Deep learning for reliable detection of epileptogenic lesions

Deep learning techniques are ideally suited to find patterns in high-dimensional and complex imaging data, which means that it has the potential to change medicine and healthcare. In this chapter, we will highlight how deep learning can improve the lives of people living with treatment-resistant focal epilepsy by identifying brain surgery targets in this burdensome neurological disease. We will discuss topics that should be investigated to ensure a successful clinical translation of deep learning, by reliably detecting epileptogenic lesions-that is, targets for brain surgery-in high-dimensional brain imaging data such as magnetic resonance imaging or positron emission tomography. As deep learning techniques perform well with imaging data, we envisage that deep learning can support the identification of epileptogenic lesions and may therefore aid the presurgical process in treatment-resistant focal epilepsy.