Electroencephalogram (EEG)-based emotion recognition holds great potential in intelligent human-computer interaction and brain-computer interface systems, as the brain generates distinct electrical activity patterns under different emotional states. However, EEG information often contains data from numerous channels, leading to high computational cost and potential redundancy. Existing channel selection methods often rely on uniform rules, lacking frequency-specific adaptability and inter-channel modeling, which can cause information loss and reduced performance during dimensionality reduction. To address this issue, we propose a novel framework that combines discriminative channel selection with hierarchical spatial-temporal modeling to enhance both performance and efficiency. In preprocessing, wavelet coherence and mutual information are used to adaptively select informative channels across multiple frequency bands. The selected signals are then processed by a Spatial-Temporal Graph-aware Network (STG-Net), which models spatial relationships between channels through graph convolution, extracting spatial features from each time frame. Coupled with a temporal modeling module, the network further captures the evolving temporal patterns of emotional states across consecutive frames. Finally, frequency-spatial-temporal features are fused for emotion classification. Compared to the state-of-the-art methods, our approach achieves superior performance in both recognition accuracy and model efficiency.

In recent years, non-invasive electroencephalography (EEG) has been widely used to decode high-level cognitive functions, such as visual perception and imagination. The processes of visual perception and imagination in the human brain have been shown to share similar neural circuits and activation patterns in cognitive science. However, current research predominantly focuses on single cognitive processes, overlooking the natural commonalities between these processes and the insights that multimodal approaches can provide. To address this, this study proposes a novel multimodal model, MMPI Net, for jointly decoding EEG signals of visual image perception and imagination. MMPI Net comprises four components: Primitive Feature Extraction for Perception and Imagination (PFE), Cross-Semantic Feature Fusion (CSFF), Joint Semantic Feature Decoder (JSFD), and Semantic Classification (SC). To ensure the effectiveness of PFEM, an Improved Channel Attention Mechanism is introduced, which employs multiple parallel convolutional branches to enhance the extraction of important information and utilizes a Diverse Branch Block approach to reduce the parameter count. In the CSFF module, a cross-attention-based fusion method is designed to effectively capture and utilize intermodal information. In the JSFD phase, a Kolmogorov-Arnold Network is incorporated and coupled with linear layers to improve classification performance. Finally, a linear layer with Softmax is used as the SC module. Experimental results on two publicly available datasets show that, compared to models that use a single cognitive process, MMPI Net achieves average accuracy improvements of 14.22% and 106.1%, demonstrating its effectiveness.

Metaverse is an initiative, immersive virtual-reality digital space that installs to the next-generation Internet for a platform of network human-computer interaction mode. As the access point number of routing/switching process in non-linear heterogeneous scheme is increased, the original method for the word, voice, or image transmission can not conform to the node and link delay needs of multi-lead electroencephalogram, and multi-media audio/video operating in brain-computer interface or virtual-reality terminal. The parallelization and distribution of data streams with routing/switching capability was not only embed in software-defined network but also plays a critical role in software-defined radio. The purpose of this manuscript is to enable the Metaverse to get an initiative control and immersive perceptual abilities, we introduce an non-linear heterogeneous scheme to make a novel initiative, immersive human-computer interactions: Software-defined memristive neural networks. In the link/network layer, we use the function of software-defined approach in control of a threshold during the early high-speed switching fabric in EEG input or audio/video output. However, the multiple control or perceptual data streams of our thought and sense in conventional routing nodes remain a poorly serial running state all the time. Here, we design a large-scale switch parallelization of memristor-based FPGA system on a basis of previous non-linear heterogeneous scheme for the designing of a routing node in memristive neural network. To conclude, our results verify a high performance in the novel human-computer interaction with multi-lead/multi-media information.

Electroencephalogram (EEG)-based emotion recognition has become a research hotspot in the field of brain-computer interface. Previous emotion recognition methods have overlooked the fusion of multi-domain emotion-specific information to improve performance, and faced the challenge of insufficient interpretability. In this paper, we proposed a novel EEG emotion recognition model that combined the asymmetry of the brain hemisphere, and the spatial, spectral, and temporal multi-domain properties of EEG signals, aiming to improve emotion recognition performance. Based on the 10-20 standard system, a global spatial projection matrix (GSPM) and a bi-hemisphere discrepancy projection matrix (BDPM) are constructed. A dual-stream spatial-spectral-temporal convolution neural network is designed to extract depth features from the two matrix paradigms. Finally, the transformer-based fusion module is used to learn the dependence of fused features, and to retain the discriminative information. We conducted extensive experiments on the SEED, SEED-IV, and DEAP public datasets, achieving excellent average results of 98.33/2.46 %, 92.15/5.13 %, 97.60/1.68 %(valence), and 97.48/1.42 %(arousal) respectively. Visualization analysis supports the interpretability of the model, and ablation experiments validate the effectiveness of multi-domain and bi-hemisphere discrepancy information fusion.

Massive multiple-input multiple-output (MIMO) assisted grant-free access is a compelling approach to support uplink ultra-reliable low-latency communications (URLLCs). It is a natural idea to combine massive MIMO and repetition transmission to further improve the reliability. However, excessive repetitions may increase the collision probability in multiuser scenarios and thus limit the reliability improvement. In this paper, we propose a novel semi-probabilistic repetition (SPRe) scheme for the massive multiuser MIMO (MU-MIMO) systems with grant-free access. By introducing a transmission probability for the replicas, the SPRe is more flexible than the popular repetition scheme, i.e.,$K$-repetition, and covers the$K$-repetition as a special case. Considering practical MU-MIMO systems with finite block-length coding, channel estimation errors, pilot collisions and the sporadic arrivals of URLLC traffic, the reliability of the proposed SPRe is thoroughly analyzed and derived in closed form. Through further asymptotic analysis, the optimal transmission probability to maximize the reliability of MU-MIMO is derived for the high-SNR regime and low-load scenarios. Moreover, the diversity gain of the proposed SPRe is also analyzed and it is proved that the optimal SPRe is superior to the$K$-repetition in terms of reliability. The analysis and superiority of the SPRe are finally verified through computer simulations.

In this research, we aimed to design a architecture of brain-computer interaction (BCI) in routing nodes routing and forwarding mechanism by exploring the memristive neuron networks (MNN) among heterogeneous multi-agent system (HMAS). In recent years, field programmable gate array (FPGA) has become a popular choice to construct a heterogeneous network system for routing and forwarding mechanism due to its ability to describe a hardware information system in a software defined method. However, most BCI approaches focus only on capturing the dynamics of EEG signal by simply receiving all wave of individual or making up them as a single system. Such methods neglect the existing problems of traditional BCI mode. To this moment, we propose a novel Heterogeneous Multi-Agent System (HMAS) to model the MNN among a group of logic units for realizing the multi-agent system in BCI approach. Specifically, we first feed each agent's non-linear feature into a routing node to model the routing and forwarding functions. Subsequently, at one time step, the outputs of all testing routers are fed into a heterogeneous network space, which mainly consists of multiple sub-agents, a new forward port, and some memory memristive neuron which build a non-linear network system to set up a multiple agent in MNN environment. In the FPGA logic unit, each sub-memory unit stores individual motion information, while these programmable units selectively integrates and stores a huge number of motion information between multiple interacting persons from multiple sub-memory units via the forwarding ports and MNN storage unit, respectively. According to the extensive experimental results on several experimental datasets validate the effectiveness of the proposed HMAS-BCI by comparing against a optimization method of the experimental scene.

The optimal design of the energy-efficient multiple-input multiple-output (MIMO) aided uplink ultra-reliable low-latency communications (URLLC) system is an important but unsolved problem. For such a system, we propose a novel absorbing-Markov-chain-based analysis framework to shed light on the puzzling relationship between the delay and reliability, as well as to quantify the system energy efficiency. We derive the transition probabilities of the absorbing Markov chain considering the Rayleigh fading, the channel estimation error, the zero-forcing multi-user-detection (ZF-MUD), the grant-free access, the ACK-enabled retransmissions within the delay bound and the interactions among these technical ingredients. Then, the delay-constrained reliability and the system energy efficiency are derived based on the absorbing Markov chain formulated. Finally, we study the optimal number of user equipments (UEs) and the optimal number of receiving antennas that maximize the system energy efficiency, while satisfying the reliability and latency requirements of URLLC simultaneously. Simulation results demonstrate the accuracy of our theoretical analysis and the effectiveness of massive MIMO in supporting large-scale URLLC systems.