Accurate Decoding Research Articles

Studying the function of tRNA modifications: experimental challenges and opportunities.

tRNAs are essential molecules in protein synthesis, responsible for translating the four-nucleotide genetic code into the corresponding amino acid sequence. RNA modifications play a crucial role in influencing tRNA folding, structure, and function. These modifications, ranging from simple methylations to complex hypermodified species, are distributed throughout the tRNA molecule. Depending on their type and position, they contribute to the accuracy and efficiency of decoding by participating in a complex network of interactions. The enzymatic processes introducing these modifications are equally intricate and diverse, adding further complexity. As a result, studying tRNA modifications faces limitations at multiple levels. This review addresses the challenges involved in manipulating and studying the function of tRNA modifications and discusses experimental strategies and possibilities to overcome these obstacles.

Acoustic Exaggeration Enhances Speech Discrimination in Young Autistic Children.

Child-directed speech (CDS), which amplifies acoustic and social features of speech during interactions with young children, promotes typical phonetic and language development. In autism, both behavioral and brain data indicate reduced sensitivity to human speech, which predicts absent, decreased, or atypical benefits of exaggerated speech signals such as CDS. This study investigates the impact of exaggerated fundamental frequency (F0) and voice-onset time on the neural processing of speech sounds in 22 Chinese-speaking autistic children aged 2-7 years old with a history of speech delays, compared with 25 typically developing (TD) peers. Electroencephalography (EEG) data were collected during passive listening to exaggerated and non-exaggerated syllables. A time-resolved multivariate pattern analysis (MVPA) was used to evaluate the potential effects of acoustic exaggeration on syllable discrimination in terms of neural decoding accuracy. For non-exaggerated syllables, neither the autism nor the TD group achieved above-chance decoding accuracy. In contrast, for exaggerated syllables, both groups achieved above-chance decoding, indicating significant syllable discrimination, with no difference in accuracy between the autism and TD groups. However, the temporal generalization patterns in the MVPA results revealed distinct neural mechanisms supporting syllable discrimination between the groups. Although the TD group demonstrated a left-hemisphere advantage for decoding and generalization, the autism group displayed similar decoding patterns between hemispheres. These findings highlight the potential of selective acoustic exaggeration to support speech learning in autistic children, underscoring the importance of tailored, sensory-based interventions.

Edge-Aware Dual-Task Image Watermarking Against Social Network Noise

In the era of widespread digital image sharing on social media platforms, deep-learning-based watermarking has shown great potential in copyright protection. To address the fundamental trade-off between the visual quality of the watermarked image and the robustness of watermark extraction, we explore the role of structural features and propose a novel edge-aware watermarking framework. Our primary innovation lies in the edge-aware secret hiding module (EASHM), which achieves adaptive watermark embedding by aligning watermarks with image structural features. To realize this, the EASHM leverages knowledge distillation from an edge detection teacher and employs a dual-task encoder that simultaneously performs edge detection and watermark embedding through maximal parameter sharing. The framework is further equipped with a social network noise simulator (SNNS) and a secret recovery module (SRM) to enhance robustness against common image noise attacks. Extensive experiments on three public datasets demonstrate that our framework achieves superior watermark imperceptibility, with PSNR and SSIM values exceeding 40.82 dB and 0.9867, respectively, while maintaining an over 99% decoding accuracy under various noise attacks, outperforming existing methods by significant margins.

Motion Cognitive Decoding of Cross-Subject Motor Imagery Guided on Different Visual Stimulus Materials.

Motor imagery (MI) plays an important role in brain-computer interfaces, especially in evoking event-related desynchronization and synchronization (ERD/S) rhythms in electroencephalogram (EEG) signals. However, the procedure for performing a MI task for a single subject is subjective, making it difficult to determine the actual situation of an individual's MI task and resulting in significant individual EEG response variations during motion cognitive decoding. To explore this issue, we designed three visual stimuli (arrow, human, and robot), each of which was used to present three MI tasks (left arm, right arm, and feet), and evaluated differences in brain response in terms of ERD/S rhythms. To compare subject-specific variations of different visual stimuli, a novel cross-subject MI-EEG classification method was proposed for the three visual stimuli. The proposed method employed a covariance matrix centroid alignment for preprocessing of EEG samples, followed by a model agnostic meta-learning method for cross-subject MI-EEG classification. The experimental results showed that robot stimulus materials were better than arrow or human stimulus materials, with an optimal cross-subject motion cognitive decoding accuracy of 79.04%. Moreover, the proposed method produced robust classification of cross-subject MI-EEG signal decoding, showing superior results to conventional methods on collected EEG signals.

Visible light visual indoor positioning system for based on residual convolutional networks and image restoration

Abstract Visible light positioning technology, with its advantages of low cost, strong anti-interference, and high precision, is widely researched and applied in various different scenarios. In this paper, for the complexity of indoor environments, considering the problem of occlusion by various obstacles that may exist in indoor spaces, which may lead to incomplete imaging of CMOS image sensors, a maximum gray value-based occlusion recovery and decoding scheme is proposed. This scheme effectively solves the problem of visible light transmission channel being blocked and accomplishes LED-ID decoding. In addition, the overflow effect due to uneven light irradiation gathered in each pixel row affects the accuracy of decoding LED-ID, which in turn leads to poor positioning accuracy. In this paper, we propose to use an adaptive gamma correction method to eliminate the influence of the overflow effect and to improve the accuracy of decoding. In order to improve the positioning accuracy, a visible light positioning algorithm based on residual convolutional network (VisiResNet) is proposed to achieve high accuracy positioning. The experimental results show that the average positioning error is 9.7 cm in a space of 9m× 3m× 3m, and a decoding accuracy of 90% within 1.4m is achieved in the face of different occlusion situations. The system can achieve centimeter-level positioning accuracy and meet the indoor positioning requirements.

Research on Enhanced Belief Propagation List Decoding Algorithm for Polar Codes in UAV Communications for 6G

The introduction of sixth-generation mobile communication technology (6G) poses new requirements for the capacity, rate, latency, and reliability of communication systems. As a vital component of 6G technology, unmanned aerial vehicle (UAV) communications also face various challenges, such as noise interference and limited hardware resources. To meet the high demands of 6G, advanced channel coding techniques need to be adopted. Polar codes, due to their theoretically achievable Shannon limit performance, have potential applications in UAV communication systems. Constructing reliable polar decoding schemes is currently a research hotspot in the field of communications. The Belief Propagation List (BPL) decoding algorithm for polar codes can effectively enhance the accuracy of polar code BP decoding. However, existing BPL decoding algorithms for polar codes face issues such as high hardware resource consumption and unsatisfactory decoding accuracy. Addressing the aforementioned issues, this paper proposes a BPL decoding algorithm for polar codes based on information geometry. An information geometry framework is constructed, where the soft information output by the BP decoder is treated as points on a statistical manifold, and their geometric properties are calculated. By introducing the concept of the soft information centroid and a path selection criterion based on the soft information centroid, combined with geometric distance as a weight, the decoding performance is improved, and hardware overhead is reduced. Simulation results show that under the conditions of a maximum of 60 iterations and 5 decoders, the proposed algorithm reduces the bit error rate by 16.2–74.9% compared to the classic BPL algorithm, providing strong technical support for the application of polar codes in scenarios such as UAV communications.

Simultaneous encoding of speed, distance, and direction in discrete reaching: an EEG study

Objective.The complicated processes of carrying out a hand reach are still far from fully understood. In order to further the understanding of the kinematics of hand movement, the simultaneous representation of speed, distance, and direction in the brain is explored.Approach.We utilized electroencephalography (EEG) signals and hand position recorded during a four-direction center-out reaching task with either quick or slow speed, near and far distance. Linear models were employed in two modes: decoding and encoding. First, to test the discriminability of speed, distance, and direction. Second, to find the contribution of the cortical sources via the source localization. Additionally, we compared the decoding accuracy when using features obtained from EEG signals and source-localized EEG signals based on the results from the encoding model.Main results.Speed, distance, and direction can be classified better than chance. The accuracy of the speed was also higher than the distance, indicating a stronger representation of the speed than the distance. The speed and distance showed similar significant sources in the central regions related to the movement initiation, while the direction indicated significant sources in the parieto-occipital regions related to the movement preparation. The combination of the features from EEG and source localized signals improved the classification.Significance.Directional and non-directional information are represented in two separate networks. The quick movement resulted in improvement in the direction classification. Our results enhance our understanding of hand movement in the brain and help us make informed decisions when designing an improved paradigm in the future.

Unraveling the Differential Efficiency of Dorsal and Ventral Pathways in Visual Semantic Decoding.

Visual semantic decoding aims to extract perceived semantic information from the visual responses of the human brain and convert it into interpretable semantic labels. Although significant progress has been made in semantic decoding across individual visual cortices, studies on the semantic decoding of the ventral and dorsal cortical visual pathways remain limited. This study proposed a graph neural network (GNN)-based semantic decoding model on a natural scene dataset (NSD) to investigate the decoding differences between the dorsal and ventral pathways in process various parts of speech, including verbs, nouns, and adjectives. Our results indicate that the decoding accuracies for verbs and nouns with motion attributes were significantly higher for the dorsal pathway as compared to those for the ventral pathway. Comparative analyses reveal that the dorsal pathway significantly outperformed the ventral pathway in terms of decoding performance for verbs and nouns with motion attributes, with evidence showing that this superiority largely stemmed from higher-level visual cortices rather than lower-level ones. Furthermore, these two pathways appear to converge in their heightened sensitivity toward semantic content related to actions. These findings reveal unique visual neural mechanisms through which the dorsal and ventral cortical pathways segregate and converge when processing stimuli with different semantic categories.

An adaptive session-incremental broad learning system for continuous motor imagery EEG classification.

Motor imagery electroencephalography (MI-EEG) is usually used as a driving signal in neuro-rehabilitation systems, and its feature space varies with the recovery progress. It is required to endow the recognition model with continuous learning and self-updating capability. Broad learning system (BLS) can be remodeled in an efficient incremental learning way. However, its architecture is intractable to change automatically to adapt to new incoming MI-EEG with time-varying and complex temporal-spatial characteristics. In this paper, an adaptive session-incremental BLS (ASiBLS) is proposed based on mutual information theory and BLS. For the initial session data, a compact temporal-spatial feature extractor (CTS) is designed to acquire the temporal-spatial features, which are input to a baseline BLS (bBLS). Furthermore, for new session data, a mutual information maximization constraint (MIMC) is introduced into the loss function of CTS to make the features' probability distribution sufficiently similar to that of the previous session, a new incremental BLS sequence (iBLS) is obtained by adding a small number of nodes to the previous model, and so on. Experiments are conducted based on the BCI Competition IV-2a dataset with two sessions and IV-2b dataset with five sessions,ASiBLS achieves average decoding accuracies of 79.89% and 87.04%, respectively. The kappa coefficient and forgetting rate are also used to evaluate the model performance. The results show that ASiBLS can adaptively generate an optimized and reduced model for each session successively, which has better plasticity in learning new knowledge and stability in retaining old knowledge as well.

GSyncCode: Geometry Synchronous Hidden Code for One-step Photography Decoding

Invisible hyperlinks and hidden barcodes have recently emerged as a hot topic in offline-to-online messaging, where an invisible message or barcode is embedded in an image and can be decoded via camera shooting. Current schemes involve a two-step decoding process: starting with vertex localization of the embedded region to correct the perspective distortion introduced by shooting, followed by decoding the message from the corrected region. However, vertex localization can be complex and time-consuming, which affects the efficiency and accuracy of message decoding. To address this issue, this paper proposes a geometry synchronous decoding scheme called GSyncCode, allowing for one-step extraction of a Data Matrix code from the photograph. Instead of correction before decoding, GSyncCode directly decodes a geometry-transformed Data Matrix that is synchronized with the embedded region. A barcode scanner is then used to efficiently retrieve messages. We design a Haar-transform based encoder HaarUNet and a HaarLoss visual function to select the key component of the Data Matrix for embedding. They improve the visual quality of the embedded image by reducing redundant embedding signals. Extensive simulated and real-world experiments demonstrate the superiority of GSyncCode in both decoding efficiency and accuracy. Our codes are published at: https://github.com/zcx-language/GSyncCode.

A neuronal marker of eye contact spontaneously activated in neurotypical subjects but not in autistic spectrum disorders

Attention to faces and eye contact are key behaviors for establishing social bonds in humans. In Autism Spectrum Disorders (ASD), a disturbance in neurodevelopment, impaired face processing and gaze avoidance are key clinical features for ASD diagnosis. The biological alterations underlying these impairments are not yet clearly established. Using high-density electroencephalography coupled with multi-variate pattern classification and group blind source separation methods we searched for face- and-face components-related neural signals that could best discriminate visual processing of neurotypical subjects (N = 38) from ASD participants (N = 27). We isolated a face-specific neural signal in the superior temporal sulcus peaking at 240 msec after face-stimulus onset. A machine learning algorithm applied on the extracted neural component reached 74% decoding accuracy at the same latencies, discriminating the neurotypical population from ASD subjects in whom this signal was weak. By manipulating attention on different parts of the face, we also found that the power of the evoked signal in neurotypical subjects varied depending on the region observed: it was strong when the eye region fell on the fovea to decrease on regions further away and outside the stimulus face. Such face and face-components selective neural modulations were not found in ASD, although they did show typical early face-related P100 and N170 signals. These results show that specialized cortical mechanisms for face perception show higher responses for eyes when attention is focused on gaze and that these mechanisms may be particularly affected in autism spectrum disorders.

MSMGE-CNN: a multi-scale multi-graph embedding convolutional neural network for motor related EEG decoding

Abstract Deep learning (DL) technique has been widely used for decoding motor related electroencephalography (EEG) signals, which has considerably driven the development of motor related brain-computer interfaces (BCIs). However, traditional convolutional neural networks (CNNs) cannot fully represent spatial topology information and dynamic temporal characteristics of multi-channel EEG signals, resulting in limited decoding accuracy. To address such challenges, a novel multi-scale multi-graph embedding convolutional neural network (MSMGE-CNN) is proposed in this study. The proposed MSMGE-CNN contains two crucial components: multi-scale time convolution and multi-graph embedding. Specifically, we design a multi-branch CNN architecture with mixed-scale time convolutions based on EEGNet to sufficiently extract robust time domain features. Afterward, we embed multi-graph information obtained based on physical distance proximity and functional connectivity of multi-channel EEG signals into the time-domain features to capture rich spatial topological dependencies via multi-graph convolution operation. We extensively evaluated the proposed method on three benchmark EEG datasets commonly used for motor imagery/execution (MI/ME) classification and obtained accuracies of 79.59 % (BCICIV-2a Dataset), 69.77 % (OpenBMI Dataset) and 96.34 % (High Gamma Dataset), respectively. These results powerfully demonstrate that MSMGE-CNN outperforms several state-of-the-art algorithms. In addition, we further conducted a series of ablation experiments to validate the rationality of our network architecture. Overall, the proposed MSMGE-CNN method dramatically improves the accuracy and robustness of MI/ME-EEG decoding, which can effectively enhance the performance of motor related BCI system.

Direction and velocity kinematic features of point-light displays grasping actions are differentially coded within the action observation network

The processing of kinematic information embedded in observed actions is an essential ability for understanding others' behavior. Previous research showed that the action observation network (AON) may encode some action kinematic features. However, our understanding of how direction and velocity are encoded within the AON is still limited. In this study, we employed event-related fMRI to investigate the neural substrates specifically activated during observation of hand grasping actions presented as point-light displays, performed with different directions (right, left) and velocities (fast, slow). Twenty-three healthy adult participants took part in the study. To identify brain regions differentially recruited by grasping direction and velocity, univariate and multivariate pattern analysis (MVPA) were performed. The results of univariate analysis demonstrate that direction is encoded in occipito-temporal and posterior visual areas, while velocity recruits lateral occipito-temporal, superior parietal and intraparietal areas. Results of MVPA further show: a) a significant decoding accuracy of both velocity and direction at the network level; b) the possibility to decode within lateral occipito-temporal and parietal areas both direction and velocity; c) a contribution of bilateral premotor areas to velocity decoding models. These results indicate that posterior parietal nodes of the AON are mainly involved in coding grasping direction and that premotor regions are crucial for coding grasping velocity, while lateral occipito-temporal cortices play a key role in encoding both parameters. The current findings could have implications for observational-based rehabilitation treatments of patients with motor disorders and artificial intelligence-based hand action recognition models.

Attention model of EEG signals based on reinforcement learning.

Applying convolutional neural networks to a large number of EEG signal samples is computationally expensive because the computational complexity is linearly proportional to the number of dimensions of the EEG signal. We propose a new Gated Recurrent Unit (GRU) network model based on reinforcement learning, which considers the implementation of attention mechanisms in Electroencephalogram (EEG) signal processing scenarios as a reinforcement learning problem. The model can adaptively select target regions or position sequences from inputs and effectively extract information from EEG signals of different resolutions at multiple scales. Just as convolutional neural networks benefit from translation invariance, our proposed network also has a certain degree of translation invariance, making its computational complexity independent of the EEG signal dimension, thus maintaining a lower learning cost. Although the introduction of reinforcement learning makes the model non differentiable, we use policy gradient methods to achieve end-to-end learning of the model. We evaluated our proposed model on publicly available EEG dataset (BCI Competition IV-2a). The proposed model outperforms the current state-of-the-art techniques in the BCI Competition IV- 2a dataset with an accuracy of 86.78 and 71.54% for the subject-dependent and subject-independent modes, respectively. In the field of EEG signal processing, attention models that combine reinforcement learning principles can focus on key features, automatically filter out noise and redundant data, and improve the accuracy of signal decoding.

Gamma oscillation optimally predicts finger movements

Our fingers are the most dexterous and complicated parts of our body and play a significant role in our daily activities. Non-invasive techniques, such as Electroencephalography (EEG) and Electromyography (EMG) can be used to collect neural and muscular signals related to finger movements. In this study, we combined an 8-channel EMG and a 31-channel EEG while the human subject moved one of the five fingers on the right hand. To identify the best EEG frequency features that encode distinct finger movements, we systematically examined the decoding accuracies of the slow-cortical potentials and three types of sensorimotor rhythms, namely the Mu, beta, and gamma oscillations. For both EMG and EEG, we came up with a simple and unified root mean square or power approach that avoided the complex signal features used by previous studies. The signal features were then fed into a feedforward artificial-neural-network (ANN) classifier. We found that the low-gamma oscillation provided the best decoding performance over the other frequency bands, ranging from 65.0 % to 89.0 %, which was comparable to the EMG performance. Combining EMG and low gamma into a single ANN can further improve the outcome for subjects who had showed suboptimal performances with EMG or EEG alone. This study provided a simple and efficient algorithm for prosthetics that assist patients with sensorimotor impairments.

A data compression algorithm with the improved SRLE for high-throughput neural signal acquisition device.

Multi-channel acquisition systems of brain neural signals can provide a powerful tool with a wide range of information for the clinical application of brain computer interfaces. High-throughput implantable systems are limited by size and power consumption, posing challenges to system design. To acquire more comprehensive neural signals and wirelessly transmit high-throughput brain neural signals, a FPGA-based acquisition system for multi-channel brain nerve signals has been developed. And the Bluetooth transmission with low-power technology are utilized. To wirelessly transmit large amount of data with limited Bluetooth bandwidth and improve the accuracy of neural signal decoding, an improved sharing run length encoding (SRLE) is proposed to compress the spike data of brain neural signal to improve the transmission efficiency of the system. The functional prototype has been developed, which consists of multi-channel data acquisition chips, FPGA main control module with the improved SRLE, a wireless data transmitter, a wireless data receiver and an upper computer. And the developed functional prototype was tested for spike detection of brain neural signal by animal experiments. From the animal experiments, it shows that the system can successfully collect and transmit brain nerve signals. And the improved SRLE algorithm has an excellent compression effect with the average compression rate of 5.94%, compared to the double run-length encoding, the FDR encoding, and the traditional run-length encoding. The developed system, incorporating the improved SRLE algorithm, is capable of wirelessly capturing spike signals with 1024 channels, thereby realizing the implantable systems of High-throughput brain neural signals.

Multitasking practice eliminates modality-based interference by separating task representations in sensory brain regions.

The debate on the neural basis of multitasking costs evolves around neural overlap between concurrently performed tasks. Recent evidence suggests that training-related reductions in representational overlap in fronto-parietal brain regions predict multitasking improvements. Cognitive theories assume that overlap of task representations may lead to unintended information exchange between tasks (i.e., crosstalk). Modality-based crosstalk was suggested as a source for multitasking costs in multisensory settings. Robust findings of increased costs for certain modality mappings may be explained by crosstalk between the stimulus modality in one task and sensory action consequences in the concurrently performed task. Whether modality-based crosstalk emerges from representational overlap in general fronto-parietal multitasking regions or modality-specific regions is not known yet. In this functional neuroimaging study, we investigate neural overlap during multitasking performance in humans, focusing on modality compatibility by employing multivariate pattern analysis and modality-specific practice interventions in three groups (total N = 54, 24 females). We observed significant differences between modality compatible and modality incompatible single-task representations, specifically in the auditory cortex but not in fronto-parietal regions. Notably, improved auditory decoding accuracy related to modality incompatible tasks was predictive of performance gains in the corresponding dual task along with complete elimination of modality-specific dual-task costs. This predictive relationship was evident only in the group practicing modality incompatible mappings, suggesting that specific practice on task sets with modality overlap influenced both neural representations and subsequent multitasking performance. This study contributes to the integration of cognitive theory and neuroscience and the role of task representations in dual-task interference.Significance Statement In a society dominated by multitasking, understanding its neurocognitive basis and plasticity is crucial for key aspects of everyday tasks. We investigate the neural mechanisms behind multitasking limitations, offering insights for targeted cognitive interventions. The study builds upon established theories of cognitive multitasking and imaging research, addressing the concept of modality-based crosstalk - the unintended exchange of modality-based information between tasks. Through functional brain imaging and pattern analysis, we examined how neural task representations contribute to performance costs in dual tasks with varying degrees of modality overlap. Notably, our findings demonstrate a practice-related decrease in neural overlap which is associated with substantial multitasking improvements, specifically in the auditory cortex, emphasizing the contribution of sensory regions to flexible multidimensional task representations.

Synchronization-based fusion of EEG and eye blink signals for enhanced decoding accuracy

Decoding locomotor tasks is crucial in cognitive neuroscience for understanding brain responses to physical tasks. Traditional methods like EEG offer brain activity insights but may require additional modalities for enhanced interpretative precision and depth. The integration of EEG with ocular metrics, particularly eye blinks, presents a promising avenue for understanding cognitive processes by combining neural and ocular behaviors. However, synchronizing EEG and eye blink activities poses a significant challenge due to their frequently inconsistent alignment. Our study with 35 participants performing various locomotor tasks such as standing, walking, and transversing obstacles introduced a novel methodology, pcEEG+, which fuses EEG principal components (pcEEG) with aligned eye blink data (syncBlink). The results demonstrated that pcEEG+ significantly improved decoding accuracy in locomotor tasks, reaching 78% in some conditions, and surpassed standalone pcEEG and syncBlink methods by 7.6% and 22.7%, respectively. The temporal generalization matrix confirmed the consistency of pcEEG+ across tasks and times. The results were replicated using two driving simulator datasets, thereby confirming the validity of our method. This study demonstrates the efficacy of the pcEEG+ method in decoding locomotor tasks, underscoring the importance of temporal synchronization for accuracy and offering a deeper insight into brain activity during complex movements.

Small-sample cucumber disease identification based on multimodal self-supervised learning

It is difficult and costly to obtain large-scale, labeled crop disease data in the field of agriculture. How to use small samples of unlabeled data for feature learning has become an urgent problem that needs to be solved. The emergence of self-supervised contrastive learning methods and self-supervised mask learning methods can solve the problem of missing labels on the training data. However, each of these paradigms comes with its own advantages and drawbacks. At the same time, the features learned by dataset in a single modality are limited, ignoring the correlation with other modal information. Hence, this paper introduced an effective framework for multimodal self-supervised learning, denoted as MMSSL, to address the task of identifying cucumber diseases with small sample sizes. Integrating image self-supervised mask learning, image self-supervised contrastive learning, and multimodal image-text contrastive learning, the model can not only learn disease feature information from different modalities, but also capture global and local disease feature information. Simultaneously, the mask learning branch was enhanced by introducing a prompt learning module based on a cross-attention network. This module aided in approximately locating the masked regions in the image data in advance, facilitating the decoder in making accurate decoding predictions. Experimental results demonstrate that the proposed method achieves a 95% accuracy in cucumber disease identification in the absence of labels. The approach effectively uncovers high-level semantic features within multimodal small-sample cucumber disease data. GradCAM is also employed for visual analysis to further understand the decision-making process of the model in disease identification. In conclusion, the proposed method in this paper is advantageous for enhancing the classification accuracy of small-sample cucumber data in a multimodal, unlabeled context, demonstrating good generalization performance.

Diffusion models-based motor imagery EEG sample augmentation via mixup strategy

Deep representation learning has been widely explored for decoding motor imagery electroencephalogram (MI-EEG) to build EEG-tailored brain-computer interfaces. Due to the labor-intensive and time-consuming recording to MI-EEG, recently BCIs were suffered from sample scarcity, especially for the limitation of sample diversity. To solve this issue, we proposed a novel sample augmentation method, Diffusion models-based EEG Sample Augmentation method via Mixup strategy (DESAM), to solve sample scarcity and improve sample diversity for both multivariate time-series and time–frequency images forms. To reduce the characteristics conflicts brought by the augmented samples, a mixup strategy with data weighting was proposed for both forms. To validate the efficacy of sample augmentation by DESAM, we conducted comparative experiments on three publicly available MI-EEG datasets, and reported the average decoding accuracies and Kappa values on augmented samples by various deep learning models. For the BCI Competition IV datasets 2a, 2b, and 1, the two forms of DESAM sample augmentations achieved improvements for both accuracies and Kappa values. Ablation studies have demonstrated the necessity and significance of our DESAM method, and it makes a possible manner to solve sample scarcity issue of MI-EEG.