The reliable monitoring of the depth of anesthesia (DoA) is essential to control the anesthesia procedure. Electroencephalography (EEG) has been widely used to estimate DoA since EEG could reflect the effect of anesthetic drugs on the central nervous system (CNS). In this study, we propose that a deep learning model consisting mainly of a deep residual shrinkage network (DRSN) and a 1 × 1 convolution network could estimate DoA in terms of patient state index (PSI) values. First, we preprocessed the four raw channels of EEG signals to remove electrical noise and other physiological signals. The proposed model then takes the preprocessed EEG signals as inputs to predict PSI values. Then we extracted 14 features from the preprocessed EEG signals and implemented three conventional feature-based models as comparisons. A dataset of 18 patients was used to evaluate the models’ performances. The results of the five-fold cross-validation show that there is a relatively high similarity between the ground-truth PSI values and the predicted PSI values of our proposed model, which outperforms the conventional models, and further, that the Spearman’s rank correlation coefficient is 0.9344. In addition, an ablation experiment was conducted to demonstrate the effectiveness of the soft-thresholding module for EEG-signal processing, and a cross-subject validation was implemented to illustrate the robustness of the proposed method. In summary, the procedure is not merely feasible for estimating DoA by mimicking PSI values but also inspired us to develop a precise DoA-estimation system with more convincing assessments of anesthetization levels.

We study the interplay of topology and dynamics in a neural network connected with spike-timing-dependent plasticity (STDP) synapses. Stimulated with periodic spike trains, the STDP-driven network undergoes a synaptic pruning process and evolves to a residual network. We examine the variation of topological and dynamical properties of the residual network by varying two key parameters of STDP: synaptic delay and the ratio between potentiation and depression. Our extensive numerical simulations of the leaky integrate-and-fire model show that there exists two regions in the parameter space. The first corresponds to fixed-point configurations, where the distribution of peak synaptic conductances and the firing rate of neurons remain constant over time. The second corresponds to oscillating configurations, where both topological and dynamical properties vary periodically, which is a result of a fixed point becoming a limit cycle via a Hopf bifurcation. This leads to interesting questions regarding the implications of these rhythms in the topology and dynamics of the network for learning and cognitive processing.

There has been a recent explosion of interest in neuromorphic computing capable of processing sophisticated and large-scale information based on spike coding, which has advantages in the implementation on electronic or optical neuromorphic systems. The neural engineering framework (NEF), a formal method for mapping attractor networks and control-theoretic algorithms to spiking neural networks, provides us a way to implement neuromorphic or numerical computing. In this paper, we implement the NEF on an optoelectronic architecture based on a photonic neuromorphic system. We discuss how electronic signals can be encoded to patterns of optical pulses, processed in the dynamic neural activity of recurrent network, and finally decoded back to the desired signals. These methods take advantage of the mechanism of spiking laser neurons and the wavelength division multiplexing protocol and explore the rapidity of optical pulses. Our work is mainly simulation with qualitative analysis. Simulation studies demonstrate three NEF principles on the proposed architecture with signals on nanosecond time scale, which run about six orders of magnitude faster than electronic counterparts on the NEF implementation. The present architecture can be also used in a broad domain of applications where more complex neuromorphic or numerical computing is necessary.

Timely detection and response to Intraoperative Hypotension (IOH) during surgery is crucial to avoid severe postoperative complications. Although several methods have been proposed to predict IOH using machine learning, their performance still has space for improvement. In this paper, we propose a ResNet-BiLSTM model based on multitask training and attention mechanism for IOH prediction. We trained and tested our proposed model using bio-signal waveforms obtained from patient monitoring of non-cardiac surgery. We selected three models (WaveNet, CNN, and TCN) that process time-series data for comparison. The experimental results demonstrate that our proposed model has optimal MSE (43.83) and accuracy (0.9224) compared to other models, including WaveNet (51.52, 0.9087), CNN (318.52, 0.5861), and TCN (62.31, 0.9045), which suggests that our proposed model has better regression and classification performance. We conducted ablation experiments on the multitask and attention mechanisms, and the experimental results demonstrated that the multitask and attention mechanisms improved MSE and accuracy. The results demonstrate the effectiveness and superiority of our proposed model in predicting IOH.