Relation of Brain Perfusion Patterns to Sudden Unexpected Death Risk Stratification: A Study in Drug Resistant Focal Epilepsy.

To explore the role of the interictal and ictal SPECT to identity functional neuroimaging biomarkers for SUDEP risk stratification in patients with drug-resistant focal epilepsy (DRFE). Twenty-nine interictal-ictal Single photon emission computed tomography (SPECT) scans were obtained from nine DRFE patients. A methodology for the relative quantification of cerebral blood flow of 74 cortical and sub-cortical structures was employed. The optimal number of clusters (K) was estimated using a modified v-fold cross-validation for the use of K means algorithm. The two regions of interest (ROIs) that represent the hypoperfused and hyperperfused areas were identified. To select the structures related to the SUDEP-7 inventory score, a data mining method that computes an automatic feature selection was used. During the interictal and ictal state, the hyperperfused ROIs in the largest part of patients were the bilateral rectus gyrus, putamen as well as globus pallidus ipsilateral to the seizure onset zone. The hypoperfused ROIs included the red nucleus, substantia nigra, medulla, and entorhinal area. The findings indicated that the nearly invariability in the perfusion pattern during the interictal to ictal transition observed in the ipsi-lateral putamen F = 12.60, p = 0.03, entorhinal area F = 25.80, p = 0.01, and temporal middle gyrus F = 12.60, p = 0.03 is a potential biomarker of SUDEP risk. The results presented in this paper allowed identifying hypo- and hyperperfused brain regions during the ictal and interictal state potentially related to SUDEP risk stratification.

EEG-Derived Functional Connectivity Patterns Associated with Mild Cognitive Impairment in Parkinson's Disease.

OBJECTIVE: To evaluate EEG-derived functional connectivity (FC) patterns associated with mild cognitive impairment (MCI) in Parkinson's disease (PD). METHODS: A sample of 15 patients without cognitive impairment (PD-WCI), 15 with MCI (PD-MCI), and 26 healthy subjects were studied. The EEG was performed in the waking functional state with eyes closed, for the functional analysis it was used the synchronization likelihood (SL) and graph theory (GT). RESULTS: PD-MCI patients showed decreased FC in frequencies alpha, in posterior regions, and delta with a generalized distribution. Patients, compared to the healthy people, presented a decrease in segregation (lower clustering coefficient in alpha p = 0.003 in PD-MCI patients) and increased integration (shorter mean path length in delta (p = 0.004) and theta (p = 0.002) in PD-MCI patients). There were no significant differences in the network topology between the parkinsonian groups. In PD-MCI patients, executive dysfunction correlated positively with global connectivity in beta (r = 0.47) and negatively with the mean path length at beta (r = -0.45); alterations in working memory were negatively correlated with the mean path length at beta r = -0.45. CONCLUSIONS: PD patients present alterations in the FC in all frequencies, those with MCI show less connectivity in the alpha and delta frequencies. The neural networks of the patients show a random topology, with a similar organization between patients with and without MCI. In PD-MCI patients, alterations in executive function and working memory are related to beta integration.

Nonconvulsive epileptic seizure monitoring with incremental learning.

Nonconvulsive epileptic seizures (NCSz) and nonconvulsive status epilepticus (NCSE) are two neurological entities associated with increment in morbidity and mortality in critically ill patients. In a previous work, we introduced a method which accurately detected NCSz in EEG data (referred here as 'Batch method'). However, this approach was less effective when the EEG features identified at the beginning of the recording changed over time. Such pattern drift is an issue that causes failures of automated seizure detection methods. This paper presents a support vector machine (SVM)-based incremental learning method for NCSz detection that for the first time addresses the seizure evolution in EEG records from patients with epileptic disorders and from ICU having NCSz. To implement the incremental learning SVM, three methodologies are tested. These approaches differ in the way they reduce the set of potentially available support vectors that are used to build the decision function of the classifier. To evaluate the suitability of the three incremental learning approaches proposed here for NCSz detection, first, a comparative study between the three methods is performed. Secondly, the incremental learning approach with the best performance is compared with the Batch method and three other batch methods from the literature. From this comparison, the incremental learning method based on maximum relevance minimum redundancy (MRMR_IL) obtained the best results. MRMR_IL method proved to be an effective tool for NCSz detection in a real-time setting, achieving sensitivity and accuracy values above 99%.