RESUMEN

The ethnopharmacological relevance of the Millingtonia hortensis (M. hortensis) flower's aqueous extract lies in its traditional use as a herbal remedy in Southeast Asia. With a rich history in folk medicine, this aqueous has been esteemed for its purported sedative and anxiolytic properties. Our research delves into the scientific basis of these traditional claims, exploring the potential mechanisms underlying the observed effects of M. hortensis flower's aqueous extract on sleep promotion and mood regulation. This study aimed to explore the sleep-promoting effects of M. hortensis dried flower in mice, using an aqueous concentration equivalent to a human concentration of 2.7 mg/mL. Anxiolytic and antidepressant properties were evaluated using behavioural tests, while electroencephalography (EEG) analysis probed the neural mechanisms underlying sleep promotion post-consumption. The aqueous extract of M. hortensis dried flower administered to mice showed a decrease in immobility in the forced swimming test, demonstrating antidepressant-like effects. Moreover, mice treated with M. hortensis aqueous exhibited increased non-rapid eye movement (NREM) sleep duration, corroborating sleep-promoting potential. EEG analysis of mice treated with M. hortensis aqueous revealed heightened beta oscillations in the frontal and parietal cortex, while pre-treatment with M. hortensis aqueous or diazepam enhanced rapid eye movement (REM) sleep after thiopental administration. Interestingly, M. hortensis aqueous pre-treatment augmented delta frequency ranges in the frontal cortex. Overall, these findings indicate that M. hortensis dried flower's aqueous extract, at a human-equivalent dosage, exerts significant behavioural and neural effects specifically, sedative and hypnotic aspects in mice, corroborating its potential as a natural remedy to promote sleep and regulate mood.

RESUMEN

STUDY OBJECTIVES: To assess whether the frequency content of electroencephalography (EEG) and electrooculography (EOG) during nocturnal polysomnography (PSG) can predict all-cause mortality. METHODS: Power spectra from PSGs of 8,716 participants, included from the MrOS Sleep Study and the Sleep Heart Health Study (SHHS), were analyzed in deep learning-based survival models. The best-performing model was further examined using SHapley Additive Explanation (SHAP) for data-driven sleep-stage specific definitions of power bands, which were evaluated in predicting mortality using Cox Proportional Hazards models. RESULTS: Survival analyses, adjusted for known covariates, identified multiple EEG frequency bands across all sleep stages predicting all-cause mortality. For EEG, we found an all-cause mortality hazard ratio (HR) of 0.90 (CI95% 0.85-0.96) for 12-15 Hz in N2, 0.86 (CI95% 0.82-0.91) for 0.75-1.5 Hz in N3, and 0.87 (CI95% 0.83-0.92) for 14.75-33.5 Hz in REM sleep. For EOG, we found several low-frequency effects including an all-cause mortality HR of 1.19 (CI95% 1.11-1.28) for 0.25 Hz in N3, 1.11 (CI95% 1.03-1.21) for 0.75 Hz in N1, and 1.11 (CI95% 1.03-1.20) for 1.25-1.75 Hz in Wake. The gain in the concordance index (C-index) for all-cause mortality is minimal, with only a 0.24% increase: The best single mortality predictor was EEG N3 (0-0.5 Hz) with C-index of 77.78% compared to 77.54% for confounders alone. CONCLUSION: Spectral power features, possibly reflecting abnormal sleep microstructure, are associated with mortality risk. These findings add to a growing literature suggesting that sleep contains incipient predictors of health and mortality.

RESUMEN

AIM: Opioids might be harmful to the developing brain and dosing accuracy is important. We aimed at investigating fentanyl effects on cortical activity in infants using computational re-analysis of bedside recorded EEG signals. METHODS: Fifteen infants born at median 26.4 gestational weeks (range 23.3-34.1), with a birth weight 740 grams (530-1420) and postnatal age 7 days (5-11) received fentanyl 0.5 or 2 µg/kg intravenously before a skin-breaking procedure or tracheal intubation, respectively. Cortical activity was continuously recorded using amplitude-integrated electroencephalography (aEEG).  Analyses using three computational EEG features representing cortical synchrony and signal power, were conducted five minutes pre- and 10 minutes post the drug administration. RESULTS: Visual assessment of trends displayed from the EEG metrics did not indicate systematic changes. However, the magnitude of the changes in the parietal and right hemisphere signals after the dose was significantly correlated (ρ < -0.5, p < 0.05) to the EEG amplitude and frequency power level before drug administration. This effect started after 3-4 min. CONCLUSION: Fentanyl, even in small doses, may affect cortical activity in the preterm brain. The effect is robustly related to the state of cortical activity prior to drug treatment, which must be taken into account when analysing the effects of sedative drugs.

RESUMEN

Cell replacement therapies using medial ganglionic eminence (MGE)-derived GABAergic precursors reduce seizures by restoring inhibition in animal models of epilepsy. However, how MGE-derived cells affect abnormal neuronal networks and consequently brain oscillations to reduce ictogenesis is still under investigation. We performed quantitative analysis of pre-ictal local field potentials (LFP) of cortical and hippocampal CA1 areas recorded in vivo in the pilocarpine rat model of epilepsy, with or without intrahippocampal MGE-precursor grafts (PILO and PILO+MGE groups, respectively). The PILO+MGE animals had a significant reduction in the number of seizures. The quantitative analysis of pre-ictal LFP showed decreased power of cortical and hippocampal delta, theta and beta oscillations from the 5 min. interictal baseline to the 20 s. pre-ictal period in both groups. However, PILO+MGE animals had higher power of slow and fast oscillations in the cortex and lower power of slow and fast oscillations in the hippocampus compared to the PILO group. Additionally, PILO+MGE animals exhibited decreased cortico-hippocampal synchrony for theta and gamma oscillations at seizure onset and lower hippocampal CA1 synchrony between delta and theta with slow gamma oscillations compared to PILO animals. These findings suggest that MGE-derived cell integration into the abnormally rewired network may help control ictogenesis.

Asunto(s)

RESUMEN

PURPOSE: Previous research has established that food intake is a biological regulator of the human sleep-wake cycle. As such, the timing of eating relative to sleep may influence the quality of sleep, including daytime naps. Here, we examine whether the timing of lunch (1 h vs. 2 h interval between lunch and a napping opportunity) impacts the quality of an afternoon nap. METHODS: Using a randomized within-subject design over two separate experimental sessions (7 days apart), participants (n = 40, mean age = 25.8 years) consumed lunch 1 h and 2 h prior to an afternoon nap opportunity. Polysomnography and subjective self-reports were used to assess sleep architecture, sleepiness levels, and nap quality. RESULTS: Results revealed no significant differences in subjective ratings of sleep quality and sleepiness, or in sleep architecture (total sleep time, sleep efficiency, sleep onset latency, sleep stages) between the 1 h and 2-h lunch conditions. CONCLUSIONS: All sleep measures were similar when napping followed eating by either 1 h or 2 h, suggesting that eating closer to nap onset may not negatively impact sleep architecture and quality. Future research should continue to identify conditions that improve nap quality, given the well-documented benefits of naps to reduce sleep pressure and improve human performance.

RESUMEN

OBJECTIVE: The automated interpretation of clinical electroencephalograms (EEGs) using artificial intelligence (AI) holds the potential to bridge the treatment gap in resource-limited settings and reduce the workload at specialized centers. However, to facilitate broad clinical implementation, it is essential to establish generalizability across diverse patient populations and equipment. We assessed whether SCORE-AI demonstrates diagnostic accuracy comparable to that of experts when applied to a geographically different patient population, recorded with distinct EEG equipment and technical settings. METHODS: We assessed the diagnostic accuracy of a "fixed-and-frozen" AI model, using an independent dataset and external gold standard, and benchmarked it against three experts blinded to all other data. The dataset comprised 50% normal and 50% abnormal routine EEGs, equally distributed among the four major classes of EEG abnormalities (focal epileptiform, generalized epileptiform, focal nonepileptiform, and diffuse nonepileptiform). To assess diagnostic accuracy, we computed sensitivity, specificity, and accuracy of the AI model and the experts against the external gold standard. RESULTS: We analyzed EEGs from 104 patients (64 females, median age = 38.6 [range = 16-91] years). SCORE-AI performed equally well compared to the experts, with an overall accuracy of 92% (95% confidence interval [CI] = 90%-94%) versus 94% (95% CI = 92%-96%). There was no significant difference between SCORE-AI and the experts for any metric or category. SCORE-AI performed well independently of the vigilance state (false classification during awake: 5/41 [12.2%], false classification during sleep: 2/11 [18.2%]; p = .63) and normal variants (false classification in presence of normal variants: 4/14 [28.6%], false classification in absence of normal variants: 3/38 [7.9%]; p = .07). SIGNIFICANCE: SCORE-AI achieved diagnostic performance equal to human experts in an EEG dataset independent of the development dataset, in a geographically distinct patient population, recorded with different equipment and technical settings than the development dataset.

RESUMEN

The first night in an unfamiliar environment is marked by reduced sleep quality and changes in sleep architecture. This so-called First-Night Effect (FNE) is well established for two consecutive nights and lays the foundation for including an adaptation night in sleep research to counteract FNEs. However, adaptation nights rarely happen immediately before experimental nights, which raises the question of how sleep adapts over non-consecutive nights. Furthermore, it is yet unclear, how environmental familiarity and hemispheric asymmetry of slow-wave sleep (SWS) contribute to the explanation of FNEs. To address this gap, 45 healthy participants spent two weekly separated nights in the sleep laboratory. In a separate study, we investigated the influence of environmental familiarity on 30 participants who spent two non-consecutive nights in the sleep laboratory and two nights at home. Sleep was recorded by polysomnography. Results of both studies show that FNEs also occur in non-consecutive nights, particularly affecting wake after sleep onset, sleep onset latency, and total sleep time. Sleep disturbances in the first night happen in both familiar and unfamiliar environments. The degree of asymmetric SWS was not correlated with the FNE but rather tended to vary over the course of several nights. Our findings suggest that non-consecutive adaptation nights are effective in controlling for FNEs, justifying the current practice in basic sleep research. Further research should focus on trait- and fluctuating state-like components explaining interhemispheric asymmetries.

RESUMEN

OBJECTIVE: Brivaracetam (BRV) is a recent antiseizure medication (ASM) approved as an add-on therapy for people with focal epilepsy. BRV has a good efficacy and safety profile compared to other ASMs. However, its specific effects on resting-state EEG activity and connectivity are unknown. The aim of this study is to evaluate quantitative EEG changes induced by BRV therapy in a population of adult people with drug-resistant epilepsy (PwE) compared to healthy controls (HC). METHODS: We performed a longitudinal, retrospective, pharmaco-EEG study on a population of 23 PwE and a group of 25 HC. Clinical outcome was dichotomized into drug-responders (i.e., >50% reduction in seizures' frequency; RES) and non-responders (N-RES) after two years of BRV. EEG parameters were compared between PwE and HC at baseline (pre-BRV) and after three months of BRV therapy (post-BRV). We investigated BRV-related variations in EEG connectivity using the phase locking value (PLV). RESULTS: BRV therapy did not induce modifications in power spectrum density across different frequency bands. PwE presented lower PLV connectivity values compared to HC in all frequency bands. RES exhibited lower theta PLV connectivity compared to HC before initiating BRV and experienced an increase after BRV, eliminating the significant difference from HC. CONCLUSIONS: This study shows that BRV does not alter the EEG power spectrum in PwE, supporting its favourable neuropsychiatric side-effect profile, and induces the disappearance of EEG connectivity differences between PwE and HC. SIGNIFICANCE: The integration of EEG quantitative analysis in epilepsy can provide insights into the efficacy, mechanism of action, and side effects of ASMs.

Asunto(s)

RESUMEN

STUDY OBJECTIVES: Post-hoc analysis to evaluate the effect of daridorexant on sleep architecture in people with insomnia, focusing on features associated with hyperarousal. METHODS: We studied sleep architecture in adults with chronic insomnia disorder from two randomized Phase 3 clinical studies (Clinicaltrials.gov: NCT03545191 and NCT03575104) investigating 3 months of daridorexant treatment (placebo, daridorexant 25 mg, daridorexant 50 mg). We analyzed sleep-wake transition probabilities, EEG spectra and sleep spindle properties including density, dispersion, and slow oscillation phase coupling. The Wake EEG Similarity Index (WESI) was determined using a machine learning algorithm analyzing the spectral profile of the EEG. RESULTS: At Month 3, daridorexant 50 mg decreased Wake-to-Wake transition probabilities (P<0.05) and increased the probability of transitions from Wake-to-N1 (P<0.05), N2 (P<0.05), and REM sleep (P<0.05), as well as from N1-to-N2 (P<0.05) compared to baseline and placebo. Daridorexant 50 mg decreased relative beta power during Wake (P=0.011) and N1 (P<0.001) compared to baseline and placebo. During Wake, relative alpha power decreased (P<0.001) and relative delta power increased (P<0.001) compared to placebo. Daridorexant did not alter EEG spectra bands in N2, N3, and REM stages or in sleep spindle activity. Daridorexant decreased the WESI score during Wake compared to baseline (P=0.004). Effects with 50 mg were consistent between Month 1 and Month 3 and less pronounced with 25 mg. CONCLUSION: Daridorexant reduced EEG features associated with hyperarousal as indicated by reduced Wake-to-Wake transition probabilities and enhanced spectral features associated with drowsiness and sleep during Wake and N1.

RESUMEN

In disorders of consciousness, verticalization is considered an effective type of treatment to improve motor and cognitive recovery. Our purpose is to investigate neurophysiological effects of robotic verticalization training (RVT) in patients with minimally conscious state (MCS). Thirty subjects affected by MCS due to traumatic or vascular brain injury, attending the intensive Neurorehabilitation Unit of the IRCCS Neurolesi (Messina, Italy), were included in this retrospective study. They were equally divided into two groups: the control group (CG) received traditional verticalization with a static bed and the experimental group (EG) received advanced robotic verticalization using the Erigo device. Each patient was evaluated using both clinical scales, including Levels of Cognitive Functioning (LCF) and Functional Independence Measure (FIM), and quantitative EEG pre (T0) and post each treatment (T1). The treatment lasted for eight consecutive weeks, and sessions were held three times a week, in addition to standard neurorehabilitation. In addition to a notable improvement in clinical parameters, such as functional (FIM) (p < 0.01) and cognitive (LCF) (p < 0.01) outcomes, our findings showed a significant modification in alpha and beta bands post-intervention, underscoring the promising effect of the Erigo device to influence neural plasticity and indicating a noteworthy difference between pre-post intervention. This was not observed in the CG. The observed changes in alpha and beta bands underscore the potential of the Erigo device to induce neural plasticity. The device's custom features and programming, tailored to individual patient needs, may contribute to its unique impact on brain responses.

RESUMEN

BACKGROUND: Self-recorded EEG by patients at home might present a viable alternative to inpatient epilepsy evaluations. OBJECTIVES AND METHODS: We developed a novel telemonitoring system comprising seamlessly integrated hard- and software with automated AI-based EEG analysis. RESULTS: The first complete study participation results demonstrate feasibility and clinical utility. CONCLUSION: Our telemonitoring solution potentially improves treatment of patients with epilepsy and moreover might help to better distribute resources in the healthcare system.

Asunto(s)

RESUMEN

Electroencephalogram (EEG) analysis plays an indispensable role across contemporary medical applications, which encompasses diagnosis, monitoring, drug discovery, and therapeutic assessment. This work puts forth an end-to-end deep learning framework that is uniquely tailored for versatile EEG analysis tasks by directly operating on raw waveform inputs. It aims to address the challenges of manual feature engineering and the neglect of spatial interrelationships in existing methodologies. Specifically, a spatial channel attention module is introduced to emphasize the critical inter-channel dependencies in EEG signals through channel statistics aggregation and multi-layer perceptron operations. Furthermore, a sparse transformer encoder is used to leverage selective sparse attention in order to efficiently process long EEG sequences while reducing computational complexity. Distilling convolutional layers further concatenates the temporal features and retains only the salient patterns. As it was rigorously evaluated on key EEG datasets, our model consistently accomplished a superior performance over the current approaches in detection and classification assignments. By accounting for both spatial and temporal relationships in an end-to-end paradigm, this work facilitates a versatile, automated EEG understanding across diseases, subjects, and objectives through a singular yet customizable architecture. Extensive empirical validation and further architectural refinement may promote broader clinical adoption prospects.

RESUMEN

Operator fatigue poses a major concern in safety-critical industries such as aviation, potentially increasing the chances of errors and accidents. To better understand this risk, there is a need for noninvasive objective measures of fatigue. This study aimed to evaluate the performance of cEEGrids, a type of ear-EEG, for fatigue detection by analyzing the alpha and theta power before and after sleep restriction in four sessions on two separate days, employing a within-participants design. Results were compared to traditional, highly validated methods: the Karolinska Sleepiness Scale (KSS) and Psychomotor Vigilance Task (PVT). After sleep restriction and an office workday, 12 participants showed increased alpha band power in multiple electrode channels, but no channels correlated with KSS scores and PVT response speed. These findings indicate that cEEGrids can detect differences in alpha power following mild sleep loss. However, it should be noted that this capability was limited to specific channels, and no difference in theta power was observed. The study shows the potential and limitations of ear-EEG for fatigue detection as a less invasive alternative to cap-EEG. Further design and electrode configuration adjustments are necessary before ear-EEG can be implemented for fatigue detection in the field.

RESUMEN

OBJECTIVE: Electroencephalographic (EEG) microstate abnormalities have been documented in different neurological disorders. We aimed to assess whether EEG microstates are altered also in patients with temporal epilepsy (TLE) and whether they show different activations in patients with unilateral TLE (UTLE) and bilateral TLE (BTLE). METHODS: Nineteen patients with UTLE, 12 with BTLE, and 15 healthy controls were enrolled. Resting state high-density electroencephalography (128 channels) was recorded for 15 min with closed eyes. We obtained a set of stable scalp maps representing the EEG activity, named microstates, from which we acquired the following variables: global explained variance (GEV), mean duration (MD), time coverage (TC), and frequency of occurrence (FO). Two-way repeated measures analysis of variance was used to compare groups, and Spearman correlation was performed to study the maps in association with the clinical and neuropsychological data. RESULTS: Patients with BTLE and UTLE showed differences in most of the parameters (GEV, MD, TC, FO) of the four microstate maps (A-D) compared to controls. Patients with BTLE showed a significant increase in all parameters for the microstates in Map-A and a decrease in Map-D compared to UTLE and controls. We observed a correlation between Map-A, disease duration, and spatial short-term memory, whereas microstate Map-D was correlated with the global intelligence score and short-term memory performance. SIGNIFICANCE: A global alteration of the neural dynamics was observed in patients with TLE compared to controls. A different pattern of EEG microstate abnormalities was identified in BTLE compared to UTLE, which might represent a distinctive biomarker.

Asunto(s)

RESUMEN

In recent years, the use of electroencephalography (EEG) has grown as a tool for diagnostic and brain function monitoring, being a simple and non-invasive method compared with other procedures like histological sampling. Typically, in order to extract functional brain responses from EEG signals, prolonged and repeated stimuli are needed because of the artifacts generated in recordings which adversely impact the stimulus-response analysis. To mitigate the artifact effect, correlation analysis (CA) methods are applied in the literature, where the predominant approaches focus on enhancing stimulus-response correlations through the use of linear analysis methods like canonical correlation analysis (CCA). This paper introduces a novel CA framework based on a neural network with a loss function specifically designed to maximize correlation between EEG and speech stimuli. Compared with other deep learning CA approaches (DCCAs) in the literature, this framework introduces a single multilayer perceptron (MLP) network instead of two networks for each stimulus. To validate the proposed approach, a comparison with linear CCA (LCCA) and DCCA was performed, using a dataset containing the EEG traces of subjects listening to speech stimuli. The experimental results show that the proposed method improves the overall Pearson correlation by 10.56% compared with the state-of-the-art DCCA method.

Asunto(s)

RESUMEN

An EEG signal (Electroencephalogram) is a bioelectric phenomenon reflecting human brain activities. In this paper, we propose a novel deep learning framework ESML (EEG-based Subject Matching Learning) using raw EEG signals to learn latent representations for EEG-based user identification and tack classification. ESML consists of two parts: one is the ESML1 model via an LSTM-based method for EEG-user linking, and one is the ESML2 model via a CNN-based method for EEG-task linking. The new model ESML is simple, but effective and efficient. It does not require any restrictions for EEG data collection on motions and thinking for users, and it does not need any EEG preprocessing operations, such as EEG denoising and feature extraction. The experiments were conducted on three public datasets and the results show that ESML performs the best and achieves significant performance improvement when compared to baseline methods (i.e., SVM, LDA, NN, DTS, Bayesian, AdaBoost and MLP). The ESML1 model provided the best precision at 96% with 109 users and the ESML2 model achieved 99% precision at 3-Class task classification. These experimental results provide direct evidence that EEG signals can be used for user identification and task classification.

RESUMEN

Announced by WHO in 2020, the global COVID-19 pandemic caused by SARS-CoV-2 has affected many people, leading to serious health consequences. These consequences are observed in the daily lives of infected patients as various dysfunctions and limitations. More and more people are suffering post-COVID-19 complications that interfere with or completely prevent them from working or even functioning independently on a daily basis. The aim of our study was to demonstrate that innovative quantitative electroencephalography (QEEG) can be used to assess cognitive function disorders reported after the COVID-19 pandemic. It is worth noting that no similar study has been conducted to date in a group of pilots. The QEEG method we used is currently one of the basic neurological examinations, enabling easy observation of post-COVID-19 changes in the nervous system. With the innovativeness of this technique, our study shows that the use of quantitative electroencephalography can be a precursor in identifying complications associated with cognitive function disorders after COVID-19. Our study was conducted on twelve 26-year-old pilots. All participants had attended the same flight academy and had contracted SARS-CoV-2 infection. The pilots began to suspect COVID-19 infection when they developed typical symptoms such as loss of smell and taste, respiratory problems, and rapid fatigue. Quantitative electroencephalography (QEEG), which is one of the most innovative forms of diagnostics, was used to diagnose the patients. Comparison of the results between the study and control groups showed significantly higher values of all measurements of alpha, theta, and beta2 waves in the study group. In the case of the sensorimotor rhythm (SMR), the measurement results were significantly higher in the control group compared to the study group. Our study, conducted on pilots who had recovered from COVID-19, showed changes in the amplitudes of brain waves associated with relaxation and concentration. The results confirmed the issues reported by pilots as evidenced by the increased amplitudes of alfa, theta, and beta2 waves. It should be emphasized that the modern diagnostic method (QEEG) presented here has significant importance in the medical diagnosis of various symptoms and observation of treatment effects in individuals who have contracted the SARS-CoV-2 virus. The present study demonstrated an innovative approach to the diagnosis of neurological complications after COVID-19.

RESUMEN

Whole-night sleep electroencephalogram (EEG) is plagued by several types of large-amplitude artifacts. Common approaches to remove them are fraught with issues: channel interpolation, rejection of noisy intervals, and independent component analysis are time-consuming, rely on subjective user decisions, and result in signal loss. Artifact Subspace Reconstruction (ASR) is an increasingly popular approach to rapidly and automatically clean wake EEG data. Indeed, ASR adaptively removes large-amplitude artifacts regardless of their scalp topography or consistency throughout the recording. This makes ASR, at least in theory, a highly-promising tool to clean whole-night EEG. However, ASR crucially relies on calibration against a subset of relatively clean "baseline" data. This is problematic when the baseline changes substantially over time, as in whole-night EEG data. Here we tackled this issue and, for the first time, validated ASR for cleaning sleep EEG. We demonstrate that ASR applied out-of-the-box, with the parameters recommended for wake EEG, results in the dramatic removal of slow waves. We also provide an appropriate procedure to use ASR for automatic and rapid cleaning of whole-night sleep EEG data or any long EEG recording. Our procedure is freely available in Dusk2Dawn, an open-source plugin for EEGLAB.

Asunto(s)

RESUMEN

BACKGROUND AND OBJECTIVE: Epilepsy is a serious brain disorder affecting more than 50 million people worldwide. If epileptic seizures can be predicted in advance, patients can take measures to avoid unfortunate consequences. Important approaches for epileptic seizure predictions are often signal transformation and classification using electroencephalography (EEG) signals. A time-frequency (TF) transformation, such as the short-term Fourier transform (STFT), has been widely used over many years but curtailed by the Heisenberg uncertainty principle. This research focuses on decomposing epileptic EEG signals with a higher resolution so that an epileptic seizure can be predicted accurately before its episodes. METHODS: This study applies a synchroextracting transformation (SET) and singular value decomposition (SET-SVD) to improve the time-frequency resolution. The SET is a more energy-concentrated TF representation than classical TF analysis methods. RESULTS: The pre-seizure classification method employing a 1-dimensional convolutional neural network (1D-CNN) reached an accuracy of 99.71% (the CHB-MIT database) and 100% (the Bonn University database). The experiments on the CHB-MIT show that the accuracy, sensitivity and specificity from the SET-SVD method, compared with the results of the STFT, are increased by 8.12%, 6.24% and 13.91%, respectively. In addition, a multi-layer perceptron (MLP) was also used as a classifier. Its experimental results also show that the SET-SVD generates a higher accuracy, sensitivity and specificity by 5.0%, 2.41% and 11.42% than the STFT, respectively. CONCLUSIONS: The results of two classification methods (the MLP and 1D-CNN) show that the SET-SVD has the capacity to extract more accurate information than the STFT. The 1D-CNN model is suitable for a fast and accurate patient-specific EEG classification.

Asunto(s)

RESUMEN

Around one-third of epilepsy patients develop drug-resistant seizures; early detection of seizures could help improve safety, reduce patient anxiety, increase independence, and enable acute treatment. In recent years, the use of artificial intelligence techniques and machine learning algorithms in different diseases, including epilepsy, has increased significantly. The main objective of this study is to determine whether the mjn-SERAS artificial intelligence algorithm developed by MJN Neuroserveis, can detect seizures early using patient-specific data to create a personalized mathematical model based on EEG training, defined as the programmed recognition of oncoming seizures before they are primarily initiated, usually within a period of a few minutes, in patients diagnosed of epilepsy. Retrospective, cross-sectional, observational, multicenter study to determine the sensitivity and specificity of the artificial intelligence algorithm. We searched the database of the Epilepsy Units of three Spanish medical centers and selected 50 patients evaluated between January 2017 and February 2021, diagnosed with refractory focal epilepsy who underwent video-EEG monitoring recordings between 3 and 5 days, a minimum of 3 seizures per patient, lasting more than 5 s and the interval between each seizure was greater than 1 h. Exclusion criteria included age <18 years, intracranial EEG monitoring, and severe psychiatric, neurological, or systemic disorders. The algorithm identified pre-ictal and interictal patterns from EEG data using our learning algorithm and was compared to a senior epileptologist's evaluation as a gold standard. Individual mathematical models of each patient were trained using this feature dataset. A total of 1963 h of 49 video-EEG recordings were reviewed, with an average of 39.26 h per patient. The video-EEG monitoring recorded 309 seizures as subsequently analyzed by the epileptologists. The mjn-SERAS algorithm was trained on 119 seizures and split testing was performed on 188 seizures. The statistical analysis includes the data from each model and reports 10 false negatives (no detection of episodes recorded by video-EEG) and 22 false positives (alert detected without clinical correlation or abnormal EEG signal within 30 min). Specifically, the automated mjn-SERAS AI algorithm achieved a sensitivity of 94.7% (95 %; CI 94.67-94.73), and an F-Score representing specificity of 92.2% (95 %; CI 92.17-92.23) compared to the reference performance represented by a mean (harmonic mean or average) and a positive predictive value of 91%, with a false positive rate of 0.55 per 24 h in the patient-independent model. This patient-specific AI algorithm for early seizure detection shows promising results in terms of sensitivity and false positive rate. Although the algorithm requires high computational requirements on specialized servers cloud for training and computing, its computational load in real-time is low, allowing its implementation on embedded devices for online seizure detection.