Term Memory Research Articles

Pharmacological intervention of behavioural traits and brain histopathology of prenatal valproic acid-induced mouse model of autism.

Autism spectrum disorder (ASD) is one of the leading causes of distorted social communication, impaired speech, hyperactivity, anxiety, and stereotyped repetitive behaviour. The aetiology of ASD is complex; therefore, multiple drugs have been suggested to manage the symptoms. Studies with histamine H3 receptor (H3R) blockers and acetylcholinesterase (AchE) blockers are considered potential therapeutic agents for the management of various cognitive impairments. Therefore, the aim of this study was to evaluate the neuro-behavioural effects of Betahistine, an H3R antagonist, and Donepezil, an acetylcholinesterase inhibitor on Swiss albino mouse model of autism. The mice were intraperitoneally injected with valproic acid (VPA) on the embryonic 12.5th day to induce autism-like symptoms in their offspring. This induced autism-like symptoms persists throughout the life. After administration of different experimental doses, various locomotor tests: Open Field, Hole-Board, Hole Cross and behavioural tests by Y-Maze Spontaneous Alternation and histopathology of brain were performed and compared with the control and negative control (NC1) groups of mice. The behavioural Y-Maze test exhibits significant improvement (p <0.01) on the short term memory of the test subjects upon administration of lower dose of Betahistine along with MAO-B inhibitor Rasagiline once compared with the NC1 group (VPA-exposed mice). Furthermore, the tests showed significant reduction in locomotion in line crossing (p <0.05), rearing (p <0.001) of the Open Field Test, and the Hole Cross Test (p <0.01) with administration of higher dose of Betahistine. Both of these effects were observed upon administration of acetylcholinesterase inhibitor, Donepezil. Brain-histopathology showed lower neuronal loss and degeneration in the treated groups of mice in comparison with the NC1 VPA-exposed mice. Administration of Betahistine and Rasagiline ameliorates symptoms like memory deficit and hyperactivity, proving their therapeutic effects. The effects found are dose dependent. The findings suggest that H3R might be a viable target for the treatment of ASD.

A mathematical formulation of learner cognition for personalised learning experiences

The paper focuses on the assessment of cognitive skills within Virtual Learning Environments (VLEs). In response to the global shift to remote learning amid the COVID-19 pandemic, VLEs, which include learning management systems (LMS) and online collaboration platforms, gained prominence. The proposed work leverages an established Cattell–Horn–Carroll (CHC) theory to propose eight metrics, which collectively form a part of Cognitive Evaluation Metrics (CEM). The proposed metrics introduce a novel computational approach for multimode evaluation of learners’ cognitive abilities for each learning task within a learning environment. The paper details the formalism for the evaluation of the metrics and makes a contribution towards the potential of the proposed methodology to evaluate cognitive abilities. Additionally, the work implements CEM integration into the learner module of a Game-Based Learning (GBL) environment. Analysis of simulations in the GBL environment, along with statistical analysis, provides insights into the normal distribution of cognitive metrics. This reveals diverse ranges in various abilities such as long or short term memory, working memory, reasoning, attention, and processing speed. The paper also explores the impact of virtual assistants, which highlights their limited relevance to enhance cognitive abilities but serve as valuable on-demand support resources.

Network Intrusion Detection based on LSTM

Network intrusion detection, as an important means of ensuring daily network security, its accuracy and response speed are crucial for defending against network attacks. This article explores and implements deep learning based network intrusion detection techniques, particularly the application of Long Short Term Memory (LSTM) networks in detecting network intrusion behavior. The aim is to solve the problems of gradient vanishing and exploding in traditional RNNs, improve the emergency response capability of network systems, and enhance the reliability and security of networks. The study used the KDD99 dataset to demonstrate the effectiveness of the LSTM model in network intrusion detection. The experimental results show that the constructed LSTM model achieves an accuracy of 88.01% in network intrusion detection tasks, demonstrating high accuracy and feasibility.

Data-Driven Modeling of Tire-Soil Interaction with POD-Based Model Order Reduction

Abstract A data-driven model capable of predicting time-domain solutions of a high-fidelity tire-soil interaction model is developed to enable quick prediction of mobility capabilities on deformable terrain. The adaptive model order reduction based on the proper orthogonal decomposition (POD), for which the high-dimensional equations are projected onto the reduced subspace, is utilized as the basis for predicting the time-domain tire-soil interaction behavior. The projection-based model order reduction, however, requires many online matrix operations due to the successive updates of the nonlinear functions and Jacobians at every time step, thereby hindering the computational improvement. Therefore, a data-driven approach using a Long Short Term Memory (LSTM) neural network is introduced to predict the reduced order coordinates without the projection and time integration processes for computational speedup. With this model, a hybrid data-driven/physics-based off-road mobility model is proposed, where four separate LSTM-POD data-driven tire-soil interaction models are integrated into the physics-based multibody dynamics (MBD) vehicle model through a force-displacement coupling algorithm. By doing so, the individual data-driven tire-soil interaction model can be constructed efficiently, and the MBD and LSTM models are assembled as a single off-road mobility model and analyzed with existing off-road mobility solvers. The predictive ability and computational benefit of the proposed data-driven tire-soil interaction model with the POD-based model order reduction are examined with several numerical examples.

USE OF MACHINE LEARNING AND DEEP LEARNING METHODS IN HOUSING PRICE INDEX ESTIMATION: AN ANALYSIS ON ANKARA AND ISTANBUL

Factors such as supply chain difficulties, rising energy and oil prices, economic recession and production loss due to the pandemic have increased costs and inflation. All these factors have also seriously affected the construction sector. This study aims to create a deep learning and machine learning focused forecasting system based on Istanbul and Ankara monthly housing price index data for the period of January 2010 to June 2023. The system was created using approximately 13 years of housing interest rates, Consumer Price Index, XGMYO, Monthly Average Dollar and XAU data as the basis of the Istanbul and Ankara Housing Price Index forecasting process. During the research process, different RNN structures (Long and Short Term Memory, Gated Recurrent Unit) and machine learning (Random Forest) structures were tested and the effectiveness of these structures in housing price index forecasting was compared. The performances of the models were evaluated using RMSE, MSE, MAE, MAPE and R2 statistics. According to the results obtained, the method that gave the best performance for both provinces is the RF model. This is followed by LSTM and GRU models, respectively

Predicting ICU Readmission from Electronic Health Records via BERTopic with Long Short Term Memory Network Approach.

Background: The increasing rate of intensive care unit (ICU) readmissions poses significant challenges in healthcare, impacting both costs and patient outcomes. Predicting patient readmission after discharge is crucial for improving medical quality and reducing expenses. Traditional analyses of electronic health record (EHR) data have primarily focused on numerical data, often neglecting valuable text data. Methods: This study employs a hybrid model combining BERTopic and Long Short-Term Memory (LSTM) networks to predict ICU readmissions. Leveraging the MIMIC-III database, we utilize both quantitative and text data to enhance predictive capabilities. Our approach integrates the strengths of unsupervised topic modeling with supervised deep learning, extracting potential topics from patient records and transforming discharge summaries into topic vectors for more interpretable and personalized predictions. Results: Utilizing a comprehensive dataset of 36,232 ICU patient records, our model achieved an AUROC score of 0.80, thereby surpassing the performance of traditional machine learning models. The implementation of BERTopic facilitated effective utilization of unstructured data, generating themes that effectively guide the selection of relevant predictive factors for patient readmission prognosis. This significantly enhanced the model's interpretative accuracy and predictive capability. Additionally, the integration of importance ranking methods into our machine learning framework allowed for an in-depth analysis of the significance of various variables. This approach provided crucial insights into how different input variables interact and impact predictions of patient readmission across various clinical contexts. Conclusions: The practical application of BERTopic technology in our hybrid model contributes to more efficient patient management and serves as a valuable tool for developing tailored treatment strategies and resource optimization. This study highlights the significance of integrating unstructured text data with traditional quantitative data to develop more accurate and interpretable predictive models in healthcare, emphasizing the importance of individualized care and cost-effective healthcare paradigms.

Chicken swarm algorithm with deep convolutional neural network based tongue image analysis for gastric cancer classification

Tongue images (the size, shape, and colour of tongue and the thickness, colour, and moisture content of tongue coating), reflecting the medical condition of entire body based on the model of traditional Chinese medicine (TCM) are extremely utilized in China for millions of years. Gastric cancer (GC) is great lethal kind of cancer in countries and societies. The screening and analysis of GC yet depend on gastroscopy, however its application was significantly restricted due to its invasive, maximum rate and the requirement for expert endoscopists. Early recognition in GC patients and direct treatment contribute significantly to safety for health. Consequently, this study introduces a Chicken Swarm Algorithm with Deep learningbased Tongue Image Analysis for Gastric Cancer Classification (CSADL-TIAGCC) system. The projected CSADL-TIAGCC approach studies the input tongue images for the identification and classification of GC. To accomplish this, the CSADL-TIAGCC system uses improved U-Net segmentation approach. Besides, residual network (ResNet-34) model-based feature extractor is used. Furthermore, long short term memory (LSTM) approach was exploited for GC classification and its hyperparameters are selected by the CSA. The simulation outcome of the CSADL-TIAGCC algorithm was examined under tongue image database. The experimental outcomes illustrate the enhanced results of the CSADL-TIAGCC technique with respect of different evaluation measures.

Bi-directional Long Short-Term Memory with Bird Mating Optimizer based Spectrum Sensing Technique for Cognitive Radio Networks

Cognitive radio networks (CRN) enable the wireless devices to sense the radio spectrum, determine the frequency state channels, and reconfigure the communication variables for satisfying the QoS needs by reducing the energy utilization. In cognitive radio, the detection of principal user signals is crucial for secondary users in order to make the best use of available spectrum (CR). The problem with conventional spectrum sensing approaches is that they have a high rate of missed detections and false alarms, which makes it difficult to make effective use of the spectrum. In instruction to recover the correctness of the detection of free spectrum, deep learning-based spectrum sensing is employed. For resolving the drawbacks of traditional energy detection models, this paper presents a new spectral sensing technique for cognitive radio networks (SST-CRN). Recently published research in spectrum sensing has placed a high value on deep learning that is model-agnostic as a result of this. In particular, long-short term memory (LSTM) networks perform exceptionally well at extracting spatial and temporal information from input data in deep learning. The proposed model, Bidirectional Long Short-Term Memory (Bi-LSTM) with Bird Mating Optimization (BMO), makes it possible to create nonlinear threshold-based systems more quickly and easily than previously possible. The proposed Bi-LSTM with BMO technique involves two stages of operations namely offline and online. The offline stage creates the non-linear threshold value to detect energy. In addition, the online stage automated selects a decision function saved in offline stage for determining the existence of primary user. The experiments were carried out and the results were analyzed using the help of the RadioML2016.10b dataset. A greater spectrum detection performance over previous sensing models has been obtained using a combination of the B-LSTM model and the BMO model, it has been discovered.

Self-initiating stimulus in Short Term Memory: differential effects of temporal prediction and motor control

Self-initiating stimulus in Short Term Memory: differential effects of temporal prediction and motor control

Working memory capacity relates to reduced negative emotion in daily life

ABSTRACT Working memory capacity (WMC) refers to the ability to maintain information in short–term memory while attending to the immediate environment, and has been associated with emotional states. Yet, research on the link between WMC and emotion in naturalistic settings is growing and inconsistencies have been observed. In the current study (N = 109), we directly replicated the procedures of a prior experience sampling study (Garrison & Schmeichel, 2022), which found that higher WMC attenuates the relationship between stressful events in daily life and negative affect. We measured WMC in the laboratory and then measured the occurrence of stressful events, momentary emotional states, and coping responses to stress several times a day for six days. Higher WMC was associated with reduced momentary negative emotion, but this relationship did not depend on the occurrence of a stressful event. Exploratory analyses found that higher WMC was associated with a greater likelihood of planning as a coping response to stress and greater number of coping strategies reported per stressful event. However, coping did not mediate the link between WMC and momentary negative emotion. Our results contribute to the robustness and ecological validity of the link between WMC and reduced negative emotion in daily life.

Large-scale data-driven uniformity analysis and sensory prediction of commercial banana ripening process

Artificial intelligence (AI) and machine learning (ML) have found prominent yet mostly academic applications in the food supply chain specifically to preserve and optimize the quality of fresh produce and achieve uniformity across the various stages of the cold chain. Nevertheless, the practical use of AI/ML for predictive analytics within real and large-scale commercial food processes, such as banana ripening, is sparse. This study proposes a novel data-driven approach tested and validated on two new large-scale datasets to automate and optimize the banana ripening process in refrigerated marine containers by successfully employing ML in uniformity analysis of the peel color and the pulp temperature of bananas based on atmospheric conditions. The results demonstrate high correlations between the gas concentrations and the uniformity of the process, suggesting that the uniformity of the peel color and the pulp temperature of fruit can be achieved by controlling the concentrations of the CO2 and O2 gas levels. Furthermore, this study, for the first time, achieves accurate algorithmic predictions of oxygen levels from other atmospheric variables to provide an alternative approach for continuous, improved and more cost-effective monitoring of the atmospheric conditions during ripening. A wide-range of predictive models are tested and validated where the Long Short Term Memory regression provides the lowest root-mean-square-errors (0.033 and 0.202) with robust R-squared values (0.999 and 0.959) for two datasets.

A - 117 Double Trouble: a Case Study on Cardiac Amyloidosis and Alzheimer’s Disease

Abstract Objective Transthyretin amyloidosis (ATTR) is an increasingly recognized cause of heart failure amongst older adults. The median age upon diagnosis is 75 with 90% of patients reported as men. The amyloidogenic protein is transthyretin (TTR), which is primarily produced in the liver, but is also produced in the choroid plexus and neurons under distress. There are two types of ATTR: an autosomal dominant and a wild-type (wt) variant. ATTRwt is the amyloid precursor in the cardiac depositions found in older adults. We present the first neuropsychological evaluation of a patient with ATTRwt. Method An 87-year-old right-handed White woman reported a 2 year history of slowly progressive short term memory decline. Her family noted rapid cognitive decline and paranoia approximately 6 months ago. Medical history was notable for heart failure, along with lumbar stenosis, carpal tunnel syndrome, arthritis, scoliosis, and hyperlipidemia. ATTRwt diagnosis was made during cardiac clearance for lumbar surgery. Neuroimaging documented age-related cerebral atrophy and chronic microvascular ischemic disease. Results Inspection of data revealed an amnestic presentation, with impaired confrontational naming, spatial judgment, and processing speed performances. She denied depressogenic and anxiolytic symptoms while family endorsed symptoms of apathy and executive dysfunction. Conclusions The relationship between cardiac and nervous systems has garnished increasing attention in the literature and in clinical planning. Overall, the relationship between cardiac TTR amyloid and brain amyloid beta deposition in humans is unclear. Neuropsychologists should be aware of this heart condition that is increasingly recognized in older adults and the intriguing findings implicated around the cardiac-brain axis.

A Deep Learning-Enhanced Compartmental Model and Its Application in Modeling Omicron in China.

The mainstream compartmental models require stochastic parameterization to estimate the transmission parameters between compartments, whose calculation depend upon detailed statistics on epidemiological characteristics, which are expensive, economically and resource-wise, to collect. In addition, infectious diseases spread in three dimensions: temporal, spatial, and mobile, i.e., they affect a population through not only the time progression of infection, but also the geographic distribution and physical mobility of the population. However, the parameterization process for the mainstream compartmental models does not effectively capture the spatial and mobile dimensions. As an alternative, deep learning techniques are utilized in estimating these stochastic parameters with greatly reduced dependency on data particularity and with a built-in temporal-spatial-mobile process that models the geographic distribution and physical mobility of the population. In particular, we apply DNN (Deep Neural Network) and LSTM (Long-Short Term Memory) techniques to estimate the transmission parameters in a customized compartmental model, then feed the estimated transmission parameters to the compartmental model to predict the development of the Omicron epidemic in China over the 28 days for the period between 4 June and 1 July 2022. The average levels of predication accuracy of the model are 98% and 92% for the number of infections and deaths, respectively. We establish that deep learning techniques provide an alternative to the prevalent compartmental modes and demonstrate the efficacy and potential of applying deep learning methodologies in predicting the dynamics of infectious diseases.

I see artifacts: ICA-based EEG artifact removal does not improve deep network decoding across three BCI tasks

Abstract \textit{Objective.}&amp;#xD;In this paper, we conduct a detailed investigation on the effect of IC-based noise rejection methods in neural network classifier-based decoding of electroencephalography (EEG) data in different task datasets.&amp;#xD;\textit{Approach.}&amp;#xD;We apply a pipeline matrix of two popular different Independent Component (IC) decomposition methods (Infomax, AMICA) with three different component rejection strategies (none, ICLabel, and MARA) on three different EEG datasets (Motor imagery, long-term memory formation, and visual memory). We cross-validate processed data from each pipeline with three architectures commonly used for EEG classification (two convolutional neural networks (CNN) and one long short term memory (LSTM) based model. We compare decoding performances on within-participant and within-dataset levels. &amp;#xD;\textit{Main Results.}&amp;#xD;Our results show that the benefit from using IC-based noise rejection for decoding analyses is at best minor, as component-rejected data did not show consistently better performance than data without rejections---especially given the significant computational resources required for ICA computations.&amp;#xD;\textit{Significance.}&amp;#xD;With ever growing emphasis on transparency and reproducibility, as well as the obvious benefits arising from streamlined processing of large-scale datasets, there has been an increased interest in automated methods for pre-processing EEG data. One prominent part of such pre-processing pipelines consists of identifying and potentially removing artifacts arising from extraneous sources. This is typically done via Independent Component (IC) based correction for which numerous methods have been proposed, differing not only in the decomposition of the raw data into ICs, but also in how they reject the computed ICs. While the benefits of these methods are well established in univariate statistical analyses, it is unclear whether they help in multivariate scenarios, and specifically in neural network based decoding studies. As computational costs for pre-processing large-scale datasets are considerable, it is important to consider whether the tradeoff between model performance and available resources is worth the effort.

Value Relevance of Key Accounting Information in Predicting Market Performance: A Machine Learning Approach

This study aims to examine the value relevance of key accounting information in predicting market performance and to explore which accounting information has the most value relevance in predicting market performance. The dataset is comprised of a total of 1401 observations from 117 companies over a period of 2010 to 2021. To conduct the study, total 12 indicators were taken from financial statements to represent the major accounting information (AI) and year-end market price and Tobin’s Q were taken as a proxy for market performance. Under the principal component analysis (PCA), it was found that only 3 AIs namely earnings per share (EPS), book value per share (BVPS) and cash flow per share (CFPS) could contribute to the whole explained variance ratio of market performance. The study employs several machine learning approaches: random forest regression (RFR) was chosen as the base line model and the result was compared with decision tree regression (DTR), long short term memory (LSTM), neural network model and multivariate regression model. After splitting the total observation into a 70:30 training-testing dataset and controlling for noise reduction, it was found that over any other models, LSTM and RFR models can predict the market performance with higher R-square value of 62% and 65% respectively along with the lowest MSE of all other models. It was found that EPS has the highest value relevance (factor importance) in predicting market profitability whereas BVPS and CFPS were found to have less than 10% factor importance in predicting market profitability.

Classification of tomato seedling chilling injury based on chlorophyll fluorescence imaging and DBO-BiLSTM

IntroductionTomatoes are sensitive to low temperatures during their growth process, and low temperatures are one of the main environmental limitations affecting plant growth and development in Northeast China. Chlorophyll fluorescence imaging technology is a powerful tool for evaluating the efficiency of plant photosynthesis, which can detect and reflect the effects that plants are subjected to during the low temperature stress stage, including early chilling injury.MethodsThis article primarily utilizes the chlorophyll fluorescence image set of tomato seedlings, applying the dung beetle optimization (DBO) algorithm to enhance the deep learning bidirectional long short term memory (BiLSTM) model, thereby improving the accuracy of classification prediction for chilling injury in tomatoes. Firstly, the proportion of tomato chilling injury areas in chlorophyll fluorescence images was calculated using a threshold segmentation algorithm to classify tomato cold damage into four categories. Then, the features of each type of cold damage image were filtered using SRCC to extract the data with the highest correlation with cold damage. These data served as the training and testing sample set for the BiLSTM model. Finally, DBO algorithm was applied to enhance the deep learning BiLSTM model, and the DBO-BiLSTM model was proposed to improve the prediction performance of tomato seedling category labels.ResultsThe results showed that the DBO-BiLSTM model optimized by DBO achieved an accuracy, precision, recall, and F1 score with an average of over 95%.DiscussionCompared to the original BiLSTM model, these evaluation parameters improved by 9.09%, 7.02%, 9.16%, and 8.68%, respectively. When compared to the commonly used SVM classification model, the evaluation parameters showed an increase of 6.35%, 7.33%, 6.33%, and 6.5%, respectively. This study was expected to detect early chilling injury through chlorophyll fluorescence imaging, achieve automatic classification and labeling of cold damage data, and lay a research foundation for in-depth research on the cold damage resistance of plants themselves and exploring the application of deep learning classification methods in precision agriculture.

SPIDOL study protocol for the assessment of intrathecal ziconotide antalgic efficacy for severe refractory neuropathic pain due to spinal cord lesions

RationaleCentral neuropathic pain resulting from spinal cord injury is notoriously debilitating and difficult to treat with few currently available treatments. A novel molecule with intrathecal administration: Ziconotide has been approved for treatment of refractory neuropathic pain in general. It acts as a presynaptic calcium channel blocker. A pilot study has shown its potential in SCI neuropathic pain patients.ObjectiveThe aim of this study is to determine the long-term (6 months) efficacy of chronic intrathecal ziconotide for the treatment of neuropathic SCI pain.Study designMulticenter, Randomized, Comparative, Placebo controlled, Double blind clinical trial, with a crossover of random alternated periods of 6 months (placebo or ITZ) for a total of 15 months including a total of 44 patients.Study population• Patients with SCI of various etiologies exhibiting neuropathic pain refractory to non-invasive treatments.• > 18 years.InterventionIntrathecal administration of ziconotide via an implanted pump.Study outcomesPrimary study outcomeDifference in pain intensity for all patients between effective treatment and placebo periods.Secondary study outcomes1. Continuous evaluation of pain intensity.2. Percentage of patients with at least 30% of pain reduction.3. Satisfaction level of the patient pain relief.4. Declarations of serious adverse events.5. Duration and intensity of spontaneous and provoked pain.6. Quality of life.7. Patient global impression of change.8. Quantification of daily dosages of analgesic drug intake.9. Long term memory and neurocognitive effects.10. Assessment of the patient’s physical and emotional distress.Nature and extent of the burden and risks associated with participation, benefit, and group relatednessParticipation in this study is in accordance with current treatment protocols for SCI neuropathic pain in France therefore it proposes a treatment that would currently be considered regular practice even though no RCT evidence is yet available. The study gives patients the advantage of directly testing versus placebo a treatment that otherwise entails significant constraints.A Data Safety Monitoring board (DSMB) will be created for continuous safety analysis. Furthermore, patients will be followed in specialized pain centers offering the possibility of continuing their treatment after the study period.

Enhancing Stock Price Prediction in the Indonesian Market: a Concave LSTM Approach with RunReLU

This study addresses the pressing need for improved stock price prediction models in the financial markets, focusing on the Indonesian stock market. It introduces an innovative approach that utilizes the custom activation function RunReLU within a Concave Long Short Term Memory (LSTM) framework. The primary objective is to enhance prediction accuracy, ultimately assisting investors and market participants in making more informed decisions. The research methodology used historical stock price data from ten prominent companies listed on the Indonesia Stock Exchange, covering the period from July 6, 2015, to October 14, 2021. Evaluation metrics such as RMSE, MAE, MAPE, and R2 were employed to assess model performance. The results consistently favored the RunReLU-based model over the ReLU-based model, showcasing lower RMSE and MAE values, higher R2 values, and notably reduced MAPE values. These findings underscore the practical applicability of custom activation functions for financial time series data, providing valuable tools for enhancing prediction precision in the dynamic landscape of the Indonesian stock market.

SleepGCN: A transition rule learning model based on Graph Convolutional Network for sleep staging

Background and Objective:Automatic sleep staging is essential for assessing and diagnosing sleep disorders, serving millions of people who suffer from them. Numerous sleep staging models have been proposed recently, but most of them have not fully explored the sleep transition rules that are essential for sleep experts to identify sleep stages. Therefore, one objective of this paper is to develop an automatic sleep staging model to capture the transition rules between sleep stages. Methods:In this paper, we propose a novel sleep staging model named SleepGCN. It utilizes the deep features of electroencephalogram (EEG) and electrooculogram (EOG) signals extracted by the sleep representation learning (SRL) module, in conjunction with the transition rules learned by the sleep transition rule learning (STRL) module to identify sleep stages. Specifically, the SRL module utilizes the residual network (ResNet) and Long Short Term Memory (LSTM) structure to capture the deep time-invariant features and temporal information of each sleep stage from the two-channel EEG-EOG, and then applies a feature enhancement block to obtain the refined features. The STRL module employs a Graph Convolutional Network (GCN) and a transition rule matrix to capture transition rules between sleep stages based on the sequence labels of the input signals. Results:We evaluate SleepGCN on five public datasets: SleepEDF-20, SleepEDF-78, SHHS, DOD-H and DOD-O. Overall, SleepGCN achieves an accuracy of 89.70%, 87.70%, 86.16%, 82.07%, and 81.20%, alongside a macro-average F1-score of 85.20%, 82.70%, 77.69%, 72.44%, and 72.93% across these datasets, respectively. Conclusions:The results achieved by our proposed model are much better than those of all other compared models. The ablation study validates the contributions of the SRL and STRL modules proposed in SleepGCN to the sleep staging tasks. Additionally, it shows that the sleep staging model using two-channel EEG-EOG outperforms the model using single-channel EEG or EOG. Overall, SleepGCN is an effective solution for sleep staging using two-channel EEG-EOG.

A Comprehensive Study on Implementation of Deep Learning on Autonomous Vehicle for Steering Angle Prediction and Stability

The development of autonomous vehicles has recently enhanced the transportation industry and opened up a variety of opportunities and problems that can be solved with the aid of current methods and technology. In this study, three separate algorithms were used to predict the steering angle with a track image: Artificial Neural Network (ANN), Convolution Neural Network (CNN), and a combination of CNN and Long-Short Term Memory (CNN-LSTM). The PID controller was employed for benchmarking, which takes Cross-Track Error (CTE) provided by the simulation to steer the vehicle. To achieve improved performance, the standard NVIDIA CNN self-driving model was slightly altered by feeding it with sequential frames. The comparison analysis was conducted using the OpenAI Gym Donkey Simulator.