Search

Pathological articulation exploration, especially the study of the kinematic characteristics of motor organ, is helpful to further reveal the essence of motor dysarthria. Due to the scarcity of the available pathological pronunciation database, there has little research working on the statistic distribution analysis for patients and normal controlled people. This paper applied the distribution method on TORGO database to discover the cognitive and motor rules of dysarthria patients. Single phoneme analysis is effective for locating the specific tongue muscle but ignoring cognitive ability assessment, particularly for the content understanding and the fluency degree of expression by patients. The paper focused on the word/sentence level rather than single phoneme analysis. The reaction time was designed to reveal the relationship between the brain cognition and motor neuron activation. The statistic distribution tells that the cerebral palsy or amyotrophic lateral sclerosis does affect people's reflection and make the patients hard to control the tongue muscles effectively, resulting in unstable reaction time. The articulation velocity of patients appears 5mm/s faster than normal people, at 85mm/s, perhaps due to the factors of the word/sentence data and the big proportion of extra large displacement. It illustrates that the tongue moves relatively coherently and fluently once patients active the muscles, but hard to slow down as the muscle control ability decreases. The spatial occupancy was represented by the maximum articulation movement range (MAMR). We adapted the logarithmic normal distribution to find out the significant threshold for the diagnosis of dysarthria with the MAMR exceeding 7mm along left and right direction and the number of abnormal ranges surpassing 10% of total number. Primary test of MAMR, as an articulatory feature, for speech classification was carried out and it achieves 81% accuracy. These explorations convinced us to apply them to the pathological speech recognition task for improvement in future.

Emotion plays a significant role in perceiving external events or situations in daily life. Due to ease of use and relative accuracy, Electroencephalography (EEG)-based emotion recognition has become a hot topic in the affective computing field. However, scalp EEG is a mixed-signal and cannot directly indicate the exact information about active cortex sources of different emotions. In this paper, we analyze the significant differences of active source regions and frequency bands for pairs of emotions-based reconstructed EEG sources using sLORETA, and 26 Brodmann areas are selected as the regions of interest (ROI). And then, six kinds of time- and frequency-domain features from significant active regions and frequency bands are extracted to classify different emotions using support vector machines. Furthermore, we compare the classification performances of emotion features extracted from active source regions and EEG sensors. We have demonstrated that the features from selected source regions can improve the classification accuracy by extensive experiments on the DEAP and TYUT 2.0 EEG-based datasets.

The recommendation algorithm based on Singular Value Decomposition (SVD)++ is a widely used algorithm for its good prediction performance. However, with the rapid increase of data in smart societies, the poor computational performance of the SVD++ recommendation algorithm becomes a prominent disadvantage, for it takes a longer time to optimize the objective function during constructing the prediction model. The learning rate function is a significant factor in the prediction model based on the SVD++ recommendation algorithm. It can directly affect the convergence speed of the prediction model and the performance of the model. The traditional model uses an exponential function, natural exponential function or piecewise constant as its learning rate function. In this paper, a novel adaptive learning rate (ALR) function is proposed, which combines the exponential with linear functions, and the function is applied to the SVD++ recommendation algorithm. The highlights of the paper are as follows. First, with a larger initial value, the proposed function descends quicker and tends to the end with a less step. Second, the theoretical properties of the proposed learning rate function are verified through theoretical analysis, including the theoretical proof of its convergence and the iteration speed comparison. Compared to the existing learning rate functions, the proposed ALR function works better on the convergence speed through mathematical derivation. Finally, the novel ALR function is applied to the SVD++ recommendation algorithm as recommendation model ALRSVD++. Some existing learning rate methods are used as benchmarks for illustrating the computation and prediction performances of proposed ALR function and its ALRSVD++ model. Experimental results demonstrated that the SVD++ recommendation algorithm based on the proposed ALR function improved computational efficiency of the training model ALRSVD++ significantly. Especially, to the larger size training dataset, the iterations and training time based on the proposed ALR function and ALRSVD++ model reduced in a great deal, without greatly sacrificing the recommendation performance.

Accurate behavioral assessments of consciousness are crucial in guiding management as they provide diagnostic and prognostic information, but they are challenging since different clinical scales lead to varying results for the same patient. This paper aimed to investigate the phase synchrony index in differentiating the states of consciousness (wakefulness, somnolence, stupor, light coma, middle coma, and deep coma) in ischemic stroke. We evaluated whether the quantitative electroencephalography (EEG) measure, phase synchrony index of the left and right hemispheres, can facilitate the assessment of consciousness in stroke patients. This paper included 82 patients with ischemic stroke admitted for inpatient rehabilitation. The phase synchrony index of the left and right hemispheres was computed in the alpha band (8–12 Hz) and beta band (13–30 Hz), respectively. The associations between the phase synchrony index of the left and right hemispheres in two frequency bands with the clinical states, including the level of consciousness and the National Institutes of Health Stroke Scale score were analyzed. For further assessments of phase synchrony of the left and right hemispheres (PSI-LR), four local phase synchrony indexes in the beta band were also evaluated. The experiments result showed that PSI-LRs in the beta band correlated significantly with the level of consciousness of ischemic stroke patient and the National Institutes of Health Stroke Scale score, and they can identify the impaired consciousness in ischemic stroke patients with an accuracy of 84.15%. Compared with local phase synchrony, the phase synchrony index of channel FP1 and channel FP2 exhibited a more significant correlation with the level of consciousness than the other three local phase synchrony measures. These results suggest that the novel quantitative EEG measure, PSI-LR provides a new objective way to assess the level of consciousness in ischemic stroke patients.

Consciousness is frequently assessed in acutely brain-injured patients. However, a direct measure of consciousness based on neurological examinations and behavioral observations is time-consuming and costly in medical labors. This paper aims to explore effective quantitative electroencephalography (EEG) measures to facilitate the objective assessment of consciousness. We do two studies in this paper. In the first study, the EEG signals recorded from 27 cerebral hemorrhage patients are analyzed to explore whether the phase synchrony index is effective in distinguishing the states of consciousness in cerebral hemorrhage patients. In the second study, 14 EEG recordings from four ischemic stroke patients, who have experienced more than two types of conscious disturbance, are studied to complement the analysis of the longitudinal value of the phase synchrony index. In both studies, a weighted sum of the six quantitative EEG features is analyzed on its correlation with the level of consciousness, and a linear regression model is built for the correlation analysis. In the first study, the phase synchrony index of the left and the right hemispheres (PSI-LR) in the beta band has shown a significant correlation with the level of consciousness in cerebral hemorrhage patients (Pearson correlation coefficients ρ = 0.500 and Spearman correlation coefficients ρ = 0.400 ). In the second study, PSI-LR also shows a high correlation with the longitudinal changes of consciousness in ischemic stroke patients (Pearson correlation coefficients ρ = 0.884 and Spearman correlation coefficients ρ = 0.887 ). Furthermore, the linear regression model combining the six quantitative EEG (QEEG) features shows a determination coefficient of 44.2% in cerebral hemorrhage patients and 83.8% in longitudinal ischemic stroke investigations, respectively. The PSI-LR is an effective indicator in not only identifying the disturbance of consciousness in cerebral hemorrhage patients, but also monitoring the changes of consciousness in ischemic stroke patients. A linear combination of the six QEEG features investigated in this paper is a new measure to improve the assessment of consciousness.

Random projection based on compressed sensing can reduce the amount of data transmitted in a wireless sensor network (WSN), and efficient routing can reduce the network traffic. Thus, this paper presents a Random projection-Polar coordinate-Chain routing (RPC) method. The method uses polar coordinates to locate nodes, establishes a chain structure to form a route, and applies random projection to achieve the compressed data collection. In a WSN, the sink is the center of the data collection. With the sink as the pole, the polar coordinates can be used to determine the orientation of each node relative to the sink so that nodes can be searched under certain conditions. By adopting the chain topological structure, the establishment of the chain through the greedy algorithm reduces the energy consumption and complexity. Based on the comparative analysis, a four-quadrant chain routing method combining the polar radius and polar angle is proposed for smaller networks. For large-scale networks, a routing algorithm combining the sector and inner circle is proposed. Then, according to the random projection theory, the weighted sum of the random projections of each row of the corresponding measurement matrix in each partition is transmitted to the sink. The sink has collected all measurements of each partition to complete the signal reconstruction. In this method, the route is formed by searching in the zones according to the polar radius and the polar angle, which avoids the roundabout route between distant nodes and reduces the energy consumption of the network. By comparing the RPC method with other related methods and the simulation experiments for different types of routing, the proposed method is proved to be time and energy efficient.

Compressive data gathering (CDG) based on compressed sensing (CS) theory for wireless sensor networks (WSNs) greatly reduces the amount of data transmitted compared with the traditional acquisition method that each node forwards the collected data directly to the next node. CDG combined with sparse random projection can further reduce the amount of data and thus prolong the lifetime of the WSN. The method of randomly selecting projection nodes as cluster heads to collect the weighted sum of sensor nodes outperforms the non-CS (without using CS) and hybrid-CS (applying CS only to relay nodes that are overloaded) schemes in decreasing the communication cost and distributing the energy consumption loads. However, the random selection of projection nodes causes the overall energy consumption of the network to be unstable and unbalanced. In this paper, we propose two compressive data gathering methods of balanced projection nodes. For WSN with uniform distribution of nodes, an even clustering method based on spatial locations is proposed to distribute the projection nodes evenly and balance the network energy consumption. For WSN with unevenly distributed nodes, an even clustering method based on node density is proposed, taking into account the location and density of nodes together, balancing the network energy and prolonging the network lifetime. The simulation results show that compared with the random projection node method and the random walk method, our proposed methods have better network connectivity and more significantly increased overall network lifetime.