Artefact detection and removal algorithms for EEG diagnostic systems

The electroencephalogram (EEG) is a medical technology that is used in the monitoring of the brain and in the diagnosis of many neurological illnesses. Although coarse in its precision, the EEG is a non-invasive tool that requires minimal set-up times, and is suitably unobtrusive and mobile to allow continuous monitoring of the patient, either in clinical or domestic environments. Consequently, the EEG is the current tool-of-choice with which to continuously monitor the brain where temporal resolution, ease-of- use and mobility are important. Traditionally, EEG data are examined by a trained clinician who identifies neurological events of interest. However, recent advances in signal processing and machine learning techniques have allowed the automated detection of neurological events for many medical applications. In doing so, the burden of work on the clinician has been significantly reduced, improving the response time to illness, and allowing the relevant medical treatment to be administered within minutes rather than hours. However, as typical EEG signals are of the order of microvolts (μV ), contamination by signals arising from sources other than the brain is frequent. These extra-cerebral sources, known as artefacts, can significantly distort the EEG signal, making its interpretation difficult, and can dramatically disimprove automatic neurological event detection classification performance. This thesis therefore, contributes to the further improvement of auto- mated neurological event detection systems, by identifying some of the major obstacles in deploying these EEG systems in ambulatory and clinical environments so that the EEG technologies can emerge from the laboratory towards real-world settings, where they can have a real-impact on the lives of patients. In this context, the thesis tackles three major problems in EEG monitoring, namely: (i) the problem of head-movement artefacts in ambulatory EEG, (ii) the high numbers of false detections in state-of-the-art, automated, epileptiform activity detection systems and (iii) false detections in state-of-the-art, automated neonatal seizure detection systems. To accomplish this, the thesis employs a wide range of statistical, signal processing and machine learning techniques drawn from mathematics, engineering and computer science. The first body of work outlined in this thesis proposes a system to automatically detect head-movement artefacts in ambulatory EEG and utilises supervised machine learning classifiers to do so. The resulting head-movement artefact detection system is the first of its kind and offers accurate detection of head-movement artefacts in ambulatory EEG. Subsequently, addtional physiological signals, in the form of gyroscopes, are used to detect head-movements and in doing so, bring additional information to the head- movement artefact detection task. A framework for combining EEG and gyroscope signals is then developed, offering improved head-movement arte- fact detection. The artefact detection methods developed for ambulatory EEG are subsequently adapted for use in an automated epileptiform activity detection system. Information from support vector machines classifiers used to detect epileptiform activity is fused with information from artefact-specific detection classifiers in order to significantly reduce the number of false detections in the epileptiform activity detection system. By this means, epileptiform activity detection which compares favourably with other state-of-the-art systems is achieved. Finally, the problem of false detections in automated neonatal seizure detection is approached in an alternative manner; blind source separation techniques, complimented with information from additional physiological signals are used to remove respiration artefact from the EEG. In utilising these methods, some encouraging advances have been made in detecting and removing respiration artefacts from the neonatal EEG, and in doing so, the performance of the underlying diagnostic technology is improved, bringing its deployment in the real-world, clinical domain one step closer.

Low-power DWT-based quasi-averaging algorithm and architecture for epileptic seizure detection

In this paper, we have developed a low-complexity algorithm for epileptic seizure detection with a high degree of accuracy. The algorithm has been designed to be feasibly implementable as battery-powered low-power implantable epileptic seizure detection system or epilepsy prosthesis. This is achieved by utilizing design optimization techniques at different levels of abstraction. Particularly, user-specific critical parameters are identified at the algorithmic level and are explicitly used along with multiplier-less implementations at the architecture level. The system has been tested on neural data obtained from in-vivo animal recordings and has been implemented in 90nm bulk-Si technology. The results show up to 90 % savings in power as compared to prevalent wavelet based seizure detection technique while achieving 97% average detection rate. Copyright 2010 ACM.

Reactive oxygen species generated by NADPH oxidase are involved in neurodegeneration in the pilocarpine model of temporal lobe epilepsy

Reactive oxygen species (ROS) appear to be involved in several neurodegenerative disorders. We tested the hypothesis that oxidative stress could have a role in the hippocampal neurodegeneration observed in temporal lobe epilepsy induced by pilocarpine. We first determined the spatio-temporal pattern of ROS generation, by means of detection with dihydroethidium oxidation, in the CA1 and CA3 areas and the dentate gyrus of the dorsal hippocampus during status epilepticus induced by pilocarpine. Fluoro-Jade B assays were also performed to detect degenerating neurons. ROS generation was increased in CA1, CA3 and the dentate gyrus after pilocarpine-induced seizures, which was accompanied by marked cell death. Treatment of rats with a NADPH oxidase inhibitor (apocynin) for 7 days prior to induction of status epilepticus was effective in decreasing both ROS production (by an average of 20%) and neurodegeneration (by an average of 61%). These results suggest an involvement of ROS generated by NADPH oxidase in neuronal death in the pilocarpine model of epilepsy. (C) 2010 Elsevier Ireland Ltd. All rights reserved.

Manganese-enhanced magnetic resonance imaging detects mossy fiber sprouting in the pilocarpine model of epilepsy

Purpose: Mossy fiber sprouting (MFS) is a frequent finding following status epilepticus (SE). The present study aimed to test the feasibility of using manganese-enhanced magnetic resonance imaging (MEMRI) to detect MFS in the chronic phase of the well-established pilocarpine (Pilo) rat model of temporal lobe epilepsy (TLE). Methods: To modulate MFS, cycloheximide (CHX), a protein synthesis inhibitor, was coadministered with Pilo in a subgroup of animals. In vivo MEMRI was performed 3 months after induction of SE and compared to the neo-Timm histologic labeling of zinc mossy fiber terminals in the dentate gyrus (DG). Key Findings: Chronically epileptic rats displaying MFS as detected by neo-Timm histology had a hyperintense MEMRI signal in the DG, whereas chronically epileptic animals that did not display MFS had minimal MEMRI signal enhancement compared to nonepileptic control animals. A strong correlation (r = 0.81, p < 0.001) was found between MEMRI signal enhancement and MFS. Significance: This study shows that MEMRI is an attractive noninvasive method for detection of mossy fiber sprouting in vivo and can be used as an evaluation tool in testing therapeutic approaches to manage chronic epilepsy.

Detection of regional blood perfusion changes in epileptic seizures with dynamic brain perfusion CT--a pilot study

BACKGROUND AND PURPOSE: Perfusion CT (P-CT) is used for acute stroke management, not, however, for evaluating epilepsy. To test the hypothesis that P-CT may identify patients with increased regional cerebral blood flow during subtle status epilepticus (SSE), we compared P-CT in SSE to different postictal conditions. METHODS: Fifteen patients (mean age 47 years, range 21-74) underwent P-CT immediately after evaluation in our emergency room. Asymmetry indices between affected and unaffected hemispheres were calculated for regional cerebral blood volume (rCBV), regional cerebral blood flow (rCBF), and mean transit time (MTT). Regional perfusion changes were compared to EEG findings. RESULTS: Three patients in subtle status epilepticus (group 1) had increased regional perfusion with electro-clinical correlate. Six patients showed postictal slowing on EEG corresponding to an area of regional hypoperfusion (group 2). CT and EEG were normal in six patients with a first epileptic seizure (group 3). Cluster analysis of asymmetry indices separated SSE from the other two groups in all three parameters, while rCBF helped to distinguish between chronic focal epilepsies and single events. CONCLUSION: Preliminary results indicate that P-CT may help to identify patients with SSE during emergency workup. This technique provides important information to neurologists or emergency physicians in the difficult clinical differential diagnosis of altered mental status due to subtle status epilepticus.

Comparison of voxel-based 3-D MRI analysis and subtraction ictal SPECT coregistered to MRI in focal epilepsy

While voxel-based 3-D MRI analysis methods as well as assessment of subtracted ictal versus interictal perfusion studies (SISCOM) have proven their potential in the detection of lesions in focal epilepsy, a combined approach has not yet been reported. The present study investigates if individual automated voxel-based 3-D MRI analyses combined with SISCOM studies contribute to an enhanced detection of mesiotemporal epileptogenic foci. Seven consecutive patients with refractory complex partial epilepsy were prospectively evaluated by SISCOM and voxel-based 3-D MRI analysis. The functional perfusion maps and voxel-based statistical maps were coregistered in 3-D space. In five patients with temporal lobe epilepsy (TLE), the area of ictal hyperperfusion and corresponding structural abnormalities detected by 3-D MRI analysis were identified within the same temporal lobe. In two patients, additional structural and functional abnormalities were detected beyond the mesial temporal lobe. Five patients with TLE underwent epileptic surgery with favourable postoperative outcome (Engel class Ia and Ib) after 3-5 years of follow-up, while two patients remained on conservative treatment. In summary, multimodal assessment of structural abnormalities by voxel-based analysis and SISCOM may contribute to advanced observer-independent preoperative assessment of seizure origin.

Focal Epilepsy: MR Imaging of Nonhemodynamic Field Effects by Using a Phase-cycled Stimulus-induced Rotary Saturation Approach with Spin-Lock Preparation.

Purpose To investigate whether nonhemodynamic resonant saturation effects can be detected in patients with focal epilepsy by using a phase-cycled stimulus-induced rotary saturation (PC-SIRS) approach with spin-lock (SL) preparation and whether they colocalize with the seizure onset zone and surface interictal epileptiform discharges (IED). Materials and Methods The study was approved by the local ethics committee, and all subjects gave written informed consent. Eight patients with focal epilepsy undergoing presurgical surface and intracranial electroencephalography (EEG) underwent magnetic resonance (MR) imaging at 3 T with a whole-brain PC-SIRS imaging sequence with alternating SL-on and SL-off and two-dimensional echo-planar readout. The power of the SL radiofrequency pulse was set to 120 Hz to sensitize the sequence to high gamma oscillations present in epileptogenic tissue. Phase cycling was applied to capture distributed current orientations. Voxel-wise subtraction of SL-off from SL-on images enabled the separation of T2* effects from rotary saturation effects. The topography of PC-SIRS effects was compared with the seizure onset zone at intracranial EEG and with surface IED-related potentials. Bayesian statistics were used to test whether prior PC-SIRS information could improve IED source reconstruction. Results Nonhemodynamic resonant saturation effects ipsilateral to the seizure onset zone were detected in six of eight patients (concordance rate, 0.75; 95% confidence interval: 0.40, 0.94) by means of the PC-SIRS technique. They were concordant with IED surface negativity in seven of eight patients (0.88; 95% confidence interval: 0.51, 1.00). Including PC-SIRS as prior information improved the evidence of the standard EEG source models compared with the use of uninformed reconstructions (exceedance probability, 0.77 vs 0.12; Wilcoxon test of model evidence, P < .05). Nonhemodynamic resonant saturation effects resolved in patients with favorable postsurgical outcomes, but persisted in patients with postsurgical seizure recurrence. Conclusion Nonhemodynamic resonant saturation effects are detectable during interictal periods with the PC-SIRS approach in patients with epilepsy. The method may be useful for MR imaging-based detection of neuronal currents in a clinical environment. (©) RSNA, 2016 Online supplemental material is available for this article.

Non uniform embedding based on relevance analysis with reduced computational complexity: application to the detection of pathologies from biosignal recordings

Nonlinear analysis tools for studying and characterizing the dynamics of physiological signals have gained popularity, mainly because tracking sudden alterations of the inherent complexity of biological processes might be an indicator of altered physiological states. Typically, in order to perform an analysis with such tools, the physiological variables that describe the biological process under study are used to reconstruct the underlying dynamics of the biological processes. For that goal, a procedure called time-delay or uniform embedding is usually employed. Nonetheless, there is evidence of its inability for dealing with non-stationary signals, as those recorded from many physiological processes. To handle with such a drawback, this paper evaluates the utility of non-conventional time series reconstruction procedures based on non uniform embedding, applying them to automatic pattern recognition tasks. The paper compares a state of the art non uniform approach with a novel scheme which fuses embedding and feature selection at once, searching for better reconstructions of the dynamics of the system. Moreover, results are also compared with two classic uniform embedding techniques. Thus, the goal is comparing uniform and non uniform reconstruction techniques, including the one proposed in this work, for pattern recognition in biomedical signal processing tasks. Once the state space is reconstructed, the scheme followed characterizes with three classic nonlinear dynamic features (Largest Lyapunov Exponent, Correlation Dimension and Recurrence Period Density Entropy), while classification is carried out by means of a simple k-nn classifier. In order to test its generalization capabilities, the approach was tested with three different physiological databases (Speech Pathologies, Epilepsy and Heart Murmurs). In terms of the accuracy obtained to automatically detect the presence of pathologies, and for the three types of biosignals analyzed, the non uniform techniques used in this work lightly outperformed the results obtained using the uniform methods, suggesting their usefulness to characterize non-stationary biomedical signals in pattern recognition applications. On the other hand, in view of the results obtained and its low computational load, the proposed technique suggests its applicability for the applications under study.

Non-invasive Detection of Cortical Control Signals for Brain-Computer Interfaces

Thesis (Ph.D.)--University of Washington, 2016-05

A research platform for artificial neural networks with applications in pediatric epilepsy

This dissertation established a state-of-the-art programming tool for designing and training artificial neural networks (ANNs) and showed its applicability to brain research. The developed tool, called NeuralStudio, allows users without programming skills to conduct studies based on ANNs in a powerful and very user friendly interface. A series of unique features has been implemented in NeuralStudio, such as ROC analysis, cross-validation, network averaging, topology optimization, and optimization of the activation function’s slopes. It also included a Support Vector Machines module for comparison purposes. Once the tool was fully developed, it was applied to two studies in brain research. In the first study, the goal was to create and train an ANN to detect epileptic seizures from subdural EEG. This analysis involved extracting features from the spectral power in the gamma frequencies. In the second application, a unique method was devised to link EEG recordings to epileptic and nonepileptic subjects. The contribution of this method consisted of developing a descriptor matrix that can be used to represent any EEG file regarding its duration and the number of electrodes. The first study showed that the inter-electrode mean of the spectral power in the gamma frequencies and its duration above a specific threshold performs better than the other frequencies in seizure detection, exhibiting an accuracy of 95.90%, a sensitivity of 92.59%, and a specificity of 96.84%. The second study yielded that Hjorth’s parameter activity is sufficient to accurately relate EEG to epileptic and non-epileptic subjects. After testing, accuracy, sensitivity and specificity of the classifier were all above 0.9667. Statistical tests measured the superiority of activity at over 99.99 % certainty. It was demonstrated that (1) the spectral power in the gamma frequencies is highly effective in locating seizures from EEG and (2) activity can be used to link EEG recordings to epileptic and non-epileptic subjects. These two studies required high computational load and could be addressed thanks to NeuralStudio. From a medical perspective, both methods proved the merits of NeuralStudio in brain research applications. For its outstanding features, NeuralStudio has been recently awarded a patent (US patent No. 7502763).

A Research Platform for Artificial Neural Networks with Applications in Pediatric Epilepsy

This dissertation established a state-of-the-art programming tool for designing and training artificial neural networks (ANNs) and showed its applicability to brain research. The developed tool, called NeuralStudio, allows users without programming skills to conduct studies based on ANNs in a powerful and very user friendly interface. A series of unique features has been implemented in NeuralStudio, such as ROC analysis, cross-validation, network averaging, topology optimization, and optimization of the activation function’s slopes. It also included a Support Vector Machines module for comparison purposes. Once the tool was fully developed, it was applied to two studies in brain research. In the first study, the goal was to create and train an ANN to detect epileptic seizures from subdural EEG. This analysis involved extracting features from the spectral power in the gamma frequencies. In the second application, a unique method was devised to link EEG recordings to epileptic and non-epileptic subjects. The contribution of this method consisted of developing a descriptor matrix that can be used to represent any EEG file regarding its duration and the number of electrodes. The first study showed that the inter-electrode mean of the spectral power in the gamma frequencies and its duration above a specific threshold performs better than the other frequencies in seizure detection, exhibiting an accuracy of 95.90%, a sensitivity of 92.59%, and a specificity of 96.84%. The second study yielded that Hjorth’s parameter activity is sufficient to accurately relate EEG to epileptic and non-epileptic subjects. After testing, accuracy, sensitivity and specificity of the classifier were all above 0.9667. Statistical tests measured the superiority of activity at over 99.99 % certainty. It was demonstrated that 1) the spectral power in the gamma frequencies is highly effective in locating seizures from EEG and 2) activity can be used to link EEG recordings to epileptic and non-epileptic subjects. These two studies required high computational load and could be addressed thanks to NeuralStudio. From a medical perspective, both methods proved the merits of NeuralStudio in brain research applications. For its outstanding features, NeuralStudio has been recently awarded a patent (US patent No. 7502763).