Concept-drifts adaptation for machine learning EEG epilepsy seizure prediction.

Seizure prediction remains a challenge, with approximately 30% of patients unresponsive to conventional treatments. Addressing this issue is crucial for improving patients' quality of life, as timely intervention can mitigate the impact of seizures. In this research field, it is critical to identify the preictal interval, the transition from regular brain activity to a seizure. While previous studies have explored various Electroencephalogram (EEG) based methodologies for prediction, few have been clinically applicable. Recent studies have underlined the dynamic nature of EEG data, characterised by data changes with time, known as concept drifts, highlighting the need for automated methods to detect and adapt to these changes. In this study, we investigate the effectiveness of automatic concept drift adaptation methods in seizure prediction. Three patient-specific seizure prediction approaches with a 10-minute prediction horizon are compared: a seizure prediction algorithm incorporating a window adjustment method by optimising performance with Support Vector Machines (Backwards-Landmark Window), a seizure prediction algorithm incorporating a data-batch (seizures) selection method using a logistic regression (Seizure-batch Regression), and a seizure prediction algorithm with a dynamic integration of classifiers (Dynamic Weighted Ensemble). These methods incorporate a retraining process after each seizure and use a combination of univariate linear features and SVM classifiers. The Firing Power was used as a post-processing technique to generate alarms before seizures. These methodologies were compared with a control approach based on the typical machine learning pipeline, considering a group of 37 patients with Temporal Lobe Epilepsy from the EPILEPSIAE database. The best-performing approach (Backwards-Landmark Window) achieved results of 0.75 ± 0.33 for sensitivity and 1.03 ± 1.00 for false positive rate per hour. This new strategy performed above chance for 89% of patients with the surrogate predictor, whereas the control approach only validated 46%.

Nisin variants: What makes them different and unique?

Nisin A is a lantibiotic bacteriocin typically produced by strains of Lactococcus lactis. This bacteriocin has been approved as a natural food preservative since the late 1980 s and shows antimicrobial activity against a range of food-borne spoilage and pathogenic microorganisms. The therapeutic potential of nisin A has also been explored increasingly both in human and veterinary medicine. Nisin has been shown to be effective in treating bovine mastitis, dental caries, cancer, and skin infections. Recently, it was demonstrated that nisin has an affinity for the same receptor used by SARS-CoV-2 to enter human cells and was proposed as a blocker of the viral infection. Several nisin variants produced by distinct bacterial strains or modified by bioengineering have been described since the discovery of nisin A. These variants present modifications in the peptide structure, biosynthesis, mode of action, and spectrum of activity. Given the importance of nisin for industrial and therapeutic applications, the objective of this study was to describe the characteristics of the nisin variants, highlighting the main differences between these molecules and their potential applications. This review will be useful to researchers interested in studying the specifics of nisin A and its variants.

Comparison between epileptic seizure prediction and forecasting based on machine learning.

Epilepsy affects around 1% of the population worldwide. Anti-epileptic drugs are an excellent option for controlling seizure occurrence but do not work for around one-third of patients. Warning devices employing seizure prediction or forecasting algorithms could bring patients new-found comfort and quality of life. These algorithms would attempt to detect a seizure's preictal period, a transitional moment between regular brain activity and the seizure, and relay this information to the user. Over the years, many seizure prediction studies using Electroencephalogram-based methodologies have been developed, triggering an alarm when detecting the preictal period. Recent studies have suggested a shift in view from prediction to forecasting. Seizure forecasting takes a probabilistic approach to the problem in question instead of the crisp approach of seizure prediction. In this field of study, the triggered alarm to symbolize the detection of a preictal period is substituted by a constant risk assessment analysis. The present work aims to explore methodologies capable of seizure forecasting and establish a comparison with seizure prediction results. Using 40 patients from the EPILEPSIAE database, we developed several patient-specific prediction and forecasting algorithms with different classifiers (a Logistic Regression, a 15 Support Vector Machines ensemble, and a 15 Shallow Neural Networks ensemble). Results show an increase of the seizure sensitivity in forecasting relative to prediction of up to 146% and in the number of patients that displayed an improvement over chance of up to 300%. These results suggest that a seizure forecasting methodology may be more suitable for seizure warning devices than a seizure prediction one.

EEG as a potential ground truth for the assessment of cognitive state in software development activities: A multimodal imaging study.

Cognitive human error and recent cognitive taxonomy on human error causes of software defects support the intuitive idea that, for instance, mental overload, attention slips, and working memory overload are important human causes for software bugs. In this paper, we approach the EEG as a reliable surrogate to MRI-based reference of the programmer's cognitive state to be used in situations where heavy imaging techniques are infeasible. The idea is to use EEG biomarkers to validate other less intrusive physiological measures, that can be easily recorded by wearable devices and useful in the assessment of the developer's cognitive state during software development tasks. Herein, our EEG study, with the support of fMRI, presents an extensive and systematic analysis by inspecting metrics and extracting relevant information about the most robust features, best EEG channels and the best hemodynamic time delay in the context of software development tasks. From the EEG-fMRI similarity analysis performed, we found significant correlations between a subset of EEG features and the Insula region of the brain, which has been reported as a region highly related to high cognitive tasks, such as software development tasks. We concluded that despite a clear inter-subject variability of the best EEG features and hemodynamic time delay used, the most robust and predominant EEG features, across all the subjects, are related to the Hjorth parameter Activity and Total Power features, from the EEG channels F4, FC4 and C4, and considering in most of the cases a hemodynamic time delay of 4 seconds used on the hemodynamic response function. These findings should be taken into account in future EEG-fMRI studies in the context of software debugging.

Classifier combination supported by the sleep-wake cycle improves EEG seizure prediction performance.

OBJECTIVE: Seizure prediction is a promising solution to improve the quality of life for drug-resistant patients, which concerns nearly 30% of patients with epilepsy. The present study aimed to ascertain the impact of incorporating sleep-wake information in seizure prediction. METHODS: We developed five patient-specific prediction approaches that use vigilance state information differently: i) using it as an input feature, ii) building a pool of two classifiers, each with different weights to sleep/wake training samples, iii) building a pool of two classifiers, each with only sleep/wake samples, iv) changing the alarm-threshold concerning each sleep/wake state, and v) adjusting the alarm-threshold after a sleep-wake transition. We compared these approaches with a control method that did not integrate sleep-wake information. Our models were tested with data (43 seizures and 482 hours) acquired during presurgical monitoring of 17 patients from the EPILEPSIAE database. As EPILEPSIAE does not contain vigilance state annotations, we developed a sleep-wake classifier using 33 patients diagnosed with nocturnal frontal lobe epilepsy from the CAP Sleep database. RESULTS: Although different patients may require different strategies, our best approach, the pool of weighted predictors, obtained 65% of patients performing above chance level with a surrogate analysis (against 41% in the control method). CONCLUSION: The inclusion of vigilance state information improves seizure prediction. Higher results and testing with longterm recordings from daily-life conditions are necessary to ensure clinical acceptance. SIGNIFICANCE: As automated sleep-wake detection is possible, it would be feasible to incorporate these algorithms into future devices for seizure prediction.

EEG epilepsy seizure prediction: the post-processing stage as a chronology.

Almost one-third of epileptic patients fail to achieve seizure control through anti-epileptic drug administration. In the scarcity of completely controlling a patient's epilepsy, seizure prediction plays a significant role in clinical management and providing new therapeutic options such as warning or intervention devices. Seizure prediction algorithms aim to identify the preictal period that Electroencephalogram (EEG) signals can capture. However, this period is associated with substantial heterogeneity, varying among patients or even between seizures from the same patient. The present work proposes a patient-specific seizure prediction algorithm using post-processing techniques to explore the existence of a set of chronological events of brain activity that precedes epileptic seizures. The study was conducted with 37 patients with Temporal Lobe Epilepsy (TLE) from the EPILEPSIAE database. The designed methodology combines univariate linear features with a classifier based on Support Vector Machines (SVM) and two post-processing techniques to handle pre-seizure temporality in an easily explainable way, employing knowledge from network theory. In the Chronological Firing Power approach, we considered the preictal as a sequence of three brain activity events separated in time. In the Cumulative Firing Power approach, we assumed the preictal period as a sequence of three overlapping events. These methodologies were compared with a control approach based on the typical machine learning pipeline. We considered a Seizure Prediction horizon (SPH) of 5 mins and analyzed several values for the Seizure Occurrence Period (SOP) duration, between 10 and 55 mins. Our results showed that the Cumulative Firing Power approach may improve the seizure prediction performance. This new strategy performed above chance for 62% of patients, whereas the control approach only validated 49% of its models.

Computational methods applied to syphilis: where are we, and where are we going?

Syphilis is an infectious disease that can be diagnosed and treated cheaply. Despite being a curable condition, the syphilis rate is increasing worldwide. In this sense, computational methods can analyze data and assist managers in formulating new public policies for preventing and controlling sexually transmitted infections (STIs). Computational techniques can integrate knowledge from experiences and, through an inference mechanism, apply conditions to a database that seeks to explain data behavior. This systematic review analyzed studies that use computational methods to establish or improve syphilis-related aspects. Our review shows the usefulness of computational tools to promote the overall understanding of syphilis, a global problem, to guide public policy and practice, to target better public health interventions such as surveillance and prevention, health service delivery, and the optimal use of diagnostic tools. The review was conducted according to PRISMA 2020 Statement and used several quality criteria to include studies. The publications chosen to compose this review were gathered from Science Direct, Web of Science, Springer, Scopus, ACM Digital Library, and PubMed databases. Then, studies published between 2015 and 2022 were selected. The review identified 1,991 studies. After applying inclusion, exclusion, and study quality assessment criteria, 26 primary studies were included in the final analysis. The results show different computational approaches, including countless Machine Learning algorithmic models, and three sub-areas of application in the context of syphilis: surveillance (61.54%), diagnosis (34.62%), and health policy evaluation (3.85%). These computational approaches are promising and capable of being tools to support syphilis control and surveillance actions.

Digital Alternative Communication for Individuals with Amyotrophic Lateral Sclerosis: What We Have.

Amyotrophic Lateral Sclerosis is a disease that compromises the motor system and the functional abilities of the person in an irreversible way, causing the progressive loss of the ability to communicate. Tools based on Augmentative and Alternative Communication are essential for promoting autonomy and improving communication, life quality, and survival. This Systematic Literature Review aimed to provide evidence on eye-image-based Human-Computer Interaction approaches for the Augmentative and Alternative Communication of people with Amyotrophic Lateral Sclerosis. The Systematic Literature Review was conducted and guided following a protocol consisting of search questions, inclusion and exclusion criteria, and quality assessment, to select primary studies published between 2010 and 2021 in six repositories: Science Direct, Web of Science, Springer, IEEE Xplore, ACM Digital Library, and PubMed. After the screening, 25 primary studies were evaluated. These studies showcased four low-cost, non-invasive Human-Computer Interaction strategies employed for Augmentative and Alternative Communication in people with Amyotrophic Lateral Sclerosis. The strategies included Eye-Gaze, which featured in 36% of the studies; Eye-Blink and Eye-Tracking, each accounting for 28% of the approaches; and the Hybrid strategy, employed in 8% of the studies. For these approaches, several computational techniques were identified. For a better understanding, a workflow containing the development phases and the respective methods used by each strategy was generated. The results indicate the possibility and feasibility of developing Human-Computer Interaction resources based on eye images for Augmentative and Alternative Communication in a control group. The absence of experimental testing in people with Amyotrophic Lateral Sclerosis reiterates the challenges related to the scalability, efficiency, and usability of these technologies for people with the disease. Although challenges still exist, the findings represent important advances in the fields of health sciences and technology, promoting a promising future with possibilities for better life quality.

Osteoporosis screening using machine learning and electromagnetic waves.

Osteoporosis is a disease characterized by impairment of bone microarchitecture that causes high socioeconomic impacts in the world because of fractures and hospitalizations. Although dual-energy X-ray absorptiometry (DXA) is the gold standard for diagnosing the disease, access to DXA in developing countries is still limited due to its high cost, being present only in specialized hospitals. In this paper, we analyze the performance of Osseus, a low-cost portable device based on electromagnetic waves that measures the attenuation of the signal that crosses the medial phalanx of a patient's middle finger and was developed for osteoporosis screening. The analysis is carried out by predicting changes in bone mineral density using Osseus measurements and additional common risk factors used as input features to a set of supervised classification models, while the results from DXA are taken as target (real) values during the training of the machine learning algorithms. The dataset consisted of 505 patients who underwent osteoporosis screening with both devices (DXA and Osseus), of whom 21.8% were healthy and 78.2% had low bone mineral density or osteoporosis. A cross-validation with k-fold = 5 was considered in model training, while 20% of the whole dataset was used for testing. The obtained performance of the best model (Random Forest) presented a sensitivity of 0.853, a specificity of 0.879, and an F1 of 0.859. Since the Random Forest (RF) algorithm allows some interpretability of its results (through the impurity check), we were able to identify the most important variables in the classification of osteoporosis. The results showed that the most important variables were age, body mass index, and the signal attenuation provided by Osseus. The RF model, when used together with Osseus measurements, is effective in screening patients and facilitates the early diagnosis of osteoporosis. The main advantages of such early screening are the reduction of costs associated with exams, surgeries, treatments, and hospitalizations, as well as improved quality of life for patients.

The goal of explaining black boxes in EEG seizure prediction is not to explain models' decisions.

Many state-of-the-art methods for seizure prediction, using the electroencephalogram, are based on machine learning models that are black boxes, weakening the trust of clinicians in them for high-risk decisions. Seizure prediction concerns a multidimensional time-series problem that performs continuous sliding window analysis and classification. In this work, we make a critical review of which explanations increase trust in models' decisions for predicting seizures. We developed three machine learning methodologies to explore their explainability potential. These contain different levels of model transparency: a logistic regression, an ensemble of 15 support vector machines, and an ensemble of three convolutional neural networks. For each methodology, we evaluated quasi-prospectively the performance in 40 patients (testing data comprised 2055 hours and 104 seizures). We selected patients with good and poor performance to explain the models' decisions. Then, with grounded theory, we evaluated how these explanations helped specialists (data scientists and clinicians working in epilepsy) to understand the obtained model dynamics. We obtained four lessons for better communication between data scientists and clinicians. We found that the goal of explainability is not to explain the system's decisions but to improve the system itself. Model transparency is not the most significant factor in explaining a model decision for seizure prediction. Even when using intuitive and state-of-the-art features, it is hard to understand brain dynamics and their relationship with the developed models. We achieve an increase in understanding by developing, in parallel, several systems that explicitly deal with signal dynamics changes that help develop a complete problem formulation.

Removing artefacts and periodically retraining improve performance of neural network-based seizure prediction models.

The development of seizure prediction models is often based on long-term scalp electroencephalograms (EEGs) since they capture brain electrical activity, are non-invasive, and come at a relatively low-cost. However, they suffer from major shortcomings. First, long-term EEG is usually highly contaminated with artefacts. Second, changes in the EEG signal over long intervals, known as concept drift, are often neglected. We evaluate the influence of these problems on deep neural networks using EEG time series and on shallow neural networks using widely-used EEG features. Our patient-specific prediction models were tested in 1577 hours of continuous EEG, containing 91 seizures from 41 patients with temporal lobe epilepsy who were undergoing pre-surgical monitoring. Our results showed that cleaning EEG data, using a previously developed artefact removal method based on deep convolutional neural networks, improved prediction performance. We also found that retraining the models over time reduced false predictions. Furthermore, the results show that although deep neural networks processing EEG time series are less susceptible to false alarms, they may need more data to surpass feature-based methods. These findings highlight the importance of robust data denoising and periodic adaptation of seizure prediction models.

Unsupervised EEG preictal interval identification in patients with drug-resistant epilepsy.

Typical seizure prediction models aim at discriminating interictal brain activity from pre-seizure electrographic patterns. Given the lack of a preictal clinical definition, a fixed interval is widely used to develop these models. Recent studies reporting preictal interval selection among a range of fixed intervals show inter- and intra-patient preictal interval variability, possibly reflecting the heterogeneity of the seizure generation process. Obtaining accurate labels of the preictal interval can be used to train supervised prediction models and, hence, avoid setting a fixed preictal interval for all seizures within the same patient. Unsupervised learning methods hold great promise for exploring preictal alterations on a seizure-specific scale. Multivariate and univariate linear and nonlinear features were extracted from scalp electroencephalography (EEG) signals collected from 41 patients with drug-resistant epilepsy undergoing presurgical monitoring. Nonlinear dimensionality reduction was performed for each group of features and each of the 226 seizures. We applied different clustering methods in searching for preictal clusters located until 2 h before the seizure onset. We identified preictal patterns in 90% of patients and 51% of the visually inspected seizures. The preictal clusters manifested a seizure-specific profile with varying duration (22.9 ± 21.0 min) and starting time before seizure onset (47.6 ± 27.3 min). Searching for preictal patterns on the EEG trace using unsupervised methods showed that it is possible to identify seizure-specific preictal signatures for some patients and some seizures within the same patient.

Development of a Cyclic Voltammetry-Based Method for the Detection of Antigens and Antibodies as a Novel Strategy for Syphilis Diagnosis.

The improvement of laboratory diagnosis is a critical step for the reduction of syphilis cases around the world. In this paper, we present the development of an impedance-based method for detecting T. pallidum antigens and antibodies as an auxiliary tool for syphilis laboratory diagnosis. We evaluate the voltammetric signal obtained after incubation in carbon or gold nanoparticle-modified carbon electrodes in the presence or absence of Poly-L-Lysine. Our results indicate that the signal obtained from the electrodes was sufficient to distinguish between infected and non-infected samples immediately (T0') or 15 min (T15') after incubation, indicating its potential use as a point-of-care method as a screening strategy.

Similarity Analysis in Understanding Online News in Response to Public Health Crisis.

BACKGROUND: The "Syphilis No!" campaign the Brazilian Ministry of Health (MoH) launched between November 2018 and March 2019, brought forward the concept "Test, Treat and Cure" to remind the population of the importance of syphilis prevention. In this context, this study aims to analyze the similarity of syphilis online news to comprehend how public health communication interventions influence media coverage of the syphilis issue. METHODS: This paper presented a computational approach to assess the effectiveness of communication actions on a public health problem. Data were collected between January 2015 and December 2019 and processed using the Hermes ecosystem, which utilizes text mining and machine learning algorithms to cluster similar content. RESULTS: Hermes identified 1049 google-indexed web pages containing the term 'syphilis' in Brazil. Of these, 619 were categorized as news stories. In total, 157 were grouped into clusters of at least two similar news items and a single cluster with 462 news classified as "single" for not featuring similar news items. From these, 19 clusters were identified in the pre-campaign period, 23 during the campaign, and 115 in the post-campaign. CONCLUSIONS: The findings presented in this study show that the volume of syphilis-related news reports has increased in recent years and gained popularity after the SNP started, having been boosted during the campaign and escalating even after its completion.

How Reliable Are Ultra-Short-Term HRV Measurements during Cognitively Demanding Tasks?

Ultra-short-term HRV features assess minor autonomous nervous system variations such as variations resulting from cognitive stress peaks during demanding tasks. Several studies compare ultra-short-term and short-term HRV measurements to investigate their reliability. However, existing experiments are conducted in low cognitively demanding environments. In this paper, we propose to evaluate these measurements' reliability under cognitively demanding tasks using a near real-life setting. For this purpose, we selected 31 HRV features, extracted from data collected from 21 programmers performing code comprehension, and compared them across 18 different time frames, ranging from 3 min to 10 s. Statistical significance and correlation tests were performed between the features extracted using the larger window (3 min) and the same features extracted with the other 17 time frames. We paired these analyses with Bland-Altman plots to inspect how the extraction window size affects the HRV features. The main results show 13 features that presented at least 50% correlation when using 60-second windows. The HF and mNN features achieved around 50% correlation using a 30-second window. The 30-second window was the smallest time frame considered to have reliable measurements. Furthermore, the mNN feature proved to be quite robust to the shortening of the time resolution.

Automatic segmentation and classification of mice ultrasonic vocalizations.

This paper addresses the development of a system for classifying mouse ultrasonic vocalizations (USVs) present in audio recordings. The automatic labeling process for USVs is usually divided into two main steps: USV segmentation followed by the matching classification. Three main contributions can be highlighted: (i) a new segmentation algorithm, (ii) a new set of features, and (iii) the discrimination of a higher number of classes when compared to similar studies. The developed segmentation algorithm is based on spectral entropy analysis. This novel segmentation approach can detect USVs with 94% and 74% recall and precision, respectively. When compared to other methods/software, our segmentation algorithm achieves a higher recall. Regarding the classification phase, besides the traditional features from time, frequency, and time-frequency domains, a new set of contour-based features were extracted and used as inputs of shallow machine learning classification models. The contour-based features were obtained from the time-frequency ridge representation of USVs. The classification methods can differentiate among ten different syllable types with 81.1% accuracy and 80.5% weighted F1-score. The algorithms were developed and evaluated based on a large dataset, acquired on diverse social interaction conditions between the animals, to stimulate a varied vocal repertoire.

EPIC: Annotated epileptic EEG independent components for artifact reduction.

Scalp electroencephalogram is a non-invasive multi-channel biosignal that records the brain's electrical activity. It is highly susceptible to noise that might overshadow important data. Independent component analysis is one of the most used artifact removal methods. Independent component analysis separates data into different components, although it can not automatically reject the noisy ones. Therefore, experts are needed to decide which components must be removed before reconstructing the data. To automate this method, researchers have developed classifiers to identify noisy components. However, to build these classifiers, they need annotated data. Manually classifying independent components is a time-consuming task. Furthermore, few labelled data are publicly available. This paper presents a source of annotated electroencephalogram independent components acquired from patients with epilepsy (EPIC Dataset). This dataset contains 77,426 independent components obtained from approximately 613 hours of electroencephalogram, visually inspected by two experts, which was already successfully utilised to develop independent component classifiers.

On the clinical acceptance of black-box systems for EEG seizure prediction.

Seizure prediction may be the solution for epileptic patients whose drugs and surgery do not control seizures. Despite 46 years of research, few devices/systems underwent clinical trials and/or are commercialized, where the most recent state-of-the-art approaches, as neural networks models, are not used to their full potential. The latter demonstrates the existence of social barriers to new methodologies due to data bias, patient safety, and legislation compliance. In the form of literature review, we performed a qualitative study to analyze the seizure prediction ecosystem to find these social barriers. With the Grounded Theory, we draw hypotheses from data, while with the Actor-Network Theory we considered that technology shapes social configurations and interests, being fundamental in healthcare. We obtained a social network that describes the ecosystem and propose research guidelines aiming at clinical acceptance. Our most relevant conclusion is the need for model explainability, but not necessarily intrinsically interpretable models, for the case of seizure prediction. Accordingly, we argue that it is possible to develop robust prediction models, including black-box systems to some extent, while avoiding data bias, ensuring patient safety, and still complying with legislation, if they can deliver human- comprehensible explanations. Due to skepticism and patient safety reasons, many authors advocate the use of transparent models which may limit their performance and potential. Our study highlights a possible path, by using model explainability, on how to overcome these barriers while allowing the use of more computationally robust models.

Interpretable EEG seizure prediction using a multiobjective evolutionary algorithm.

Seizure prediction might be the solution to tackle the apparent unpredictability of seizures in patients with drug-resistant epilepsy, which comprise about a third of all patients with epilepsy. Designing seizure prediction models involves defining the pre-ictal period, a transition stage between inter-ictal brain activity and the seizure discharge. This period is typically a fixed interval, with some recent studies reporting the evaluation of different patient-specific pre-ictal intervals. Recently, researchers have aimed to determine the pre-ictal period, a transition stage between regular brain activity and a seizure. Authors have been using deep learning models given the ability of such models to automatically perform pre-processing, feature extraction, classification, and handling temporal and spatial dependencies. As these approaches create black-box models, clinicians may not have sufficient trust to use them in high-stake decisions. By considering these problems, we developed an evolutionary seizure prediction model that identifies the best set of features while automatically searching for the pre-ictal period and accounting for patient comfort. This methodology provides patient-specific interpretable insights, which might contribute to a better understanding of seizure generation processes and explain the algorithm's decisions. We tested our methodology on 238 seizures and 3687 h of continuous data, recorded on scalp recordings from 93 patients with several types of focal and generalised epilepsies. We compared the results with a seizure surrogate predictor and obtained a performance above chance for 32% patients. We also compared our results with a control method based on the standard machine learning pipeline (pre-processing, feature extraction, classifier training, and post-processing), where the control marginally outperformed our approach by validating 35% of the patients. In total, 54 patients performed above chance for at least one method: our methodology or the control one. Of these 54 patients, 21 ([Formula: see text]38%) were solely validated by our methodology, while 24 ([Formula: see text]44%) were only validated by the control method. These findings may evidence the need for different methodologies concerning different patients.

Ensemble Deep Neural Network for Automatic Classification of EEG Independent Components.

OBJECTIVE: Independent component analysis (ICA) is commonly used to remove noisy artifacts from multi-channel scalp electroencephalogram (EEG) signals. ICA decomposes EEG into different independent components (ICs) and then, experts remove the noisy ones. This process is highly time-consuming and experts are not always available. To surpass this drawback, research is going on to develop models to automatically conduct IC classification. Current state-of-the-art models use power spectrum densities (PSDs) and topoplots to classify ICs. The performance of these methods may be limited by disregarding the IC time-series that would contribute to fully simulate the visual inspection performed by experts. METHODS: We present a novel ensemble deep neural network that combines time-series, PSDs, and topoplots to classify ICs. Moreover, we study the ability to use our model in transfer learning approaches. RESULTS: Experimental results showed that using time-series improves IC classification. Results also indicated that transfer learning obtained higher performance than simply training a new model from scratch. CONCLUSION: Researchers should develop IC classifiers using the three sources of information. Moreover, transfer learning approaches should be considered when producing new deep learning models. SIGNIFICANCE: This work improves IC classification, enhancing the automatic removal of EEG artifacts. Additionally, since labelled ICs are generally not publicly available, the possibility of using our model in transfer learning studies may motivate other researchers to develop their own classifiers.