International Journal of Neural Systems

Identifying a core set of features is one of the most important steps in the development of an automated seizure detector. In most of the published studies describing features and seizure classifiers, the features were hand-engineered, which may not be optimal. The main goal of the present paper is using deep convolutional neural networks (CNNs) and random forest to automatically optimize feature selection and classification. The input of the proposed classifier is raw multi-channel EEG and the output is the class label: seizure/nonseizure. By training this network, the required features are optimized, while fitting a nonlinear classifier on the features. After training the network with EEG recordings of 26 neonates, five end layers performing the classification were replaced with a random forest classifier in order to improve the performance. This resulted in a false alarm rate of 0.9 per hour and seizure detection rate of 77% using a test set of EEG recordings of 22 neonates that also included dubious seizures. The newly proposed CNN classifier outperformed three data-driven feature-based approaches and performed similar to a previously developed heuristic method.

The aim of this study was to develop methods for detecting the nonstationary periodic characteristics of neonatal electroencephalographic (EEG) seizures by adapting estimates of the correlation both in the time (spike correlation; SC) and time–frequency domain (time–frequency correlation; TFC). These measures were incorporated into a seizure detection algorithm (SDA) based on a support vector machine to detect periods of seizure and nonseizure. The performance of these nonstationary correlation measures was evaluated using EEG recordings from 79 term neonates annotated by three human experts. The proposed measures were highly discriminative for seizure detection (median ${\mathrm{\text{AUC}}}_{\mathrm{\text{SC}}}$: 0.933 IQR: 0.821–0.975, median ${\mathrm{\text{AUC}}}_{\mathrm{\text{TFC}}}$: 0.883 IQR: 0.707–0.931). The resultant SDA applied to multi-channel recordings had a median AUC of 0.988 (IQR: 0.931–0.998) when compared to consensus annotations, outperformed two state-of-the-art SDAs $(p<0.001)$ and was noninferior to the human expert for 73/79 of neonates.

Automatic seizure detection is extremely important in the monitoring and diagnosis of epilepsy. The paper presents a novel method based on dictionary pair learning (DPL) for seizure detection in the long-term intracranial electroencephalogram (EEG) recordings. First, for the EEG data, wavelet filtering and differential filtering are applied, and the kernel function is performed to make the signal linearly separable. In DPL, the synthesis dictionary and analysis dictionary are learned jointly from original training samples with alternating minimization method, and sparse coefficients are obtained by using of linear projection instead of costly $l_0$-norm or $l_1$-norm optimization. At last, the reconstructed residuals associated with seizure and nonseizure sub-dictionary pairs are calculated as the decision values, and the postprocessing is performed for improving the recognition rate and reducing the false detection rate of the system. A total of 530h from 20 patients with 81 seizures were used to evaluate the system. Our proposed method has achieved an average segment-based sensitivity of 93.39%, specificity of 98.51%, and event-based sensitivity of 96.36% with false detection rate of 0.236/h.

Brain function has been proposed to arise as a result of the coordinated activity between distributed brain areas. An important issue in the study of brain activity is the characterization of the synchrony among these areas and the resulting complexity of the system. However, the variety of ways to define and, hence, measure brain synchrony and complexity has sometimes led to inconsistent results. Here, we study the relationship between synchrony and commonly used complexity estimators of electroencephalogram (EEG) activity and we explore how simulated lesions in anatomically based cortical networks would affect key functional measures of activity. We explored this question using different types of neural network lesions while the brain dynamics was modeled with a time-delayed set of 66 Kuramoto oscillators. Each oscillator modeled a region of the cortex (node), and the connectivity and spatial location between different areas informed the creation of a network structure (edges). Each type of lesion consisted on successive lesions of nodes or edges during the simulation of the neural dynamics. For each type of lesion, we measured the synchrony among oscillators and three complexity estimators (Higuchi’s Fractal Dimension, Sample Entropy and Lempel-Ziv Complexity) of the simulated EEGs. We found a general negative correlation between EEG complexity metrics and synchrony but Sample Entropy and Lempel-Ziv showed a positive correlation with synchrony when the edges of the network were deleted. This suggests an intricate relationship between synchrony of the system and its estimated complexity. Hence, complexity seems to depend on the multiple states of interaction between the oscillators of the system. Our results can contribute to the interpretation of the functional meaning of EEG complexity.

The study of connectivity patterns of a system’s variables, such as multi-channel electroencephalograms (EEG), is of utmost importance towards a better understanding of its internal evolutionary mechanisms. Here, the problem of estimating the connectivity network from multivariate time series in the presence of prominent unobserved variables is addressed. The causality measure of partial mutual information from mixed embedding (PMIME), designed to estimate direct lag-causal effects in the presence of many observed variables, is adapted to estimate also zero-lag effects, the so-called instantaneous causality. We term the proposed advanced method, PMIME0. The estimation of instantaneous causality by PMIME0 is a signature of the presence of hidden source in the observed system, as demonstrated analytically in a toy model. It is further demonstrated that the PMIME0 identifies the true instantaneous with great accuracy in a variety of high-dimensional dynamical systems. The method is applied to EEG data with epileptiform discharges (EDs), and the results imply a strong impact of unobserved confounders during the EDs. This finding comes as a possible explanation for the increased levels of causality during epileptic seizures estimated by some measures affected by the presence of a common source.

Deep brain stimulation (DBS) of the anterior nucleus of the thalamus (ANT) is a promising treatment for patients with refractory epilepsy. However, therapy response varies and precise positioning of the DBS lead is potentially essential for maximizing therapeutic efficacy. We investigate if single-cell recordings acquired by microelectrode recordings can aid targeting of the ANT during surgery and hypothesize that the neuronal firing properties of the target region relate to clinical outcome. We prospectively included 10 refractory epilepsy patients and performed microelectrode recordings under general anesthesia to identify the change in neuronal signals when approaching and transecting the ANT. The neuronal firing properties of the target region, anatomical locations of microelectrode recordings and active contact positions of the DBS lead along the recorded trajectory were compared between responders and nonresponders to DBS. We obtained 19 sets of recordings from 10 patients (five responders and five nonresponders). Amongst the 403 neurons detected, 365 (90.6%) were classified as bursty. Entry into the ANT was characterized by an increase in firing rate while exit of the ANT was characterized by a decrease in firing rate. Comparing the trajectories of responders to nonresponders, we found differences neither in the neuronal firing properties themselves nor in their locations relative to the position of the active contact. Single-cell firing rate acquired by microelectrode recordings under general anesthesia can thus aid targeting of the ANT during surgery, but is not related to clinical outcome in DBS for patients with refractory epilepsy.