defects. Magnetic resonance imaging (MRI) plays a critical role in diagnosing FCD, as it provides high-contrast, detailed images of soft tissues. In this study, we propose an automated technique for the detection of FCD. Our method first involves preprocessing MRI images by aligning them to a standard brain atlas. Following this, the classification of lesion vertices is carried out using advanced deep learning techniques. Specifically, we employ a 3D convolutional neural network (CNN) architecture enhanced with a 3D channel attention mechanism to effectively classify MRI images and detect FCD.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

MRI-Based Detection of Focal Cortical Dysplasia Using 3D Convolutional Neural Networks and Attention Mechanisms

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • M. C. Binish,
  • M. A. Abini,
  • Vinu Thomas

摘要

Cortical dysplasia is a developmental disorder that affects the brain’s outermost layer, frequently leading to epilepsy. Among its various subtypes, focal cortical dysplasia (FCD) is marked by abnormalities in cortical structure or cellular composition, often resulting in drug-resistant seizures. FCD is one of the most prevalent causes of epilepsy, arising when the brain’s top layer does not develop properly. These malformations may occur due to irregularities in cortical architecture or cytological defects. Magnetic resonance imaging (MRI) plays a critical role in diagnosing FCD, as it provides high-contrast, detailed images of soft tissues. In this study, we propose an automated technique for the detection of FCD. Our method first involves preprocessing MRI images by aligning them to a standard brain atlas. Following this, the classification of lesion vertices is carried out using advanced deep learning techniques. Specifically, we employ a 3D convolutional neural network (CNN) architecture enhanced with a 3D channel attention mechanism to effectively classify MRI images and detect FCD.