Epoch Selection Research Articles

Unsupervised GAN epoch selection for biomedical data synthesis

Abstract Supervised Neural Networks are used for segmentation in many biological and biomedical applications. To omit the time-consuming and tiring process of manual labeling, unsupervised Generative Adversarial Networks (GANs) can be used to synthesize labeled data. However, the training of GANs requires extensive computation and is often unstable. Due to the lack of established stopping criteria, GANs are usually trained multiple times for a heuristically fixed number of epochs. Early stopping and epoch selection can lead to better synthetic datasets resulting in higher downstream segmentation quality on biological or medical data. This article examines whether the Frechet Inception Distance (FID), the Kernel Inception Distance (KID), or the WeightWatcher tool can be used for early stopping or epoch selection of unsupervised GANs. The experiments show that the last trained GAN epoch is not necessarily the best one to synthesize downstream segmentation data. On complex datasets, FID and KID correlate with the downstream segmentation quality, and both can be used for epoch selection.

Effect of Data Reduction Techniques on Daily Moderate to Vigorous Physical Activity Collected with ActiGraph® in People with COPD.

ActiGraph® is a valid, frequently used, accelerometer to quantify moderate to vigorous physical activities (MVPA) in people with COPD. The impact of ActiGraph processing techniques on this population is unknown. This study aimed to explore the effect of data reduction techniques on MVPA in people with COPD. MVPA/day, through ActiGraph GT3X+, was estimated using: Troiano, Freedson 98 and FreedsonVM3 cutoffs, 15-s and 60-s epochs, and normal and low-frequency extension (LFE) filters. Cutoff, epoch, and filter effects were explored with Aligned Rank Transform-ANOVA. Lin's concordance correlation coefficients and Bland-Altman plots were used to evaluate agreement and bias between different techniques. The analysis included 136 people with COPD (79% male; 68 ± 8 years; FEV1 51 ± 17% predicted). MVPA/day differed according to cutoff, filter, and epoch selection (p-value < 0.001). FreedsonVM3 cutoff, 15-s epochs, and LFE yielded the highest MVPA (45 min/day, 68% of physically active participants). Troiano cutoff, 60-s epochs, and normal filter yielded the lowest MVPA (8 min/day, 20% of physically active participants). Only comparisons between Troiano and Freedson98 cutoffs presented an almost perfect agreement. ActiGraph data reduction techniques affected MVPA/day estimates and their interpretation at the individual and group level. Studies using different processing criteria should not be compared in people with COPD. Future studies with a gold standard are required to ascertain which processing technique produces the most accurate MVPA estimates in COPD. Meanwhile, future trials employing the ActiGraph GT3X+ may consider estimating MVPA based on Freedson VM3 cutofffs, 60-s epochs, and normal filter.

Emotion Recognition Based on the Human's Brain

There are several different algorithms to train models in machine learning and for purpose of enhancing the accuracy of the models for emotion recognition systems by using EEG datasets, we can compare different algorithms and methods. To experience more intuitively the different impacts caused by different algorithms and methods, Emotion classification by applying Convolutional Neural Networks, Sparse Auto-encoder, Deep Neural Network, and ZTW-based epoch selection algorithm, these three models of emotion recognition will be mentioned and compared in this essay to find out a way to improve the accuracy of the emotion recognition models based on EEG datasets, which also include the process and final results of these models.

A Phase-based EEG Epoch Selection Method for Decoding Bi-directional Hand Movement Imagination in Stroke Patients.

Electroencephalogram (EEG) based non-invasive Brain Computer Interface (BCI) system is gaining significant attention as a promising solution for stroke rehabilitation. Accurate selection of informative EEG time segment, that accommodates the specific neural activity patterns associated with the underlying mental task can help to improve the efficacy of the BCI system. In this work, we propose a phase-based EEG epoch selection algorithm to extract the discriminative EEG time segment corresponding to bi-directional hand motor imagery. The imagined center-out hand movement in two directions is decoded using the selected epoch of the EEG, recorded from 16 stroke patients with hemiparesis and specifically hand weakness. Phase Lock Value (PLV) EEG features extracted from the selected EEG epoch is used as discriminative feature for binary classification of imagined hand movement direction using Linear Discriminant Analysis. The use of selected EEG epoch yielded an improvement of 11.5% and 11.7% in the average direction classification accuracy of calibration and feedback session data respectively, compared to the baseline method employing the whole EEG signal. In addition to improvement in decoding accuracy, the epoch selection also yielded an average Information Transfer Rate (ITR) of 39.8±24.6 bits per minute, which is 86% improvement compared to the baseline method.Clinical Relevance- The proposed Motor Imagery (MI)-BCI system may be of clinical relevance as an active rehabilitation tool for stroke-affected patients, to enhance neural plasticity and recovery of centre-out activities of affected hand and forms a strong platform for MI-BCI coupled with exoskeletons or prosthesis rehabilitation.

Reliability of the P3 component in a 3-stimulus concealed information test with known and unknown stimuli

The parietal P3 amplitude of the event-related potential (ERP) has been highlighted to signal stimulus salience in concealed information tests (CITs) with known and unknown stimuli. To extend previous validity investigations in deception research, the reliability of the early parietal P3 amplitude in a 3-stimulus CIT including known and unknown stimuli has been investigated in a sample of N = 68 participants. Variations of Cronbach’s Alpha and split-half reliability (odd–even, 1st vs. 2nd half) of the early parietal P3 amplitude were considered for ERP quantification method (peak-to-peak, mean, baseline-to-peak), epoch selection strategies (chronological vs. random) and number of epochs (10–40) in three stimulus types (probe, target, irrelevant). Excellent reliability coefficients have been summarized for number of epochs, P3 quantifications, and epoch selection strategies. Moreover, the best-practice recommendations highlight mandatory combinations of epoch selection strategy, number of epochs and P3 quantifications to ensure robust and excellent reliabilities of the early parietal P3 amplitude in a 3-stimulus CIT with known (probe, target) and unknown (irrelevant) stimuli. The robustness of mandatory excellent reliability coefficients for the early parietal P3 has been discussed for two data tracks in single-case analyses.

Feasibility of Universal Anomaly Detection without Knowing the Abnormality in Medical Images.

Many anomaly detection approaches, especially deep learning methods, have been recently developed to identify abnormal image morphology by only employing normal images during training. Unfortunately, many prior anomaly detection methods were optimized for a specific "known" abnormality (e.g., brain tumor, bone fraction, cell types). Moreover, even though only the normal images were used in the training process, the abnormal images were often employed during the validation process (e.g., epoch selection, hyper-parameter tuning), which might leak the supposed "unknown" abnormality unintentionally. In this study, we investigated these two essential aspects regarding universal anomaly detection in medical images by (1) comparing various anomaly detection methods across four medical datasets, (2) investigating the inevitable but often neglected issues on how to unbiasedly select the optimal anomaly detection model during the validation phase using only normal images, and (3) proposing a simple decision-level ensemble method to leverage the advantage of different kinds of anomaly detection without knowing the abnormality. The results of our experiments indicate that none of the evaluated methods consistently achieved the best performance across all datasets. Our proposed method enhanced the robustness of performance in general (average AUC 0.956).

Scorepochs: A Computer-Aided Scoring Tool for Resting-State M/EEG Epochs.

M/EEG resting-state analysis often requires the definition of the epoch length and the criteria in order to select which epochs to include in the subsequent steps. However, the effects of epoch selection remain scarcely investigated and the procedure used to (visually) inspect, label, and remove bad epochs is often not documented, thereby hindering the reproducibility of the reported results. In this study, we present Scorepochs, a simple and freely available tool for the automatic scoring of resting-state M/EEG epochs that aims to provide an objective method to aid M/EEG experts during the epoch selection procedure. We tested our approach on a freely available EEG dataset containing recordings from 109 subjects using the BCI2000 64 channel system.

Single-pulse electrical stimulation methodology in freely moving rat

BackgroundCortico-cortical evoked potentials (CCEP) are becoming popular to infer brain connectivity and cortical excitability in implanted refractory epilepsy patients. Our goal was to transfer this methodology to the freely moving rodent. New methodCCEP were recorded on freely moving Sprague-Dawley rats, from cortical and subcortical areas using depth electrodes. Electrical stimulation was applied using 1 ms biphasic current pulse, cathodic first, at a frequency of 0.5 Hz, with intensities ranging from 0.2 to 0.8 mA. Data were then processed in a similar fashion to human clinical studies, which included epoch selection, artefact correction and smart averaging. ResultsFor a large range of tested intensities, we recorded CCEPs with very good signal to noise ratio and reproducibility between animals, without any behavioral modification. The CCEP were composed of different components according to recorded and stimulated sites, similarly to human recordings. Comparison with existing methodsWe minimally adapted a clinically-motivated methodology to a freely moving rodent model to achieve high translational relevance of future preclinical studies. ConclusionsOur results indicate that the CCEP methodology can be applied to freely moving rodents and transferred to preclinical research. This will be of interest to address various neuroscientific questions, in physiological and pathological conditions.

Improving the repeatability of Motor Unit Number Index (MUNIX) by introducing additional epochs at low contraction levels.

To evaluate the repeatability of (Motor Unit Number Index) MUNIX under repeatability conditions, specify the origin of variations and provide strategies for quality control. MUNIX calculations were performed on the bicep brachii muscles of eight healthy subjects. Negative effect of suboptimal electrode positions on MUNIX accuracy was eliminated by employing the high-density surface electromyography technique. MUNIX procedures that utilized a variety of surface interferential pattern (SIP) epoch recruitment strategies (including the original MUNIX procedure, two proposed improvement strategies and their combinations) were described. For each MUNIX procedure, ten thousands of different SIP pools were constructed by randomly recruiting necessary SIP epochs from a large SIP epoch pool (3 datasets, 9 independent electromyography recordings at different contraction levels per dataset and 10 SIP epochs per recording) and implemented for MUNIX calculation. The repeatability of each MUNIX procedure was assessed by summarizing the resulting MUNIX distribution and compared to investigate the effect of SIP epoch selection strategy on repeatability performance. SIP epochs selected at lower contraction levels have a stronger influence on the repeatability of MUNIX than those selected at higher contraction levels. MUNIX under repeatability conditions follows a normal distribution and the standard deviation can be significantly reduced by introducing more epochs near the MUNIX definition line. The MUNIX technique shows an inherent variation attributable to SIP epochs at low contraction levels. It is recommended that more epochs should be sampled at these low contraction levels to improve the repeatability. The present study thoroughly documented the inherent variation of MUNIX and the causes, and offered practical solutions to improve the repeatability of MUNIX.

EEG epochs with less alpha rhythm improve discrimination of mild Alzheimer's

Background and objectiveEyes-closed-awake electroencephalogram (EEG) is a useful tool in the diagnosis of Alzheimer's. However, there is eyes-closed-awake EEG with dominant or rare alpha rhythm. In this paper, we show that random selection of EEG epochs disregarding the alpha rhythm will lead to bias concerning EEG-based Alzheimer's Disease diagnosis. MethodsWe compared EEG epochs with more than 30% and with less than 30% alpha rhythm of mild Alzheimer's Disease patients and healthy elderly. We classified epochs as dominant alpha scenario and rare alpha scenario according to alpha rhythm (8–13 Hz) percentage in O1, O2 and Oz channels. Accordingly, we divided the probands into four groups: 17 dominant alpha scenario controls, 15 mild Alzheimer's patients with dominant alpha scenario epochs, 12 rare alpha scenario healthy elderly and 15 mild Alzheimer's Disease patients with rare alpha scenario epochs. We looked for group differences using one-way ANOVA tests followed by post-hoc multiple comparisons (p < 0.05) over normalized energy values (%) on the other four well-known frequency bands (delta, theta, beta and gamma) using two different electrode configurations (parieto-occipital and central). ResultsAfter carrying out post-hoc multiple comparisons, for both electrode configurations we found significant differences between mild Alzheimer's patients and healthy elderly on beta- and theta-energy (%) only for the rare alpha scenario. No differences were found for the dominant alpha scenario in any of the five frequency bands. ConclusionsThis is the first study of Alzheimer's awake-EEG reporting the influence of alpha rhythm on epoch selection, where our results revealed that, contrarily to what was most likely expected, less synchronized EEG epochs (rare alpha scenario) better discriminated mild Alzheimer's than those presenting abundant alpha (dominant alpha scenario). In addition, we find out that epoch selection is a very sensitive issue in qEEG research. Consequently, for Alzheimer's studies dealing with resting state EEG, we propose that epoch selection strategies should always be cautiously designed and thoroughly explained.

EEG epoch selection: lack of alpha rhythm improves discrimination of Alzheimer's disease.

In this work we propose a detailed EEG epoch selection method and compare epochs with rare and abundant alpha rhythm (AR) of patients with Alzheimer's disease (AD) and normal controls. Epochs were classified as Dominant Alpha Scenario (DAS) and Rare Alpha Scenario (RAS) according to the AR percentage (energy within the 8-13 Hz bandwidth) in O1, O2 and Oz electrodes. Participants were divided into four groups: 17 DAS controls (N1), 15 DAS mild-AD patients (AD1), 12 RAS controls (N2) and 15 RAS mild-AD patients (AD2). We found out that scenario factor (DAS vs. RAS, two-way ANOVA) is significant over a great amount of electrode-bandwidth situations. Furthermore, one-way ANOVA showed significant differences between RAS AD and RAS controls in much more situations as compared to DAS. This is the first study using AD awake EEG reporting the decisive influence of alpha rhythm on epoch selection, where our results revealed that, contrary to what was initially expected, EEG epochs with poor alpha (RAS) discriminate mild AD much better than those presenting richer alpha content (DAS).

Exploration of Prominent Frequency Wave in EEG Signals from Brain Sensors Network

We investigated the signals regularity of electroencephalography (EEG) channels separately and determined the energy of selected frequency waves, such as, δ, θ, α, β, and γ. The goal of this research is to identify the prominent frequency band from selected frequencies. We recorded the EEG signal data of 30 controlled subjects with 18 EEG channels. These subjects are all males with an average age of 24 years. Emotional stimuli related to different emotions were presented to each of selected candidates. EEG data were extracted and further processed for artifact removal, filtering, epoch selection and averaging of the signals. We designed and tested our method for exploring the frequency waves of all EEG channels. We also employed the Hjorth parameters to measure the signal regularity in time and frequency domain. The detailed physiological response of human subjects is also presented in this paper. Our results showed that the energy level of delta wave is mostly high in all cases.

DYNAMICS OF COGNITIVE FUNCTIONING IN HEALTHY CONTROLS AND PATIENTS WITH SCHIZOPHRENIA

We have developed a toolbox and graphic user interface, EEGLAB, running under the crossplatform MATLAB environment (The Mathworks, Inc.) for processing collections of single-trial and/or averaged EEG data of any number of channels. Available functions include EEG data, channel and event information importing, data visualization (scrolling, scalp map and dipole model plotting, plus multi-trial ERP-image plots), preprocessing (including artifact rejection, filtering, epoch selection, and averaging), independent component analysis (ICA) and time/frequency decompositions including channel and component cross-coherence supported by bootstrap statistical methods based on data resampling. EEGLAB functions are organized into three layers. Top-layer functions allow users to interact with the data through the graphic interface without needing to use MATLAB syntax. Menu options allow users to tune the behavior of EEGLAB to available memory. Middle-layer functions allow users to customize data processing using command history and interactive ‘pop’ functions. Experienced MATLAB users can use EEGLAB data structures and stand-alone signal processing functions to write custom and/or batch analysis scripts. Extensive function help and tutorial information are included. A ‘plug-in’ facility allows easy incorporation of new EEG modules into the main menu. EEGLAB is freely available (http://www.sccn.ucsd.edu/eeglab/) under the GNU public license for noncommercial use and open source development, together with sample data, user tutorial and extensive documentation.

EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis

We have developed a toolbox and graphic user interface, EEGLAB, running under the crossplatform MATLAB environment (The Mathworks, Inc.) for processing collections of single-trial and/or averaged EEG data of any number of channels. Available functions include EEG data, channel and event information importing, data visualization (scrolling, scalp map and dipole model plotting, plus multi-trial ERP-image plots), preprocessing (including artifact rejection, filtering, epoch selection, and averaging), independent component analysis (ICA) and time/frequency decompositions including channel and component cross-coherence supported by bootstrap statistical methods based on data resampling. EEGLAB functions are organized into three layers. Top-layer functions allow users to interact with the data through the graphic interface without needing to use MATLAB syntax. Menu options allow users to tune the behavior of EEGLAB to available memory. Middle-layer functions allow users to customize data processing using command history and interactive ‘pop’ functions. Experienced MATLAB users can use EEGLAB data structures and stand-alone signal processing functions to write custom and/or batch analysis scripts. Extensive function help and tutorial information are included. A ‘plug-in’ facility allows easy incorporation of new EEG modules into the main menu. EEGLAB is freely available ( http://www.sccn.ucsd.edu/eeglab/) under the GNU public license for noncommercial use and open source development, together with sample data, user tutorial and extensive documentation.

Single Channel QEEG Amplitudes in a Bright, Normal Young Adult Sample

This study provides normative data using the Autogenics A620 . Mean amplitudes for theta, alpha, SMR, and beta, and the theta/beta ratio, are reported for eyes-open and eyes-closed conditions. Inter-raterreliabilitys of epoch selection are reported.

Power spectra accuracy improvement by optimal signal epoch selection: an heuristic approach

The problem of epoch selection in frequency analysis of biomedical signals often involves tradeoffs between the physiological system studied stationarity, the spectral resolution required, the need for dynamic investigations implying shorter epochs, etc. The accuracy of the spectra strongly depends on the epoch length. We propose a method for optimal epoch selection introducing the notion of ‘quality of power spectra’. Two criteria for quantitative estimation of spectral quality are defined, considering a quality to be high when more power is concentrated in narrow frequency bands. A synthesized signal is used for describing the procedure developed and evaluating the accuracy. Noise sensitivity is assessed by adding intensive noise to the signal. Hundreds of synthesized signals are used for validation of the method and some initial experiments with heart rate variation data are presented.

Power spectra accuracy improvement by a method for leakage effect reduction

A method for leakage effect reduction by a modification of the classical spectra obtained Discrete Fourier Transform is proposed. The properties of the new method are investigated analytically and validated through synthesized signals, processed by the developed complete computational procedure. This combines the proposed method for leakage reduction and our previous approach for optimal epoch selection, enabling more accurate spectral analysis of relatively short epochs of various biomedical and other signals. An application of heart rate variability to power spectra is presented.