Zhilin Zhang

Spatiotemporal Sparse Bayesian Learning With Applications to Compressed Sensing of Multichannel Physiological Signals

Energy consumption is an important issue in continuous wireless telemonitoring of physiological signals. Compressed sensing (CS) is a promising framework to address it, due to its energy-efficient data compression procedure. However, most CS algorithms have difficulty in data recovery due to nonsparsity characteristic of many physiological signals. Block sparse Bayesian learning (BSBL) is an effective approach to recover such signals with satisfactory recovery quality. However, it is time-consuming in recovering multichannel signals, since its computational load almost linearly increases with the number of channels. This work proposes a spatiotemporal sparse Bayesian learning algorithm to recover multichannel signals simultaneously. It not only exploits temporal correlation within each channel signal, but also exploits inter-channel correlation among different channel signals. Furthermore, its computational load is not significantly affected by the number of channels. The proposed algorithm was applied to brain computer interface (BCI) and EEG-based drivers drowsiness estimation. Results showed that the algorithm had both better recovery performance and much higher speed than BSBL. Particularly, the proposed algorithm ensured that the BCI classification and the drowsiness estimation had little degradation even when data were compressed by 80%, making it very suitable for continuous wireless telemonitoring of multichannel signals.

Photoplethysmography-Based Heart Rate Monitoring Using Asymmetric Least Squares Spectrum Subtraction and Bayesian Decision Theory

Motion artifacts (MAs) are strong interference sources in wearable photoplethysmography (PPG) signals, significantly affecting estimation of heart rate (HR) and other physiological parameters. In this paper, a novel method called SPECTRAP is proposed for accurate motion-tolerant estimation of HR using a PPG signal and a simultaneous acceleration signal. The method first calculates the spectra of the PPG signal and the acceleration signal, and then removes the MA spectral components from the PPG spectrum using a new spectrum subtraction algorithm. The new spectrum subtraction algorithm is based on asymmetric least square and overcomes drawbacks of the conventional spectrum subtraction algorithms. To find the spectral peak corresponding to HR on the resulting spectrum, SPECTRAP formulates the problem into a pattern classification problem and uses the Bayesian decision theory to solve it. Experimental results on the PPG database used in 2015 IEEE Signal Processing Cup showed that the proposed algorithm has excellent performance. The average absolute error on the twelve training sets was 1.50 beat per minute (BPM) (standard deviation: 1.95 BPM). The average absolute error on the ten testing sets was 2.13 BPM (standard deviation: 2.77 BPM).

Identifying the Neuroanatomical Basis of Cognitive Impairment in Alzheimer’s Disease by Correlation- and Nonlinearity-Aware Sparse Bayesian Learning

Predicting cognitive performance of subjects from their magnetic resonance imaging (MRI) measures and identifying relevant imaging biomarkers are important research topics in the study of Alzheimers disease. Traditionally, this task is performed by formulating a linear regression problem. Recently, it is found that using a linear sparse regression model can achieve better prediction accuracy. However, most existing studies only focus on the exploitation of sparsity of regression coefficients, ignoring useful structure information in regression coefficients. Also, these linear sparse models may not capture more complicated and possibly nonlinear relationships between cognitive performance and MRI measures. Motivated by these observations, in this work we build a sparse multivariate regression model for this task and propose an empirical sparse Bayesian learning algorithm. Different from existing sparse algorithms, the proposed algorithm models the response as a nonlinear function of the predictors by extending the predictor matrix with block structures. Further, it exploits not only inter-vector correlation among regression coefficient vectors, but also intra-block correlation in each regression coefficient vector. Experiments on the Alzheimers Disease Neuroimaging Initiative database showed that the proposed algorithm not only achieved better prediction performance than state-of-the-art competitive methods, but also effectively identified biologically meaningful patterns.

Compression via compressive sensing: A low-power framework for the telemonitoring of multi-channel physiological signals

Telehealth and wearable equipment can deliver personal healthcare and necessary treatment remotely. One major challenge is transmitting large amount of biosignals through wireless networks. The limited battery life calls for low-power data compressors. Compressive Sensing (CS) has proved to be a low-power compressor. In this study, we apply CS on the compression of multichannel biosignals. We firstly develop an efficient CS algorithm from the Block Sparse Bayesian Learning (BSBL) framework. It is based on a combination of the block sparse model and multiple measurement vector model. Experiments on real-life Fetal ECGs showed that the proposed algorithm has high fidelity and efficiency. Implemented in hardware, the proposed algorithm was compared to a Discrete Wavelet Transform (DWT) based algorithm, verifying the proposed one has low power consumption and occupies less computational resources.

Quantized Compressive Sensing for Low-Power Data Compression and Wireless Telemonitoring

In low-power wireless telemonitoring, physiological signals must be compressed before transmission to extend battery life. In this paper, we propose a two-stage data compressor based on quantized compressive sensing (QCS), where signals are first compressed by compressive sensing with a 50% compression ratio and then quantized with 2 bits per measurement. We also develop a reconstruction algorithm, called Bayesian de quantization (BDQ), to recover signals from the quantized compressed measurements. This algorithm exploits both the model of quantization errors and the correlated structure of physiological signals, which improves the quality of recovery. We validate the proposed data compressor and the recovery algorithm on wrist-type photoplethysmography data. This data are used to estimate the heart rate during fitness training. Results show that an average estimation error of 2.596 beats per minute is achieved using QCS and BDQ. This accuracy approaches the performance on non-compressed data, but we transmit n bits instead of n samples, which is a substantial improvement for low-power telemonitoring.

Sparse Bayesian learning for identifying the neuroanatomical basis of cognitive impairment in Alzheimer’s disease

a neurodegenerative disease. Three main variants have been described: nonfluent/agrammatic, semantic and logopenic. Each variant is more closely related to the involvement of defined parts of the language network, usually in the left hemisphere. Diffusion tensor images allow study the different anatomical tracts invivopatients.Objective:Toevaluate the potential ofDiffusion tensor tractography in the diagnosis of variants of primary progressive aphasia. Methods: Fifteen patients with clinical diagnosis of PPA were enrolled in this study. They were evaluated by a neurologist and the diagnosis or PPA and its variants was determined by clinical and neurolinguistic features following the diagnostic process according Gorno Tempini et al (2011). We selected the most paradigmatic cases for each variant according to the criteria above. The selected cases underwent 3T MRI with diffusion tensor images (DTI) and tractography was performed to evaluate the inferior longitudinal, uncinate and superior longitudinal fasciculus. Results: The patient with the non-fluent/agrammatic variant showed differences in the superior longitudinal and uncinate fasciculus, logopenic variant showed fiber dispersion in