ObjectiveThe time relationship between seizure semiology and epileptic discharges during focal epileptic seizures is a crucial predictor for the localization of epileptogenic zone. Low voltage fast activities (LVFA), especially gamma band oscillations, are confirmed to play a central role in ictogenesis and semiology production. In the present study, we focus on the “electro-clinical correlation” between LVFA in agranulo-dysgranular insulo-cingulate cortices and the sign of “Chapeau de gendarme (CDG)” via detailed analysis of ictal video-stereoencephalography (video-SEEG) of focal epileptic seizures. MethodsWe retrospectively analyzed the ictal video-SEEG of the 7 cases in which CDG signs were presented in habitual seizures and intracerebral electrodes were co-implanted in agranulo-dysgranular insular and cingulate cortices. We calculate the latency of LVFA in each of cortical regions of interest, agranulo-dygranular insular cortex, agranulo-dysgranular cingulate cortex, and the latency of CDG signs via visual and spectral analysis of the ictal SEEG. Moreover, Pearson correlation analysis and linear regression were used to test the time relationship between gamma band oscillations in agranulo-dysgranular insulo-cingulate cortices and generation of CDG signs. ResultsThe co-activation of LVFA occurred in agranulo-dysgranular insulo-cingulate cortices always preceded the appearance of CDG sign in all of the 69 seizures. The LVFA were confirmed as gamma band oscillations via spectral analysis of SEEG. A linear relationship between the latencies of CDG signs and the latencies of co-activation of agranulo-dysgranular insulo-cingulate cortices in gamma band was furth confirmed by Pearson correlation analysis and linear regression. ConclusionsThere is a causal relationship between the involvement of agranulo-dysgranular insulo-cingulate cortices and the generation of CDG sign, and thus the CDG sign could be view as semiological marker of activation of emotional insulo-cingulate cortex in focal epilepsy.
ObjectiveTo summarize and explore the application of machine learning models to survival data with non-proportional hazards (NPH), and to provide a methodological reference for large-scale, high-dimensional survival data. MethodsFirst, the concept of NPH and related testing methods were outlined. Then the advantages and disadvantages of machine learning algorithm-based NPH survival analysis methods were summarized based on the relevant literature. Finally, using real-world clinical data, a case study was conducted with two ensemble machine learning models and two deep learning models in survival data with NPH: a study of the risk of death within 30 days in stroke patients in the ICU. ResultsEight commonly used machine learning model-based NPH survival analyses were identified, including five traditional machine learning models such as random survival forest and three deep learning models based on artificial neural networks (e.g., DeepHit). The case study found that the random survival forest model performed the best (C-index=0.773, IBS=0.151), and the permutation importance-based algorithm found that age was the most important characteristic affecting the risk of death in stroke patients. ConclusionSurvival big data in the era of precision medicine presenting NPH are common, and machine learning model-based survival analysis can be used when faced with more complex survival data and higher survival analysis needs.