Network controllability analysis of awake and asleep conditions in the brain / 浙江大学学报（英文版）（B辑：生物医学和生物技术）

The difference between sleep and wakefulness is critical for human health. Sleep takes up one third of our lives and remains one of the most mysterious conditions; it plays an important role in memory consolidation and health restoration. Distinct neural behaviors take place under awake and asleep conditions, according to neuroimaging studies. While disordered transitions between wakefulness and sleep accompany brain disease, further investigation of their specific characteristics is required. In this study, the difference is objectively quantified by means of network controllability. We propose a new pipeline using a public intracranial stereo-electroencephalography (stereo-EEG) dataset to unravel differences in the two conditions in terms of system neuroscience. Because intracranial stereo-EEG records neural oscillations covering large-scale cerebral areas, it offers the highest temporal resolution for recording neural behaviors. After EEG preprocessing, the EEG signals are band-passed into sub-slow (0.1‍-‍1 Hz), delta (1‍-‍4 Hz), theta (4‍-‍8 Hz), alpha (8‍-‍13 Hz), beta (13‍-‍30 Hz), and gamma (30‍-‍45 Hz) band oscillations. Then, dynamic functional connectivity is extracted from time-windowed EEG neural oscillations through phase-locking value (PLV) and non-overlapping sliding time windows. Next, average and modal network controllability are implemented on these time-varying brain networks. Based on this preliminary study, it appears that significant differences exist in the dorsolateral frontal-parietal network (FPN), salience network (SN), and default-mode network (DMN). The combination of network controllability and dynamic functional networks offers new insight for characterizing distinctions between awake and asleep stages in the brain. In other words, network controllability captures the underlying brain dynamics under both awake and asleep conditions.

Clinical efficacy of mild therapeutic hypothermia with different rewarming time on neonatal hypoxic-ischemic encephalopathy: a prospective randomized controlled study / 中国当代儿科杂志

OBJECTIVES@#To investigate the clinical efficacy of mild therapeutic hypothermia (MTH) with different rewarming time on neonatal hypoxic-ischemic encephalopathy (HIE).@*METHODS@#A prospective study was performed on 101 neonates with HIE who were born and received MTH in Zhongshan Hospital, Xiamen University, from January 2018 to January 2022. These neonates were randomly divided into two groups: MTH1 group (n=50; rewarming for 10 hours at a rate of 0.25°C/h) and MTH2 group (n=51; rewarming for 25 hours at a rate of 0.10°C/h). The clinical features and the clinical efficacy were compared between the two groups. A binary logistic regression analysis was used to identify the factors influencing the occurrence of normal sleep-wake cycle (SWC) on amplitude-integrated electroencephalogram (aEEG) at 25 hours of rewarming.@*RESULTS@#There were no significant differences between the MTH1 and MTH2 groups in gestational age, 5-minute Apgar score, and proportion of neonates with moderate/severe HIE (P>0.05). Compared with the MTH2 group, the MTH1 group tended to have a normal arterial blood pH value at the end of rewarming, a significantly shorter duration of oxygen dependence, a significantly higher proportion of neonates with normal SWC on aEEG at 10 and 25 hours of rewarming, and a significantly higher Neonatal Behavioral Neurological Assessment score on days 5, 12, and 28 after birth (P<0.05), while there was no significant difference in the incidence rate of rewarming-related seizures between the two groups (P>0.05). There were no significant differences between the two groups in the incidence rate of neurological disability at 6 months of age and the score of Bayley Scale of Infant Development at 3 and 6 months of age (P>0.05). The binary logistic regression analysis showed that prolonged rewarming time (25 hours) was not conducive to the occurrence of normal SWC (OR=3.423, 95%CI: 1.237-9.469, P=0.018).@*CONCLUSIONS@#Rewarming for 10 hours has a better short-term clinical efficacy than rewarming for 25 hours. Prolonging rewarming time has limited clinical benefits on neonates with moderate/severe HIE and is not conducive to the occurrence of normal SWC, and therefore, it is not recommended as a routine treatment method.

Automatic sleep staging algorithm for stochastic depth residual networks based on transfer learning / 生物医学工程学杂志

The existing automatic sleep staging algorithms have the problems of too many model parameters and long training time, which in turn results in poor sleep staging efficiency. Using a single channel electroencephalogram (EEG) signal, this paper proposed an automatic sleep staging algorithm for stochastic depth residual networks based on transfer learning (TL-SDResNet). Firstly, a total of 30 single-channel (Fpz-Cz) EEG signals from 16 individuals were selected, and after preserving the effective sleep segments, the raw EEG signals were pre-processed using Butterworth filter and continuous wavelet transform to obtain two-dimensional images containing its time-frequency joint features as the input data for the staging model. Then, a ResNet50 pre-trained model trained on a publicly available dataset, the sleep database extension stored in European data format (Sleep-EDFx) was constructed, using a stochastic depth strategy and modifying the output layer to optimize the model structure. Finally, transfer learning was applied to the human sleep process throughout the night. The algorithm in this paper achieved a model staging accuracy of 87.95% after conducting several experiments. Experiments show that TL-SDResNet50 can accomplish fast training of a small amount of EEG data, and the overall effect is better than other staging algorithms and classical algorithms in recent years, which has certain practical value.

Automatic sleep staging based on power spectral density and random forest / 生物医学工程学杂志

The method of using deep learning technology to realize automatic sleep staging needs a lot of data support, and its computational complexity is also high. In this paper, an automatic sleep staging method based on power spectral density (PSD) and random forest is proposed. Firstly, the PSDs of six characteristic waves (K complex wave, δ wave, θ wave, α wave, spindle wave, β wave) in electroencephalogram (EEG) signals were extracted as the classification features, and then five sleep states (W, N1, N2, N3, REM) were automatically classified by random forest classifier. The whole night sleep EEG data of healthy subjects in the Sleep-EDF database were used as experimental data. The effects of using different EEG signals (Fpz-Cz single channel, Pz-Oz single channel, Fpz-Cz + Pz-Oz dual channel), different classifiers (random forest, adaptive boost, gradient boost, Gaussian naïve Bayes, decision tree, K-nearest neighbor), and different training and test set divisions (2-fold cross-validation, 5-fold cross-validation, 10-fold cross-validation, single subject) on the classification effect were compared. The experimental results showed that the effect was the best when the input was Pz-Oz single-channel EEG signal and the random forest classifier was used, no matter how the training set and test set were transformed, the classification accuracy was above 90.79%. The overall classification accuracy, macro average F1 value, and Kappa coefficient could reach 91.94%, 73.2% and 0.845 respectively at the highest, which proved that this method was effective and not susceptible to data volume, and had good stability. Compared with the existing research, our method is more accurate and simpler, and is suitable for automation.

A research on epilepsy source localization from scalp electroencephalograph based on patient-specific head model and multi-dipole model / 生物医学工程学杂志

Accurate source localization of the epileptogenic zone (EZ) is the primary condition of surgical removal of EZ. The traditional localization results based on three-dimensional ball model or standard head model may cause errors. This study intended to localize the EZ by using the patient-specific head model and multi-dipole algorithms using spikes during sleep. Then the current density distribution on the cortex was computed and used to construct the phase transfer entropy functional connectivity network between different brain areas to obtain the localization of EZ. The experiment result showed that our improved methods could reach the accuracy of 89.27% and the number of implanted electrodes could be reduced by (19.34 ± 7.15)%. This work can not only improve the accuracy of EZ localization, but also reduce the additional injury and potential risk caused by preoperative examination and surgical operation, and provide a more intuitive and effective reference for neurosurgeons to make surgical plans.

Levetiracetam plus Oxcarbazepine Combination Treatment Downregulates Serum Multidrug Resistance Protein 1 Levels and Upregulates Neuropeptide Y Levels in Children with Epilepsy

Abstract The aim of the present study was to investigate the usefulness of multidrug resistance protein 1 (MDR1) and neuropeptide Y (NPY) levels in predicting the efficacy of levetiracetam (LEV) plus oxcarbazepine (OXC) treatment administered to children with epilepsy and to determine their prognosis. Overall, 193 children with epilepsy admitted to the hospital were enrolled and randomly divided into two groups according to different treatment methods: group A (n = 106, treated with LEV plus OXC combination) and group B (n = 87, treated with OXC only). After treatment, compared with group B, group A exhibited a remarkably higher total effective rate and a significantly lower total adverse reaction rate. Areas under the curve for MDR1 and NPY for predicting ineffective treatment were 0.867 and 0.834, whereas those for predicting epilepsy recurrence were 0.916 and 0.829, respectively. Electroencephalography abnormalities, intracranial hemorrhage, neonatal convulsion, premature delivery, and MDR1 and NPY levels were independent risk factors for poor prognosis in children with epilepsy. Serum MDR1 and NPY levels exhibited a high predictive value for early epilepsy diagnosis, treatment efficacy assessment, and prognostication in children with epilepsy treated with LEV plus OXC combination.

Fusión de datos multimodales de neurociencias generados en estudios EEG-fMRI, un análisis temático / Fusion of Multimodal Neuroscience Data Generated in EEG-fMRI Studies, a Thematic Analysis

La actividad cerebral tiene múltiples atributos, entre ellos los eléctricos, metabólicos, hemodinámicos y hormonales. Los métodos modernos para estudiar las funciones cerebrales como el PET (Tomografía por Emisión de Positrones), fMRI (Imagen de Resonancia Magnética Funcional) y MEG (Magnetoencefalograma) son ampliamente utilizados por los científicos. Sin embargo, el EEG es una herramienta utilizada para la investigación y diagnóstico debido a su bajo costo, simplicidad de uso, movilidad y la posibilidad de monitoreo a largo tiempo de adquisición. Para detectar e interpretar las características relevantes de estas señales, se describe cada proceso por su escala temporal (EEG) y espacial (fMRI). La presente investigación se enfoca en realizar una revisión bibliográfica sobre la integración de datos multimodales EEG-fMRI que propicie valorar su importancia para el desarrollo de algoritmos de fusión y su uso en el contexto cubano. Para ello se analizaron documentos con altos índices de citas en la literatura, donde se destacan autores precursores de los temas en análisis. Los estudios multimodales EEG-fMRI generan múltiples datos temporales y espaciales con alto valor para la medicina basada en evidencia. La integración de los mismos provee un valor agregado en la búsqueda de nuevos métodos diagnósticos, aplicando minería de datos, Deep learning y algoritmos de fusión. En este trabajo se pone de relieve la existencia de baja resolución temporal de fMRI y por otro lado la baja resolución espacial de EEG, por lo que la integración de ambos estudios aumentaría la calidad de su información(AU)

Brain activity has multiple attributes, including electrical, metabolic, hemodynamic, and hormonal. Modern methods for studying brain functions such as PET (Positron Emission Tomography), fMRI (Functional Magnetic Resonance Imaging), and MEG (Magnetoencephalogram) are widely used by scientists. However, the EEG is a tool used for research and diagnosis due to its low cost, simplicity of use, mobility and the possibility of long-term monitoring of acquisition. To detect and interpret the relevant characteristics of these signals, each process is described by its temporal (EEG) and spatial (fMRI) scale. The present research focuses on conducting a bibliographic review on the integration of multimodal EEG-fMRI data that favors assessing its importance for the development of fusion algorithms and their use in the Cuban context. For this, documents with high rates of citations in the literature were analyzed, where precursor authors of the topics under analysis stand out. Multimodal EEG-fMRI studies generate multiple temporal and spatial data with high value for evidence-based medicine. Their integration provides added value in the search for new diagnostic methods, applying data mining, Deep learning and fusion algorithms. This work highlights the existence of low temporal resolution of fMRI and, on the other hand, the low spatial resolution of EEG, so the integration of both studies would increase the quality of their information(AU)

Utilidad del monitoreo continuo del electroencefalograma en el paciente crítico / Utility of continuous electroencephalogram monitoring in the critical patients

Introducción: El monitoreo continuo del Electroencefalograma, es la recogida simultánea de la actividad cerebral y la conducta clínica por un período de horas a días. Por el alto costo de la técnica aún no está muy difundida. Objetivos: Evaluar la utilidad del monitoreo electroencefalográfico continuo en el paciente crítico. Métodos: Se realizó un estudio descriptivo, retrospectivo y longitudinal en 118 sujetos mayores de 19 años ingresados en las unidades de terapia del Hospital Clínico Quirúrgico Hermanos Ameijeiras; entre noviembre 2016 a octubre 2018 con indicación de un Electroencefalograma continuo. Se consideraron variables clínicas y electroencefalográficas: escala de Glasgow, ocurrencia de crisis, diagnóstico, estado al egreso, anormalidad del Electroencefalograma, descargas epileptiformes, sospecha de estatus epiléptico no convulsivo por electroencefalograma entre otras. Los datos se procesaron con test de Chi cuadrado, test de Mc Nemar y test t de student, se empleó un nivel de significación de p≤0.05. Resultados: 60 de los pacientes pertenecían al sexo femenino, la mediana de las edades fue 67,5 años. La escala de Glasgow mostró asociación significativa con el grado de anormalidad del electroencefalograma (p=0,001), es la arreactividad y la discontinuidad de la actividad de base predictores de pobre pronóstico. Se observaron descargas epileptiformes periódicas en 100 pacientes. Se definió estatus epiléptico no convulsivo en 56 sujetos (37,28 por ciento) y en 81 sujetos (68,64 por ciento) el resultado del electroencefalograma motivó una conducta médica. Conclusiones: El monitoreo continuo del electroencefalograma es útil en el diagnóstico y manejo del paciente con episodios no convulsivos, permite formular un pronóstico neurológico y orientó la conducta médica(AU)

Introduction: The continuous monitoring of the electroencephalogram is the simultaneous collection of brain activity and clinical behavior for a period of hours to days. Due to the high cost of the technique, it is not yet widely used. Objectives: To evaluate the usefulness of continuous electroencephalographic monitoring in critically ill patients. Methods: A descriptive, retrospective and longitudinal study was carried out in 118 subjects over 19 years of age admitted to the therapy units at Hermanos Ameijeiras Surgical Clinical Hospital; from November 2016 to October 2018. They were indicated a continuous electroencephalogram. Clinical and electroencephalographic variables were considered, such as Glasgow scale, seizure occurrence, diagnosis, discharge status, electroencephalogram abnormality, epileptiform discharges, suspicion of nonconvulsive status epilepticus by electroencephalogram, among others. The data was processed with the Chi square test, the Mc Nemar test and the student's t test, using significance level of p≤0.05. Results: Sixty patients were female, the median age was 67.5 years. The Glasgow scale showed significant association with the degree of electroencephalogram abnormality (p=0.001). A reactivity and discontinuity of baseline activity are predictors of poor prognosis. Periodic epileptiform discharges were observed in 100 patients. Non-convulsive status epilepticus was defined in 56 subjects (37.28 percent) and in 81 subjects (68.64 percent) the result of the electroencephalogram motivated a medical procedure. Conclusions: The continuous monitoring of the electroencephalogram is useful in the diagnosis and management of patients with non-convulsive episodes, it allows formulating a neurological prognosis and guided medical conduct(AU)

EEG education in Brazil: a national survey of adult neurology residents / Educação em EEG no Brasil: uma pesquisa nacional com residentes de neurologia de adultos

ABSTRACT Background: In light of the established challenges of resident EEG education worldwide, we sought to better understand the current state of neurology resident EEG education in Brazil. Objective: To define Brazilian EEG practices including in-residency requirements for EEG training and competency. Methods: We assessed the perspectives of adult residents (PGY1-3) on EEG education and their level of confidence interpreting EEG with a 24-question online survey. Results: We analyzed 102 responses from 52 Brazilian neurology residency programs distributed in 14 states. There were 18 PGY1s, 45 PGY2s, and 39 PGY3s. Ninety-six percent of participants reported that learning how to read EEG during residency was very or extremely important. The most commonly reported barriers to EEG education were insufficient EEG exposure (70%) and ineffective didactics (46%). Residents believed that standard EEG lectures were the most efficient EEG teaching method followed by interpreting EEG with attendings' supervision. Roughly half of residents (45%) reported not being able to read EEG even with supervision, and approximately 70% of all participants did not feel confident writing an EEG report independently. Conclusion: Despite the well-established residency EEG education requirements recommended by the Brazilian Academy of Neurology (ABN), there seems to be a significant lack of comfort interpreting EEG among Brazilian adult neurology residents. We encourage Brazilian neurology residency leadership to re-evaluate the current EEG education system in order to ensure that residency programs are following EEG education requirements and to assess whether EEG benchmarks require modifications.

RESUMO Antecedentes: Diante dos desafios da educação em EEG estabelecidos em todo o mundo, buscamos compreender melhor o estado atual da educação em EEG durante a residência de neurologia no Brasil. Objetivo: Investigar práticas de EEG no Brasil, incluindo requisitos para treinamento e competência durante a residência de neurologia. Métodos: Avaliamos as perspectivas dos residentes (R1-3) de neurologia (adulto) sobre educação em EEG e nível de confiança ao interpretá-lo através de questionário online de 24 perguntas. Resultados: Foram analisadas 102 respostas de 52 programas de residência distribuídos em 14 estados. Dezoito R1s, 45 R2s e 39 R3s responderam à pesquisa. Noventa e seis por cento dos participantes relataram que aprender a ler EEG durante a residência é muito ou extremamente importante. As barreiras mais relatadas para educação em EEG foram exposição insuficiente ao EEG (70%) e didática ineficaz (46%). Os participantes apontaram aulas como método de ensino mais eficaz, seguido pela interpretação do EEG supervisionada pelos chefes. Aproximadamente metade dos residentes (45%) relatou não ser capaz de ler EEG mesmo com supervisão e cerca de 70% não se sente confiante para escrever um laudo de EEG de forma independente. Conclusões: Apesar dos requisitos estabelecidos pela Academia Brasileira de Neurologia (ABN) sobre ensino de EEG durante a residência, há significativa falta de confiança na sua interpretação pelos residentes de neurologia (adulto). Incentivamos as lideranças a reavaliar o sistema de educação para garantir que os programas de residência sigam requisitos de educação em EEG e se os benchmarks de EEG requerem modificações.

Evaluation of the clinical effect of an artificial intelligence-assisted diagnosis and treatment system for neonatal seizures in the real world: a multicenter clinical study protocol / 中国当代儿科杂志

Neonatal seizures are the most common clinical manifestations of critically ill neonates and often suggest serious diseases and complicated etiologies. The precise diagnosis of this disease can optimize the use of anti-seizure medication, reduce hospital costs, and improve the long-term neurodevelopmental outcomes. Currently, a few artificial intelligence-assisted diagnosis and treatment systems have been developed for neonatal seizures, but there is still a lack of high-level evidence for the diagnosis and treatment value in the real world. Based on an artificial intelligence-assisted diagnosis and treatment systems that has been developed for neonatal seizures, this study plans to recruit 370 neonates at a high risk of seizures from 6 neonatal intensive care units (NICUs) in China, in order to evaluate the effect of the system on the diagnosis, treatment, and prognosis of neonatal seizures in neonates with different gestational ages in the NICU. In this study, a diagnostic study protocol is used to evaluate the diagnostic value of the system, and a randomized parallel-controlled trial is designed to evaluate the effect of the system on the treatment and prognosis of neonates at a high risk of seizures. This multicenter prospective study will provide high-level evidence for the clinical application of artificial intelligence-assisted diagnosis and treatment systems for neonatal seizures in the real world.

Expert consensus on minimum technical standards for neonatal electroencephalography operation and report writing / 中国当代儿科杂志

Electroencephalography (EEG) monitoring is an important examination method in the management of critically ill neonates, which can be used to evaluate brain function and developmental status, severity of encephalopathy, and seizures and predict the long-term neurodevelopmental outcome of high-risk neonates with brain injury. EEG monitoring for neonates is different from that for adults and children, and its operation and interpretation are easily affected by the number of recording electrodes, electrode montage, and monitoring quality. Therefore, standard operation must be followed to ensure the quality of signal acquisition and correct interpretation, thereby ensuring proper management of critically ill neonates. The Subspecialty Group of Neonatology, Society of Pediatrics, Chinese Medical Association established an expert group composed of professionals in neonatology and brain electrophysiology to perform a literature review, summarize the minimum technical standards for neonatal EEG monitoring, and develop the expert consensus on minimum technical standards for neonatal EEG operation and report writing. This consensus will provide guidance for neonatal EEG operation, including technical parameters of EEG monitoring device, operation procedures of EEG monitoring, and specifications for report writing.

Cross subject personality assessment based on electroencephalogram functional connectivity and domain adaptation / 生物医学工程学杂志

The research shows that personality assessment can be achieved by regression model based on electroencephalogram (EEG). Most of existing researches use event-related potential or power spectral density for personality assessment, which can only represent the brain information of a single region. But some research shows that human cognition is more dependent on the interaction of brain regions. In addition, due to the distribution difference of EEG features among subjects, the trained regression model can not get accurate results of cross subject personality assessment. In order to solve the problem, this research proposes a personality assessment method based on EEG functional connectivity and domain adaption. This research collected EEG data from 45 normal people under different emotional pictures (positive, negative and neutral). Firstly, the coherence of 59 channels in 5 frequency bands was taken as the original feature set. Then the feature-based domain adaptation was used to map the feature to a new feature space. It can reduce the distribution difference between training and test set in the new feature space, so as to reduce the distribution difference between subjects. Finally, the support vector regression model was trained and tested based on the transformed feature set by leave-one-out cross-validation. What's more, this paper compared the methods used in previous researches. The results showed that the method proposed in this paper improved the performance of regression model and obtained better personality assessment results. This research provides a new method for personality assessment.

Neurovascular coupling analysis of working memory based on electroencephalography and functional near-infrared spectroscopy / 生物医学工程学杂志

Working memory is an important foundation for advanced cognitive function. The paper combines the spatiotemporal advantages of electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) to explore the neurovascular coupling mechanism of working memory. In the data analysis, the convolution matrix of time series of different trials in EEG data and hemodynamic response function (HRF) and the blood oxygen change matrix of fNIRS are extracted as the coupling characteristics. Then, canonical correlation analysis (CCA) is used to calculate the cross correlation between the two modal features. The results show that CCA algorithm can extract the similar change trend of related components between trials, and fNIRS activation of frontal pole region and dorsolateral prefrontal lobe are correlated with the delta, theta, and alpha rhythms of EEG data. This study reveals the mechanism of neurovascular coupling of working memory, and provides a new method for fusion of EEG data and fNIRS data.

Progresses and prospects on frequency recognition methods for steady-state visual evoked potential / 生物医学工程学杂志

Steady-state visual evoked potential (SSVEP) is one of the commonly used control signals in brain-computer interface (BCI) systems. The SSVEP-based BCI has the advantages of high information transmission rate and short training time, which has become an important branch of BCI research field. In this review paper, the main progress on frequency recognition algorithm for SSVEP in past five years are summarized from three aspects, i.e., unsupervised learning algorithms, supervised learning algorithms and deep learning algorithms. Finally, some frontier topics and potential directions are explored.

A spatial-temporal hybrid feature extraction method for rapid serial visual presentation of electroencephalogram signals / 生物医学工程学杂志

Rapid serial visual presentation-brain computer interface (RSVP-BCI) is the most popular technology in the early discover task based on human brain. This algorithm can obtain the rapid perception of the environment by human brain. Decoding brain state based on single-trial of multichannel electroencephalogram (EEG) recording remains a challenge due to the low signal-to-noise ratio (SNR) and nonstationary. To solve the problem of low classification accuracy of single-trial in RSVP-BCI, this paper presents a new feature extraction algorithm which uses principal component analysis (PCA) and common spatial pattern (CSP) algorithm separately in spatial domain and time domain, creating a spatial-temporal hybrid CSP-PCA (STHCP) algorithm. By maximizing the discrimination distance between target and non-target, the feature dimensionality was reduced effectively. The area under the curve (AUC) of STHCP algorithm is higher than that of the three benchmark algorithms (SWFP, CSP and PCA) by 17.9%, 22.2% and 29.2%, respectively. STHCP algorithm provides a new method for target detection.

Parameter transfer learning based on shallow visual geometry group network and its application in motor imagery classification / 生物医学工程学杂志

Transfer learning is provided with potential research value and application prospect in motor imagery electroencephalography (MI-EEG)-based brain-computer interface (BCI) rehabilitation system, and the source domain classification model and transfer strategy are the two important aspects that directly affect the performance and transfer efficiency of the target domain model. Therefore, we propose a parameter transfer learning method based on shallow visual geometry group network (PTL-sVGG). First, Pearson correlation coefficient is used to screen the subjects of the source domain, and the short-time Fourier transform is performed on the MI-EEG data of each selected subject to acquire the time-frequency spectrogram images (TFSI). Then, the architecture of VGG-16 is simplified and the block design is carried out, and the modified sVGG model is pre-trained with TFSI of source domain. Furthermore, a block-based frozen-fine-tuning transfer strategy is designed to quickly find and freeze the block with the greatest contribution to sVGG model, and the remaining blocks are fine-tuned by using TFSI of target subjects to obtain the target domain classification model. Extensive experiments are conducted based on public MI-EEG datasets, the average recognition rate and Kappa value of PTL-sVGG are 94.9% and 0.898, respectively. The results show that the subjects' optimization is beneficial to improve the model performance in source domain, and the block-based transfer strategy can enhance the transfer efficiency, realizing the rapid and effective transfer of model parameters across subjects on the datasets with different number of channels. It is beneficial to reduce the calibration time of BCI system, which promote the application of BCI technology in rehabilitation engineering.

Research on electroencephalogram power spectral density of stroke patients under transcranial direct current stimulation / 生物医学工程学杂志

Transcranial direct current stimulation (tDCS) has become a new method of post-stroke rehabilitation treatment and is gradually accepted by people. However, the neurophysiological mechanism of tDCS in the treatment of stroke still needs further study. In this study, we recruited 30 stroke patients with damage to the left side of the brain and randomly divided them into a real tDCS group (15 cases) and a sham tDCS group (15 cases). The resting EEG signals of the two groups of subjects before and after stimulation were collected, then the difference of power spectral density was analyzed and compared in the band of delta, theta, alpha and beta, and the delta/alpha power ratio (DAR) was calculated. The results showed that after real tDCS, delta band energy decreased significantly in the left temporal lobes, and the difference was statistically significant ( P < 0.05); alpha band energy enhanced significantly in the occipital lobes, and the difference was statistically significant ( P < 0.05); the difference of theta and beta band energy was not statistically significant in the whole brain region ( P > 0.05). Furthermore, the difference of delta, theta, alpha and beta band energy was not statistically significant after sham tDCS ( P > 0.05). On the other hand, the DAR value of stroke patients decreased significantly after real tDCS, and the difference was statistically significant ( P < 0.05), and there was no significant difference in sham tDCS ( P > 0.05). This study reveals to a certain extent the neurophysiological mechanism of tDCS in the treatment of stroke.

Motor imagery electroencephalogram classification based on sparse spatiotemporal decomposition and channel attention / 生物医学工程学杂志

Motor imagery electroencephalogram (EEG) signals are non-stationary time series with a low signal-to-noise ratio. Therefore, the single-channel EEG analysis method is difficult to effectively describe the interaction characteristics between multi-channel signals. This paper proposed a deep learning network model based on the multi-channel attention mechanism. First, we performed time-frequency sparse decomposition on the pre-processed data, which enhanced the difference of time-frequency characteristics of EEG signals. Then we used the attention module to map the data in time and space so that the model could make full use of the data characteristics of different channels of EEG signals. Finally, the improved time-convolution network (TCN) was used for feature fusion and classification. The BCI competition IV-2a data set was used to verify the proposed algorithm. The experimental results showed that the proposed algorithm could effectively improve the classification accuracy of motor imagination EEG signals, which achieved an average accuracy of 83.03% for 9 subjects. Compared with the existing methods, the classification accuracy of EEG signals was improved. With the enhanced difference features between different motor imagery EEG data, the proposed method is important for the study of improving classifier performance.

Research on the feature representation of motor imagery electroencephalogram signal based on individual adaptation / 生物医学工程学杂志

Aiming at the problem of low recognition accuracy of motor imagery electroencephalogram signal due to individual differences of subjects, an individual adaptive feature representation method of motor imagery electroencephalogram signal is proposed in this paper. Firstly, based on the individual differences and signal characteristics in different frequency bands, an adaptive channel selection method based on expansive relevant features with label F (ReliefF) was proposed. By extracting five time-frequency domain observation features of each frequency band signal, ReliefF algorithm was employed to evaluate the effectiveness of the frequency band signal in each channel, and then the corresponding signal channel was selected for each frequency band. Secondly, a feature representation method of common space pattern (CSP) based on fast correlation-based filter (FCBF) was proposed (CSP-FCBF). The features of electroencephalogram signal were extracted by CSP, and the best feature sets were obtained by using FCBF to optimize the features, so as to realize the effective state representation of motor imagery electroencephalogram signal. Finally, support vector machine (SVM) was adopted as a classifier to realize identification. Experimental results show that the proposed method in this research can effectively represent the states of motor imagery electroencephalogram signal, with an average identification accuracy of (83.0±5.5)% for four types of states, which is 6.6% higher than the traditional CSP feature representation method. The research results obtained in the feature representation of motor imagery electroencephalogram signal lay the foundation for the realization of adaptive electroencephalogram signal decoding and its application.

Localization of epileptogenic zone based on reconstruction of dynamical epileptic network and virtual resection / 生物医学工程学杂志

Drug-refractory epilepsy (DRE) may be treated by surgical intervention. Intracranial EEG has been widely used to localize the epileptogenic zone (EZ). Most studies of epileptic network focus on the features of EZ nodes, such as centrality and degrees. It is difficult to apply those features to the treatment of individual patients. In this study, we proposed a spatial neighbor expansion approach for EZ localization based on a neural computational model and epileptic network reconstruction. The virtual resection method was also used to validate the effectiveness of our approach. The electrocorticography (ECoG) data from 11 patients with DRE were analyzed in this study. Both interictal data and surgical resection regions were used. The results showed that the rate of consistency between the localized regions and the surgical resections in patients with good outcomes was higher than that in patients with poor outcomes. The average deviation distance of the localized region for patients with good outcomes and poor outcomes were 15 mm and 36 mm, respectively. Outcome prediction showed that the patients with poor outcomes could be improved when the brain regions localized by the proposed approach were treated. This study provides a quantitative analysis tool for patient-specific measures for potential surgical treatment of epilepsy.