Spectral and nonlinear analyses of MEG background activity in patients with Alzheimer's disease

Source imaging method based on diagonal covariance bases and its applications to OPM-MEG

Magnetoencephalography (MEG) is a noninvasive imaging technique used in neuroscience and clinical research. The source estimation of MEG involves solving a highly underdetermined inverse problem, which requires additional constraints to restrict the solution space. Traditional methods tend to obscure the extent of the sources. However, an accurate estimation of the source extent is important for studying brain activity or preoperatively estimating pathogenic regions. To improve the estimation accuracy of the extended source extent, the spatial constraint of sources is employed in the Bayesian framework. For example, the source is decomposed into a linear combination of validated spatial basis functions, which is proved to improve the source imaging accuracy. In this work, we further construct the spatial properties of the source using the diagonal covariance bases (DCB), which we summarize as the source imaging method SI-DCB. In this approach, specifically, the covariance matrix of the spatial coefficients is modeled as a weighted combination of diagonal covariance basis functions. The convex analysis is used to estimate noise and model parameters under the Bayesian framework. Extensive numerical simulations showed that SI-DCB outperformed five benchmark methods in accurately estimating the location and extent of patch sources. The effectiveness of SI-DCB was verified through somatosensory stimulation experiments performed on a 31-channel OPM-MEG system. The SI-DCB correctly identified the source area where each brain response occurred. The superior performance of SI-DCB suggests that it can provide a template approach for improving the accuracy of source extent estimations under a sparse Bayesian framework.

Entropy and fractal analysis of brain-related neurophysiological signals in Alzheimer's and Parkinson's disease

Brain-related neuronal recordings, such as local field potential, electroencephalogram and magnetoencephalogram, offer the opportunity to study the complexity of the human brain at different spatial and temporal scales. The complex properties of neuronal signals are intrinsically related to the concept of 'scale-free' behavior and irregular dynamic, which cannot be fully described through standard linear methods, but can be measured by nonlinear indexes. A remarkable application of these analysis methods on electrophysiological recordings is the deep comprehension of the pathophysiology of neurodegenerative diseases, that has been shown to be associated to changes in brain activity complexity. In particular, a decrease of global complexity has been associated to Alzheimer's disease, while a local increase of brain signals complexity characterizes Parkinson's disease. Despite the recent proliferation of studies using fractal and entropy-based analysis, the application of these techniques is still far from clinical practice, due to the lack of an agreement about their correct estimation and a conclusive and shared interpretation. Along with the aim of helping towards the realization of a multidisciplinary audience to approach nonlinear methods based on the concepts of fractality and irregularity, this survey describes the implementation and proper employment of the mostly known and applied indexes in the context of Alzheimer's and Parkinson's diseases.

Computational Modeling of Therapy with the NMDA Antagonist in Neurodegenerative Disease: Information Theory in the Mechanism of Action of Memantine

(1) Background: in patients with neurodegenerative diseases, noncompetitive N-methyl-D-aspartate (NMDA) receptor antagonists provide neuroprotective advantages. We performed memantine therapy and proved mathematical and computer modeling of neurodegenerative disease in this study. (2) Methods: a computer simulation environment of the N-methyl-D-aspartate receptor incorporating biological mechanisms of channel activation by high extracellular glutamic acid concentration. In comparison to controls, pathological models were essentially treated with doses of memantine 3−30 µM. (3) Results: the mean values and 95% CI for Shannon entropy in Alzheimer’s disease (AD) and memantine treatment models were 1.760 (95% CI, 1.704−1.818) vs. 2.385 (95% CI, 2.280−2.490). The Shannon entropy was significantly higher in the memantine treatment model relative to AD model (p = 0.0162). The mean values and 95% CI for the positive Lyapunov exponent in AD and memantine treatment models were 0.125 (95% CI, NE−NE) vs. 0.058 (95% CI, 0.044−0.073). The positive Lyapunov exponent was significantly higher in the AD model relative to the memantine treatment model (p = 0.0091). The mean values and 95% CI for transfer entropy in AD and memantine treatment models were 0.081 (95% CI, 0.048−0.114) vs. 0.040 (95% CI, 0.019−0.062). The transfer entropy was significantly higher in the AD model relative to the memantine treatment model (p = 0.0146). A correlation analysis showed positive and statistically significant correlations of the memantine concentrations and the positive Lyapunov exponent (correlation coefficient R = 0.87, p = 0.0023) and transfer entropy (TE) (correlation coefficient R = 0.99, p < 0.000001). (4) Conclusions: information theory results of simulation studies show that the NMDA antagonist, memantine, causes neuroprotective benefits in patients with AD. Our simulation study opens up remarkable new scenarios in which a medical product, drug, or device, can be developed and tested for efficacy based on parameters of information theory.

Mining the Mind: Linear Discriminant Analysis of MEG Source Reconstruction Time Series Supports Dynamic Changes in Deep Brain Regions During Meditation Sessions

Meditation practices have been claimed to have a positive effect on the regulation of mood and emotions for quite some time by practitioners, and in recent times there has been a sustained effort to provide a more precise description of the influence of meditation on the human brain. Longitudinal studies have reported morphological changes in cortical thickness and volume in selected brain regions due to meditation practice, which is interpreted as an evidence its effectiveness beyond the subjective self reporting. Using magnetoencephalography (MEG) or electroencephalography to quantify the changes in brain activity during meditation practice represents a challenge, as no clear hypothesis about the spatial or temporal pattern of such changes is available to date. In this article we consider MEG data collected during meditation sessions of experienced Buddhist monks practicing focused attention (Samatha) and open monitoring (Vipassana) meditation, contrasted by resting state with eyes closed. The MEG data are first mapped to time series of brain activity averaged over brain regions corresponding to a standard Destrieux brain atlas. Next, by bootstrapping and spectral analysis, the data are mapped to matrices representing random samples of power spectral densities in [Formula: see text], [Formula: see text], [Formula: see text], and [Formula: see text] frequency bands. We use linear discriminant analysis to demonstrate that the samples corresponding to different meditative or resting states contain enough fingerprints of the brain state to allow a separation between different states, and we identify the brain regions that appear to contribute to the separation. Our findings suggest that the cingulate cortex, insular cortex and some of the internal structures, most notably the accumbens, the caudate and the putamen nuclei, the thalamus and the amygdalae stand out as separating regions, which seems to correlate well with earlier findings based on longitudinal studies.

Exploring the Interactions Between Neurophysiology and Cognitive and Behavioral Changes Induced by a Non-pharmacological Treatment: A Network Approach

Dementia due to Alzheimer's disease (AD) is a neurological syndrome which has an increasing impact on society, provoking behavioral, cognitive, and functional impairments. AD lacks an effective pharmacological intervention; thereby, non-pharmacological treatments (NPTs) play an important role, as they have been proven to ameliorate AD symptoms. Nevertheless, results associated with NPTs are patient-dependent, and new tools are needed to predict their outcome and to improve their effectiveness. In the present study, 19 patients with AD underwent an NPT for 83.1 ± 38.9 days (mean ± standard deviation). The NPT was a personalized intervention with physical, cognitive, and memory stimulation. The magnetoencephalographic activity was recorded at the beginning and at the end of the NPT to evaluate the neurophysiological state of each patient. Additionally, the cognitive (assessed by means of the Mini-Mental State Examination, MMSE) and behavioral (assessed in terms of the Dementia Behavior Disturbance Scale, DBD-13) status were collected before and after the NPT. We analyzed the interactions between cognitive, behavioral, and neurophysiological data by generating diverse association networks, able to intuitively characterize the relationships between variables of a different nature. Our results suggest that the NPT remarkably changed the structure of the association network, reinforcing the interactions between the DBD-13 and the neurophysiological parameters. We also found that the changes in cognition and behavior are related to the changes in spectral-based neurophysiological parameters. Furthermore, our results support the idea that MEG-derived parameters can predict NPT outcome; specifically, a lesser degree of AD neurophysiological alterations (i.e., neural oscillatory slowing, decreased variety of spectral components, and increased neural signal regularity) predicts a better NPT prognosis. This study provides deeper insights into the relationships between neurophysiology and both, cognitive and behavioral status, proving the potential of network-based methodology as a tool to further understand the complex interactions elicited by NPTs.

Magnetoencephalography: physics, techniques, and applications in the basic and clinical neurosciences

Magnetoencephalography (MEG) is a technique used to measure the magnetic fields generated from neuronal activity in the brain. MEG has a high temporal resolution on the order of milliseconds and provides a more direct measure of brain activity when compared with hemodynamic-based neuroimaging methods such as magnetic resonance imaging and positron emission tomography. The current review focuses on basic features of MEG such as the instrumentation and the physics that are integral to the signals that can be measured, and the principles of source localization techniques, particularly the physics of beamforming and the techniques that are used to localize the signal of interest. In addition, we review several metrics that can be used to assess functional coupling in MEG and describe the advantages and disadvantages of each approach. Lastly, we discuss the current and future applications of MEG.

Consistency of local activation parameters at sensor- and source-level in neural signals

OBJECTIVE

Although magnetoencephalography and electroencephalography (M/EEG) signals at sensor level are robust and reliable, they suffer from different degrees of distortion due to changes in brain tissue conductivities, known as field spread and volume conduction effects. To estimate original neural generators from M/EEG activity acquired at sensor level, diverse source localisation algorithms have been proposed; however, they are not exempt from limitations and usually involve time-consuming procedures. Connectivity and network-based M/EEG analyses have been found to be affected by field spread and volume conduction effects; nevertheless, the influence of the aforementioned effects on widely used local activation parameters has not been assessed yet. The goal of this study is to evaluate the consistency of various local activation parameters when they are computed at sensor- and source-level.

APPROACH

Six spectral (relative power, median frequency, and individual alpha frequency) and non-linear parameters (Lempel-Ziv complexity, sample entropy, and central tendency measure) are computed from M/EEG signals at sensor- and source-level using four source inversion methods: weighted minimum norm estimate (wMNE), standardised low-resolution brain electromagnetic tomography (sLORETA), linear constrained minimum variance (LCMV), and dynamical statistical parametric mapping (dSPM).

MAIN RESULTS

Our results show that the spectral and non-linear parameters yield similar results at sensor- and source-level, showing high correlation values between them for all the source inversion methods evaluated and both modalities of signal, EEG and MEG. Furthermore, the correlation values remain high when performing coarse-grained spatial analyses.

SIGNIFICANCE

To the best of our knowledge, this is the first study analysing how field spread and volume conduction effects impact on local activation parameters computed from resting-state neural activity. Our findings evidence that local activation parameters are robust against field spread and volume conduction effects and provide equivalent information at sensor- and source-level even when performing regional analyses.

Relationship between the Presence of the ApoE ε4 Allele and EEG Complexity along the Alzheimer's Disease Continuum

Alzheimer's disease (AD) is the most prevalent cause of dementia, being considered a major health problem, especially in developed countries. Late-onset AD is the most common form of the disease, with symptoms appearing after 65 years old. Genetic determinants of AD risk are vastly unknown, though, ε 4 allele of the ApoE gene has been reported as the strongest genetic risk factor for AD. The objective of this study was to analyze the relationship between brain complexity and the presence of ApoE ε 4 alleles along the AD continuum. For this purpose, resting-state electroencephalography (EEG) activity was analyzed by computing Lempel-Ziv complexity (LZC) from 46 healthy control subjects, 49 mild cognitive impairment subjects, 45 mild AD patients, 44 moderate AD patients and 33 severe AD patients, subdivided by ApoE status. Subjects with one or more ApoE ε 4 alleles were included in the carriers subgroups, whereas the ApoE ε 4 non-carriers subgroups were formed by subjects without any ε 4 allele. Our results showed that AD continuum is characterized by a progressive complexity loss. No differences were observed between AD ApoE ε 4 carriers and non-carriers. However, brain activity from healthy subjects with ApoE ε 4 allele (carriers subgroup) is more complex than from non-carriers, mainly in left temporal, frontal and posterior regions (p-values < 0.05, FDR-corrected Mann-Whitney U-test). These results suggest that the presence of ApoE ε 4 allele could modify the EEG complexity patterns in different brain regions, as the temporal lobes. These alterations might be related to anatomical changes associated to neurodegeneration, increasing the risk of suffering dementia due to AD before its clinical onset. This interesting finding might help to advance in the development of new tools for early AD diagnosis.

Altered complexity of resting-state BOLD activity in Alzheimer's disease-related neurodegeneration: a multiscale entropy analysis

Brain complexity, which reflects the ability of the brain to adapt to a changing environment, has been found to be significantly changed with age. However, there is less evidence on the alterations of brain complexity in neurodegenerative disorders such as Alzheimer's disease (AD). Here we investigated the altered complexity of resting-state blood oxygen level-dependent signals in AD-related neurodegeneration using multiscale entropy (MSE) analysis. All participants were recruited from the Alzheimer's Disease Neuroimaging Initiative, including healthy controls (HC, n=62), amnestic mild cognitive impairment (aMCI, n =81) patients, and Alzheimer's disease (AD, n=25) patients. Our results showed time scale-dependent MSE differences across the three groups. In scale=1, significantly changed MSE patterns (HC>aMCI>AD) were found in four brain regions, including the hippocampus, middle frontal gyrus, intraparietal lobe, and superior frontal gyrus. In scale=4, reversed MSE patterns (HC<aMCI<AD) were found in the middle frontal gyrus and middle occipital gyrus. Furthermore, the values of regional entropy were significantly associated with cognitive functions positively on the short time scale, while negatively on the longer time scale. Our findings suggest that MSE could be a reliable measure for characterizing brain deterioration in AD, and may provide insights into the neural mechanism of AD-related neurodegeneration.

Complexity Analysis of EEG, MEG, and fMRI in Mild Cognitive Impairment and Alzheimer's Disease: A Review

Alzheimer's disease (AD) is a degenerative brain disease with a high and irreversible incidence. In recent years, because brain signals have complex nonlinear dynamics, there has been growing interest in studying complex changes in the time series of brain signals in patients with AD. We reviewed studies of complexity analyses of single-channel time series from electroencephalogram (EEG), magnetoencephalogram (MEG), and functional magnetic resonance imaging (fMRI) in AD and determined future research directions. A systematic literature search for 2000-2019 was performed in the Web of Science and PubMed databases, resulting in 126 identified studies. Compared to healthy individuals, the signals from AD patients have less complexity and more predictable oscillations, which are found mainly in the left parietal, occipital, right frontal, and temporal regions. This complexity is considered a potential biomarker for accurately responding to the functional lesion in AD. The current review helps to reveal the patterns of dysfunction in the brains of patients with AD and to investigate whether signal complexity can be used as a biomarker to accurately respond to the functional lesion in AD. We proposed further studies in the signal complexities of AD patients, including investigating the reliability of complexity algorithms and the spatial patterns of signal complexity. In conclusion, the current review helps to better understand the complexity of abnormalities in the AD brain and provide useful information for AD diagnosis.

Effects of Inducing Gamma Oscillations in Hippocampal Subregions DG, CA3, and CA1 on the Potential Alleviation of Alzheimer's Disease-Related Pathology: Computer Modeling and Simulations

The aim of this study was to evaluate the possibility of the gamma oscillation function (40–130 Hz) to reduce Alzheimer’s disease related pathology in a computer model of the hippocampal network dentate gyrus, CA3, and CA1 (DG-CA3-CA1) regions. Methods: Computer simulations were made for a pathological model in which Alzheimer’s disease was simulated by synaptic degradation in the hippocampus. Pathology modeling was based on sequentially turning off the connections with entorhinal cortex layer 2 (EC2) and the dentate gyrus on CA3 pyramidal neurons. Gamma induction modeling consisted of simulating the oscillation provided by the septo-hippocampal pathway with band frequencies from 40–130 Hz. Pathological models with and without gamma induction were compared with a control. Results: In the hippocampal regions of DG, CA3, and CA1, and jointly DG-CA3-CA1 and CA3-CA1, gamma induction resulted in a statistically significant improvement in terms of increased numbers of spikes, spikes per burst, and burst duration as compared with the model simulating Alzheimer’s disease (AD). The positive maximal Lyapunov exponent was negative in both the control model and the one with gamma induction as opposed to the pathological model where it was positive within the DG-CA3-CA1 region. Gamma induction resulted in decreased transfer entropy in accordance with the information flow in DG → CA3 and CA3 → CA1. Conclusions: The results of simulation studies show that inducing gamma oscillations in the hippocampus may reduce Alzheimer’s disease related pathology. Pathologically higher transfer entropy values after gamma induction returned to values comparable to the control model.

Computer Model of Synapse Loss During an Alzheimer's Disease-Like Pathology in Hippocampal Subregions DG, CA3 and CA1-The Way to Chaos and Information Transfer

The aim of the study was to compare the computer model of synaptic breakdown in an Alzheimer’s disease-like pathology in the dentate gyrus (DG), CA3 and CA1 regions of the hippocampus with a control model using neuronal parameters and methods describing the complexity of the system, such as the correlative dimension, Shannon entropy and positive maximal Lyapunov exponent. The model of synaptic breakdown (from 13% to 50%) in the hippocampus modeling the dynamics of an Alzheimer’s disease-like pathology was simulated. Modeling consisted in turning off one after the other EC2 connections and connections from the dentate gyrus on the CA3 pyramidal neurons. The pathological model of synaptic disintegration was compared to a control. The larger synaptic breakdown was associated with a statistically significant decrease in the number of spikes (R = −0.79, P < 0.001), spikes per burst (R = −0.76, P < 0.001) and burst duration (R = −0.83, P < 0.001) and an increase in the inter-burst interval (R = 0.85, P < 0.001) in DG-CA3-CA1. The positive maximal Lyapunov exponent in the control model was negative, but in the pathological model had a positive value of DG-CA3-CA1. A statistically significant decrease of Shannon entropy with the direction of information flow DG->CA3->CA1 (R = −0.79, P < 0.001) in the pathological model and a statistically significant increase with greater synaptic breakdown (R = 0.24, P < 0.05) of the CA3-CA1 region was obtained. The reduction of entropy transfer for DG->CA3 at the level of synaptic breakdown of 35% was 35%, compared with the control. Entropy transfer for CA3->CA1 at the level of synaptic breakdown of 35% increased to 95% relative to the control. The synaptic breakdown model in an Alzheimer’s disease-like pathology in DG-CA3-CA1 exhibits chaotic features as opposed to the control. Synaptic breakdown in which an increase of Shannon entropy is observed indicates an irreversible process of Alzheimer’s disease. The increase in synapse loss resulted in decreased information flow and entropy transfer in DG->CA3, and at the same time a strong increase in CA3->CA1.

M/EEG-Based Bio-Markers to Predict the MCI and Alzheimer's Disease: A Review From the ML Perspective

This paper reviews the state-of-the-art neuromarkers development for the prognosis of Alzheimer's disease (AD) and mild cognitive impairment (MCI). The first part of this paper is devoted to reviewing the recently emerged machine learning (ML) algorithms based on electroencephalography (EEG) and magnetoencephalography (MEG) modalities. In particular, the methods are categorized by different types of neuromarkers. The second part of the review is dedicated to a series of investigations that further highlight the differences between these two modalities. First, several source reconstruction methods are reviewed and their source-level performances explored, followed by an objective comparison between EEG and MEG from multiple perspectives. Finally, a number of the most recent reports on classification of MCI/AD during resting state using EEG/MEG are documented to show the up-to-date performance for this well-recognized data collecting scenario. It is noticed that the MEG modality may be particularly effective in distinguishing between subjects with MCI and healthy controls, a high classification accuracy of more than 98% was reported recently; whereas the EEG seems to be performing well in classifying AD and healthy subjects, which also reached around 98% of the accuracy. A number of influential factors have also been raised and suggested for careful considerations while evaluating the ML-based diagnosis systems in the real-world scenarios.

Protection of Superconducting Industrial Machinery Using RNN-Based Anomaly Detection for Implementation in Smart Sensor

Sensing the voltage developed over a superconducting object is very important in order to make superconducting installation safe. An increase in the resistive part of this voltage (quench) can lead to significant deterioration or even to the destruction of the superconducting device. Therefore, detection of anomalies in time series of this voltage is mandatory for reliable operation of superconducting machines. The largest superconducting installation in the world is the main subsystem of the Large Hadron Collider (LHC) accelerator. Therefore a protection system was built around superconducting magnets. Currently, the solutions used in protection equipment at the LHC are based on a set of hand-crafted custom rules. They were proved to work effectively in a range of applications such as quench detection. However, these approaches lack scalability and require laborious manual adjustment of working parameters. The presented work explores the possibility of using the embedded Recurrent Neural Network as a part of a protection device. Such an approach can scale with the number of devices and signals in the system, and potentially can be automatically configured to given superconducting magnet working conditions and available data. In the course of the experiments, it was shown that the model using Gated Recurrent Units (GRU) comprising of two layers with 64 and 32 cells achieves 0.93 accuracy for anomaly/non-anomaly classification, when employing custom data compression scheme. Furthermore, the compression of proposed module was tested, and showed that the memory footprint can be reduced four times with almost no performance loss, making it suitable for hardware implementation.

Measuring Alterations of Spontaneous EEG Neural Coupling in Alzheimer's Disease and Mild Cognitive Impairment by Means of Cross-Entropy Metrics

Alzheimer's Disease (AD) represents the most prevalent form of dementia and is considered a major health problem due to its high prevalence and its economic costs. An accurate characterization of the underlying neural dynamics in AD is crucial in order to adopt effective treatments. In this regard, mild cognitive impairment (MCI) is an important clinical entity, since it is a risk-state for developing dementia. In the present study, coupling patterns of 111 resting-state electroencephalography (EEG) recordings were analyzed. Specifically, we computed Cross-Approximate Entropy (Cross-ApEn) and Cross-Sample Entropy (Cross-SampEn) of 37 patients with dementia due to AD, 37 subjects with MCI, and 37 healthy control (HC) subjects. Our results showed that Cross-SampEn outperformed Cross-ApEn, revealing higher number of significant connections among the three groups (Kruskal-Wallis test, FDR-corrected p-values < 0.05). AD patients exhibited statistically significant lower similarity values at θ and β1 frequency bands compared to HC. MCI is also characterized by a global decrease of similarity in all bands, being only significant at β1. These differences shows that β band might play a significant role in the identification of early stages of AD. Our results suggest that Cross-SampEn could increase the insight into brain dynamics at different AD stages. Consequently, it may contribute to develop early AD biomarkers, potentially useful as diagnostic information.

Analysis of long range dependence in the EEG signals of Alzheimer patients

Alzheimer's disease (AD), a cognitive disability is analysed using a long range dependence parameter, hurst exponent (HE), calculated based on the time domain analysis of the measured electrical activity of brain. The electroencephalogram (EEG) signals of controls and mild cognitive impairment (MCI)-AD patients are evaluated under normal resting and mental arithmetic conditions. Simultaneous low pass filtering and total variation denoising algorithm is employed for preprocessing. Larger values of HE observed in the right hemisphere of the brain for AD patients indicated a decrease in irregularity of the EEG signal under cognitive task conditions. Correlations between HE and the neuropsychological indices are analysed using bivariate correlation analysis. The observed reduction in the values of Auto mutual information and cross mutual information in the local antero-frontal and distant regions in the brain hemisphere indicates the loss of information transmission in MCI-AD patients.

Analysis of long range dependence in the EEG signals of Alzheimer patients

Alzheimer's disease (AD), a cognitive disability is analysed using a long range dependence parameter, hurst exponent (HE), calculated based on the time domain analysis of the measured electrical activity of brain. The electroencephalogram (EEG) signals of controls and mild cognitive impairment (MCI)-AD patients are evaluated under normal resting and mental arithmetic conditions. Simultaneous low pass filtering and total variation denoising algorithm is employed for preprocessing. Larger values of HE observed in the right hemisphere of the brain for AD patients indicated a decrease in irregularity of the EEG signal under cognitive task conditions. Correlations between HE and the neuropsychological indices are analysed using bivariate correlation analysis. The observed reduction in the values of Auto mutual information and cross mutual information in the local antero-frontal and distant regions in the brain hemisphere indicates the loss of information transmission in MCI-AD patients.

Comparison of background EEG activity of different groups of patients with idiopathic epilepsy using Shannon spectral entropy and cluster-based permutation statistical testing

Idiopathic epilepsy is characterized by generalized seizures with no apparent cause. One of its main problems is the lack of biomarkers to monitor the evolution of patients. The only tools they can use are limited to inspecting the amount of seizures during previous periods of time and assessing the existence of interictal discharges. As a result, there is a need for improving the tools to assist the diagnosis and follow up of these patients. The goal of the present study is to compare and find a way to differentiate between two groups of patients suffering from idiopathic epilepsy, one group that could be followed-up by means of specific electroencephalographic (EEG) signatures (intercritical activity present), and another one that could not due to the absence of these markers. To do that, we analyzed the background EEG activity of each in the absence of seizures and epileptic intercritical activity. We used the Shannon spectral entropy (SSE) as a metric to discriminate between the two groups and performed permutation-based statistical tests to detect the set of frequencies that show significant differences. By constraining the spectral entropy estimation to the [6.25–12.89) Hz range, we detect statistical differences (at below 0.05 alpha-level) between both types of epileptic patients at all available recording channels. Interestingly, entropy values follow a trend that is inversely related to the elapsed time from the last seizure. Indeed, this trend shows asymptotical convergence to the SSE values measured in a group of healthy subjects, which present SSE values lower than any of the two groups of patients. All these results suggest that the SSE, measured in a specific range of frequencies, could serve to follow up the evolution of patients suffering from idiopathic epilepsy. Future studies remain to be conducted in order to assess the predictive value of this approach for the anticipation of seizures.

Electrocardiographic Spectral Features for Long-Term Outcome Prognosis of Atrial Fibrillation Catheter Ablation

Atrial fibrillation (AF) is the most common arrhythmia in routine clinical practice. Despite many years of research, its mechanisms still are not well understood, thus reducing the effectiveness of AF treatments. Nowadays, pulmonary vein isolation by catheter ablation is the treatment of choice for AF resistant either to pharmacological or electrical cardioversion. However, given that long-term recurrences are common, an appropriate patient selection before the procedure is of paramount relevance in the improvement of AF catheter ablation outcome. The present work studies how several spectral features of the atrial activity (AA) from a single lead of the surface electrocardiogram (ECG) can become potential pre-ablation predictors of long-term (>2 months) sinus rhythm maintenance. Among all the analyzed spectral features, results indicated that the most significant single predictor of paroxysmal AF ablation treatment outcome was related to the amplitude of the first harmonic of the dominant frequency, providing sensitivity (Se), specificity (Sp) and accuracy (Ac) values of 90%, 42.86 and 77.78%, respectively. On the other hand, the AA harmonic structure was the most significant single predictor for persistent AF, with Se, Sp and Ac values of 100%, 54.55 and 77.27%, respectively. A logistic regression analysis, mainly based on spectral amplitudes as well as on the harmonic structure of the AA, provided a higher predictive ability both for paroxysmal AF (Se = 100%, Sp = 57.14% and Ac = 88.89%) and persistent AF (Se = 90.91%, Sp = 72.73 and Ac = 81.82%). In conclusion, the study of key AA spectral features from the surface ECG can provide a significant preoperative prognosis of AF catheter ablation outcome at long-term follow-up.

Spectral analysis of resting state magnetoencephalogram activity in patients with bipolar disorder

We analyzed the frequency spectrum of magnetoencephalogram (MEG) background activity in 16 bipolar disorder (BD) patients and 24 age-matched healthy control. Median frequency (MF), spectral entropy (SpEn), and relative power in delta (RPδ), theta (RPθ), alpha (RPα), beta (RPβ), and gamma (RPγ) bands were computed for all 148 MEG channels. Significant differences between the two groups were found in the average level of MF, RPδ, and RPθ in the posterior region of the scalp. Moreover, the MF, SpEn, RPδ, and RPβ values of BD patients had a different dependence on age as compared with the results of control subjects, which may suggest that BD affects how the brain activity develops with age. We conclude that the spectral analysis of the background MEG in BD patients may give insights into how this condition affects the brain activity.

Inclusion of Neuropsychological Scores in Atrophy Models Improves Diagnostic Classification of Alzheimer's Disease and Mild Cognitive Impairment

Brain atrophy in mild cognitive impairment (MCI) and Alzheimer's disease (AD) are difficult to demarcate to assess the progression of AD. This study presents a statistical framework on the basis of MRI volumes and neuropsychological scores. A feature selection technique using backward stepwise linear regression together with linear discriminant analysis is designed to classify cognitive normal (CN) subjects, early MCI (EMCI), late MCI (LMCI), and AD subjects in an exhaustive two-group classification process. Results show a dominance of the neuropsychological parameters like MMSE and RAVLT. Cortical volumetric measures of the temporal, parietal, and cingulate regions are found to be significant classification factors. Moreover, an asymmetrical distribution of the volumetric measures across hemispheres is seen for CN versus EMCI and EMCI versus AD, showing dominance of the right hemisphere; whereas CN versus LMCI and EMCI versus LMCI show dominance of the left hemisphere. A 2-fold cross-validation showed an average accuracy of 93.9%, 90.8%, and 94.5%, for the CN versus AD, CN versus LMCI, and EMCI versus AD, respectively. The accuracy for groups that are difficult to differentiate like EMCI versus LMCI was 73.6%. With the inclusion of the neuropsychological scores, a significant improvement (24.59%) was obtained over using MRI measures alone.

MEG-based detection and localization of perilesional dysfunction in chronic stroke

Post-stroke impairment is associated not only with structural lesions, but also with dysfunction in surviving perilesional tissue. Previous studies using equivalent current dipole source localization of MEG/EEG signals have demonstrated a preponderance of slow-wave activity localized to perilesional areas. Recent studies have also demonstrated the utility of nonlinear analyses such as multiscale entropy (MSE) for quantifying neuronal dysfunction in a wide range of pathologies. The current study utilized beamformer-based reconstruction of signals in source space to compare spectral and nonlinear measures of electrical activity in perilesional and healthy cortices. Data were collected from chronic stroke patients and healthy controls, both young and elderly. We assessed relative power in the delta (1–4 Hz), theta (4–7 Hz), alpha (8–12 Hz) and beta (15–30 Hz) frequency bands, and also measured the nonlinear complexity of electrical activity using MSE. Perilesional tissue exhibited a general slowing of the power spectrum (increased delta/theta, decreased beta) as well as a reduction in MSE. All measures tested were similarly sensitive to changes in the posterior perilesional regions, but anterior perilesional dysfunction was detected better by MSE and beta power. The findings also suggest that MSE is specifically sensitive to electrophysiological dysfunction in perilesional tissue, while spectral measures were additionally affected by an increase in rolandic beta power with advanced age. Furthermore, perilesional electrophysiological abnormalities in the left hemisphere were correlated with the degree of language task-induced activation in the right hemisphere. Finally, we demonstrate that single subject spectral and nonlinear analyses can identify dysfunctional perilesional regions within individual patients that may be ideal targets for interventions with noninvasive brain stimulation.

•

We assessed the spontaneous MEG activity of perilesional tissue in stroke.

•

We observed perilesional spectral slowing and reduced signal complexity.

•

We demonstrate a method to identify dysfunctional tissue within a single subject.

An SVM-based classifier for estimating the state of various rotating components in agro-industrial machinery with a vibration signal acquired from a single point on the machine chassis

The goal of this article is to assess the feasibility of estimating the state of various rotating components in agro-industrial machinery by employing just one vibration signal acquired from a single point on the machine chassis. To do so, a Support Vector Machine (SVM)-based system is employed. Experimental tests evaluated this system by acquiring vibration data from a single point of an agricultural harvester, while varying several of its working conditions. The whole process included two major steps. Initially, the vibration data were preprocessed through twelve feature extraction algorithms, after which the Exhaustive Search method selected the most suitable features. Secondly, the SVM-based system accuracy was evaluated by using Leave-One-Out cross-validation, with the selected features as the input data. The results of this study provide evidence that (i) accurate estimation of the status of various rotating components in agro-industrial machinery is possible by processing the vibration signal acquired from a single point on the machine structure; (ii) the vibration signal can be acquired with a uniaxial accelerometer, the orientation of which does not significantly affect the classification accuracy; and, (iii) when using an SVM classifier, an 85% mean cross-validation accuracy can be reached, which only requires a maximum of seven features as its input, and no significant improvements are noted between the use of either nonlinear or linear kernels.

Role of EEG as biomarker in the early detection and classification of dementia

The early detection and classification of dementia are important clinical support tasks for medical practitioners in customizing patient treatment programs to better manage the development and progression of these diseases. Efforts are being made to diagnose these neurodegenerative disorders in the early stages. Indeed, early diagnosis helps patients to obtain the maximum treatment benefit before significant mental decline occurs. The use of electroencephalogram as a tool for the detection of changes in brain activities and clinical diagnosis is becoming increasingly popular for its capabilities in quantifying changes in brain degeneration in dementia. This paper reviews the role of electroencephalogram as a biomarker based on signal processing to detect dementia in early stages and classify its severity. The review starts with a discussion of dementia types and cognitive spectrum followed by the presentation of the effective preprocessing denoising to eliminate possible artifacts. It continues with a description of feature extraction by using linear and nonlinear techniques, and it ends with a brief explanation of vast variety of separation techniques to classify EEG signals. This paper also provides an idea from the most popular studies that may help in diagnosing dementia in early stages and classifying through electroencephalogram signal processing and analysis.

Cross-Approximate Entropy parallel computation on GPUs for biomedical signal analysis. Application to MEG recordings

Cross-Approximate Entropy (Cross-ApEn) is a useful measure to quantify the statistical dissimilarity of two time series. In spite of the advantage of Cross-ApEn over its one-dimensional counterpart (Approximate Entropy), only a few studies have applied it to biomedical signals, mainly due to its high computational cost. In this paper, we propose a fast GPU-based implementation of the Cross-ApEn that makes feasible its use over a large amount of multidimensional data. The scheme followed is fully scalable, thus maximizes the use of the GPU despite of the number of neural signals being processed. The approach consists in processing many trials or epochs simultaneously, with independence of its origin. In the case of MEG data, these trials can proceed from different input channels or subjects. The proposed implementation achieves an average speedup greater than 250× against a CPU parallel version running on a processor containing six cores. A dataset of 30 subjects containing 148 MEG channels (49 epochs of 1024 samples per channel) can be analyzed using our development in about 30min. The same processing takes 5 days on six cores and 15 days when running on a single core. The speedup is much larger if compared to a basic sequential Matlab(®) implementation, that would need 58 days per subject. To our knowledge, this is the first contribution of Cross-ApEn measure computation using GPUs. This study demonstrates that this hardware is, to the day, the best option for the signal processing of biomedical data with Cross-ApEn.

Synchrony analysis of spontaneous MEG activity in Alzheimer's disease patients

The aim of this study was to analyze the magnetoencephalography (MEG) background activity in Alzheimer's disease (AD) patients using cross-approximate entropy (Cross-ApEn). Cross-ApEn is a nonlinear measure of asynchrony between time series. Five minutes of recording were acquired with a 148-channel whole-head magnetometer in 12 AD patients and 12 age-matched control subjects. We found significantly higher synchrony between MEG signals from AD patients compared with control subjects. Additionally, we evaluated the ability of Cross-ApEn to discriminate these two groups using receiver operating characteristic (ROC) curves with a leave-one-out cross-validation procedure. We obtained an accuracy of 70.83% (66.67% sensitivity, 75% specificity) and a value of area under the ROC curve of 0.83. These results provide evidence of disconnection problems in AD. Our findings show the usefulness of Cross-ApEn to detect the brain dysfunction in AD.

Characterization of early partial seizure onset: frequency, complexity and entropy

OBJECTIVE

A clear classification of partial seizures onset features is not yet established. Complexity and entropy have been very widely used to describe dynamical systems, but a systematic evaluation of these measures to characterize partial seizures has never been performed.

METHODS

Eighteen different measures including power in frequency bands up to 300 Hz, Gabor atom density (GAD), Higuchi fractal dimension (HFD), Lempel-Ziv complexity, Shannon entropy, sample entropy, and permutation entropy, were selected to test sensitivity to partial seizure onset. Intracranial recordings from 45 patients with mesial temporal, neocortical temporal and neocortical extratemporal seizure foci were included (331 partial seizures).

RESULTS

GAD, Lempel-Ziv complexity, HFD, high frequency activity, and sample entropy were the most reliable measures to assess early seizure onset.

CONCLUSIONS

Increases in complexity and occurrence of high-frequency components appear to be commonly associated with early stages of partial seizure evolution from all regions. The type of measure (frequency-based, complexity or entropy) does not predict the efficiency of the method to detect seizure onset.

SIGNIFICANCE

Differences between measures such as GAD and HFD highlight the multimodal nature of partial seizure onsets. Improved methods for early seizure detection may be achieved from a better understanding of these underlying dynamics.

Addition of EEG improves accuracy of a logistic model that uses neuropsychological and cardiovascular factors to identify dementia and MCI

To investigate whether addition of EEG would improve accuracy of a logistic model that uses neuropsychological assessment and cardiovascular history to identify dementia and mild cognitive impairment (MCI) as a single group, we collected data and constructed logistic models from a sample of 78 normal adults and 33 patients (aged 50-85 years). To determine accuracy, we compared logistic regression results to a geriatrician's diagnosis of MCI or dementia that included Alzheimer's disease, vascular dementia or mixed dementia. We found that the addition of EEG (non-linear complexity) to a logistic model that included both neuropsychological assessment (ADAS-Cog) and cardiovascular history increased overall accuracy from 80% to 92%. The logistic model identified dementia and MCI as a single group comprised of the following subgroups (with accuracies): Alzheimer's disease (92%; 12/13), vascular dementia (73%; 8/11), mixed dementia (100%; 4/4), and mild cognitive impairment (80%; 4/5). Whereas the analysis is limited by small sample sizes and mixing of diverse pathologies, the findings do provide support that the subgroups may share changes in neuropsychological, cardiovascular, and electroencephalographic factors (specifically ADAS-Cog total score, cardiovascular history, and EEG complexity). Taken together, the study results provide support that EEG might complement the clinician's evaluation of dementia and MCI.

MEG analysis in Alzheimer's disease computing approximate entropy for different frequency bands

The goal of this study was to analyze the magnetoencephalogram (MEG) background activity in patients with Alzheimer's disease (AD) using a regularity measure: approximate entropy (ApEn). This measure was computed for a broad band (0.5-40 Hz) as well as typical frequency bands (delta, theta, alpha, beta and gamma). Five minutes of recording were acquired with a 148-channel whole-head magnetometer in 15 patients with probable AD and 15 elderly control subjects. Our results showed that AD patients' MEGs were more regular than controls' recordings at all frequency bands, with the exception of beta. Additionally, there were statistically significant differences (p 〈 0.01, Student's t-test) at the broad and delta bands. Using receiver operating characteristic curves, the highest accuracy (83.33%) was reached at delta band. These results suggest the usefulness of ApEn to gain a better understanding of dynamical processes underlying the MEG recording.

Feature extraction and selection for objective gait analysis and fall risk assessment by accelerometry

BACKGROUND

Falls in the elderly is nowadays a major concern because of their consequences on elderly general health and moral states. Moreover, the aging of the population and the increasing life expectancy make the prediction of falls more and more important. The analysis presented in this article makes a first step in this direction providing a way to analyze gait and classify hospitalized elderly fallers and non-faller. This tool, based on an accelerometer network and signal processing, gives objective informations about the gait and does not need any special gait laboratory as optical analysis do. The tool is also simple to use by a non expert and can therefore be widely used on a large set of patients.

METHOD

A population of 20 hospitalized elderlies was asked to execute several classical clinical tests evaluating their risk of falling. They were also asked if they experienced any fall in the last 12 months. The accelerations of the limbs were recorded during the clinical tests with an accelerometer network distributed on the body. A total of 67 features were extracted from the accelerometric signal recorded during a simple 25 m walking test at comfort speed. A feature selection algorithm was used to select those able to classify subjects at risk and not at risk for several classification algorithms types.

RESULTS

The results showed that several classification algorithms were able to discriminate people from the two groups of interest: fallers and non-fallers hospitalized elderlies. The classification performances of the used algorithms were compared. Moreover a subset of the 67 features was considered to be significantly different between the two groups using a t-test.

CONCLUSIONS

This study gives a method to classify a population of hospitalized elderlies in two groups: at risk of falling or not at risk based on accelerometric data. This is a first step to design a risk of falling assessment system that could be used to provide the right treatment as soon as possible before the fall and its consequences. This tool could also be used to evaluate the risk several times during the revalidation procedure.

Analysis of spontaneous MEG activity in mild cognitive impairment using spectral entropies and disequilibrium measures

The aim of this study was to explore the ability of several spectral entropies and disequilibrium measures to discriminate between spontaneous magnetoencephalographic (MEG) oscillations from 18 mild cognitive impairment (MCI) patients and 24 controls. The Shannon spectral entropy (SSE), Tsallis spectral entropy (TSE), and Rényi spectral entropy (RSE) were calculated from the normalized power spectral density to evaluate the irregularity patterns. In addition, the Euclidean (ED) and Wootters (WD) distances were computed as disequilibrium measures. Results revealed statistically significant lower SSE and TSE(2) values for MCI patients than for controls (p < 0.05) in the right lateral region of the brain. ED also obtained statistically significant lower values for MCI patients than for controls using the (p < 0.05) in the right lateral region of the brain. These findings suggest that MCI is associated with a significant decrease in irregularity of MEG activity. In addition, the highest accuracy of 64.3% was achieved by the SSE. We conclude that measures from information theory can be useful to both characterize abnormal brain dynamics and help in MCI detection.

Entropy and Complexity Analyses in Alzheimer's Disease: An MEG Study

Alzheimer’s disease (AD) is one of the most frequent disorders among elderly population and it is considered the main cause of dementia in western countries. This irreversible brain disorder is characterized by neural loss and the appearance of neurofibrillary tangles and senile plaques. The aim of the present study was the analysis of the magnetoencephalogram (MEG) background activity from AD patients and elderly control subjects. MEG recordings from 36 AD patients and 26 controls were analyzed by means of six entropy and complexity measures: Shannon spectral entropy (SSE), approximate entropy (ApEn), sample entropy (SampEn), Higuchi’s fractal dimension (HFD), Maragos and Sun’s fractal dimension (MSFD), and Lempel-Ziv complexity (LZC). SSE is an irregularity estimator in terms of the flatness of the spectrum, whereas ApEn and SampEn are embbeding entropies that quantify the signal regularity. The complexity measures HFD and MSFD were applied to MEG signals to estimate their fractal dimension. Finally, LZC measures the number of different substrings and the rate of their recurrence along the original time series. Our results show that MEG recordings are less complex and more regular in AD patients than in control subjects. Significant differences between both groups were found in several brain regions using all these methods, with the exception of MSFD (p-value < 0.05, Welch’s t-test with Bonferroni’s correction). Using receiver operating characteristic curves with a leave-one-out cross-validation procedure, the highest accuracy was achieved with SSE: 77.42%. We conclude that entropy and complexity analyses from MEG background activity could be useful to help in AD diagnosis.

MEG connectivity analysis in patients with Alzheimer's disease using cross mutual information and spectral coherence

Alzheimer's disease (AD) is an irreversible brain disorder which represents the most common form of dementia in western countries. An early and accurate diagnosis of AD would enable to develop new strategies for managing the disease; however, nowadays there is no single test that can accurately predict the development of AD. In this sense, only a few studies have focused on the magnetoencephalographic (MEG) AD connectivity patterns. This study compares brain connectivity in terms of linear and nonlinear couplings by means of spectral coherence and cross mutual information function (CMIF), respectively. The variables defined from these functions provide statistically significant differences (p < 0.05) between AD patients and control subjects, especially the variables obtained from CMIF. The results suggest that AD is characterized by both decreases and increases of functional couplings in different frequency bands as well as by an increase in regularity, that is, more evident statistical deterministic relationships in AD patients' MEG connectivity. The significant differences obtained indicate that AD could disturb brain interactions causing abnormal brain connectivity and operation. Furthermore, the combination of coherence and CMIF features to perform a diagnostic test based on logistic regression improved the tests based on individual variables for its robustness.

Study of the MEG background activity in Alzheimer's disease patients with scaling analysis methods

Alzheimer's disease (AD) is one of the most prominent neurodegenerative disorders. The aim of this research work is to study the magnetoencephalogram (MEG) background activity in AD patients using two scaling analysis methods: detrended fluctuation analysis (DFA) and backward detrended moving average (BDMA). Both measures have been designed to quantify correlations in noisy and non-stationary signals. Five minutes of recording were acquired with a 148-channel whole-head magnetometer in 15 patients with probable AD and 15 control subjects. Both DFA and BDMA exhibited two scaling regions with different slopes. Significant differences between both groups were found in the second region of DFA and in the first region of BDMA (p < 0.01, Student's t-test). Using receiver operating characteristic curves, accuracies of 83.33% with DFA and of 80% with BDMA were reached. Our findings show the usefulness of these scaling analysis methods to increase our insight into AD.

Use of magnetoencephalography (MEG) to study functional brain networks in neurodegenerative disorders

The pathophysiological mechanisms underlying clinical symptoms in neurodegenerative disorders such as Parkinson's disease (PD) and Alzheimer's disease (AD) are incompletely understood. Magnetoencephalography (MEG) is a relatively new functional neuroimaging technique, which allows the simultaneous recording of the brain's magnetic activity from large arrays of sensors covering the whole head. MEG studies in PD and AD have identified characteristic patterns of abnormal oscillatory activity in different frequency bands. Furthermore, MEG studies aimed at the characterization of distributed functional networks have demonstrated distinct patterns of abnormal connectivity in demented and non-demented PD, as well as in AD. In PD abnormal oscillatory activity and disturbed connectivity may respond differently to dopaminergic treatment. Further studies in this field could benefit from new technological developments such as ultra low field MRI and from the application of a well-defined theoretical framework such as graph theory to the study of disturbed brain networks.

Blind source separation to enhance spectral and non-linear features of magnetoencephalogram recordings. Application to Alzheimer's disease

This work studied whether a blind source separation (BSS) and component selection procedure could increase the differences between Alzheimer's disease (AD) patients and control subjects' spectral and non-linear features of magnetoencephalogram (MEG) recordings. MEGs were acquired with a 148-channel whole-head magnetometer from 62 subjects (36 AD patients and 26 controls), who were divided randomly into training and test sets. MEGs were decomposed using the algorithm for multiple unknown signals extraction (AMUSE). The extracted AMUSE components were characterised with two spectral--median frequency and spectral entropy (SpecEn)--and two non-linear features: Lempel-Ziv complexity (LZC) and sample entropy (SampEn). One-way analysis of variance with age as a covariate was applied to the training set to decide which components had the most significant differences between groups. Then, partial reconstructions of the MEGs were computed with these significant components. In the test set, the accuracy and area under the ROC curve (AUC) associated with each partial reconstruction of the MEGs were compared with the case where no BSS-preprocessing was applied. This preprocessing increased the AUCs between 0.013 and 0.227, while the accuracy for SpecEn, LZC and SampEn rose between 6.4% and 22.6%, improving the separation between AD patients and control subjects.

An index of signal mode complexity based on orthogonal transformation

Irregular and complex signals are ubiquitous in nature. The principal aim of this paper is to develop an index, quantifying the complexity of such signals, which is based on the distribution of the strengths of its orthogonal oscillatory modes estimated by singular value decomposition. The index is first tested with simulated chaotic and/or stochastic maps and flows. Among neural data analysis, the index is first applied to a cognitive EEG data set recorded from two groups, musicians and non-musicians, during listening to music and resting state. In the gamma band (30-50 Hz), musicians showed robust changes in complexity, consistent over various scalp regions, during listening to music from resting condition, whereas such changes were minimal for non-musicians. Then the index is used to separate healthy participants from epileptic and manic patients based on spontaneous EEG analysis. Finally, it is used to track a tonic-clonic seizure EEG signal, and a conspicuous change was found in the complexity profiles of delta band (1-3.5 Hz) oscillations at the onset of seizure. We conclude that this index would be useful for quantification of a wide range of time series including neural oscillations.

Nonlinear analysis of electroencephalogram and magnetoencephalogram recordings in patients with Alzheimer's disease

The aim of the present study is to show the usefulness of nonlinear methods to analyse the electroencephalogram (EEG) and magnetoencephalogram (MEG) in patients with Alzheimer's disease (AD). The following nonlinear methods have been applied to study the EEG and MEG background activity in AD patients and control subjects: approximate entropy, sample entropy, multiscale entropy, auto-mutual information and Lempel-Ziv complexity. We discuss why these nonlinear methods are appropriate to analyse the EEG and MEG. Furthermore, the performance of all these methods has been compared when applied to the same databases of EEG and MEG recordings. Our results show that EEG and MEG background activities in AD patients are less complex and more regular than in healthy control subjects. In line with previous studies, our work suggests that nonlinear analysis techniques could be useful in AD diagnosis.

Nonlinear forecasting measurement of magnetoencephalogram recordings from Alzheimer's disease patients

The goal of this study was to analyze the magnetoencephalogram (MEG) background activity in patients with Alzheimer's disease (AD) using a nonlinear forecasting measure. It is a nonparametric method to quantify the predictability of time series. Five minutes of recording were acquired with a 148-channel whole-head magnetometer in 15 patients with probable AD and 15 elderly control subjects. Stationary epochs of 5 seconds (848 points, sample frequency of 169.55 Hz) were selected. Our results showed that AD patients' MEGs were more predictable than controls' recordings. Additionally, an accuracy of 76.7% (80.0% sensitivity; 73.3% specificity) was reached using a receiver operating characteristic curve. These preliminary results suggest the usefulness of nonlinear forecasting to gain a better understanding of dynamical processes underlying the MEG recording.