Dauwels Justin, Cash Sydney
Annu Int Conf IEEE Eng Med Biol Soc. 2014;2014:4439-42. doi: 10.1109/EMBC.2014.6944609.
Up to 30% of epileptic patients have seizures poorly controlled with anti-epileptic drugs alone. Surgical therapy might be beneficial to patients who respond poorly to drug treatments. It is therefore crucial to accurately localize the seizure focus. Neurologists rely heavily on seizures to determine the focus. The invasive recordings usually continue for days or weeks, which is costly and entails significant risk for the patients. In this paper, techniques are developed to localize the seizure focus using brief interictal intracranial EEG (iEEG). A supervised learning paradigm is utilized making use of features extracted from interictal iEEG on multiple referential montages. Analysis of 14 epileptic patients (implanted with depth electrodes) shows that iEEG features such as slowing, ripples, spikes, and local synchrony measures are strongly correlated to the seizure focus. These procedures may allow reliable localization of the seizure focus from brief interictal iEEG, which in turn may lead to shorter hospitalizations.
高达30%的癫痫患者仅使用抗癫痫药物治疗时癫痫发作控制不佳。手术治疗可能对药物治疗反应不佳的患者有益。因此，准确确定癫痫发作灶至关重要。神经科医生严重依赖癫痫发作来确定病灶。侵入性记录通常持续数天或数周，这成本高昂且给患者带来重大风险。在本文中，开发了利用简短发作间期颅内脑电图（iEEG）来定位癫痫发作灶的技术。利用一种监督学习范式，该范式利用从多个参考导联组合的发作间期iEEG中提取的特征。对14名植入深度电极的癫痫患者的分析表明，诸如慢波、涟漪、棘波和局部同步性测量等iEEG特征与癫痫发作灶密切相关。这些程序可能允许从简短的发作间期iEEG中可靠地定位癫痫发作灶，这反过来可能导致住院时间缩短。
