This study intended to discover the effect of education and muscle relaxation (EMR) program on anxiety, depression and care burden among caregivers of acute stroke survivors. This randomized, controlled study enrolled a total of 110 caregivers of first-ever acute stroke patients, and randomly assigned to EMR (N = 55) and control (N = 55) groups. The caregivers in the EMR group received 12-month health education and progressive muscle relaxation, and those in control group were provided common rehabilitation advices. Hospital Anxiety and Depression Scale (HADS) and Zarit Caregiver Burden Scale in caregivers were evaluated at the time of patients’ discharge from hospital (M0), then at month(M) 3, M6 and M12 after the discharge. HADS-anxiety score, anxiety rate and severity were similar at M0, M3, while were reduced at M6 and M12 in EMR group compared to control group. Furthermore, HADS-depression score was similar at M0 and M3 but was decreased at M6 and M12 in EMR group compared with control group, however, there was no difference of depression rate and severity between the 2 groups at each time point. Moreover, Zarit Caregiver Burden Scale score was similar at M0 and M3, but was decreased at M6 and M12; meanwhile, degree of care burden was similar at M0, M3 and M6, but was reduced at M12 in EMR group compared to control group. EMR program decreases anxiety, depression and care burden in caregivers of acute stroke survivors, suggesting its potential in improving mental health and further promoting quality of lives in these caregivers.

To examine the association and the mediating effect among medication beliefs, perception of illness, and medication adherence in ischemic stroke patients. Patients and Methods: This is a cross-sectional study, 306 ischemic stroke patients recruited from The Second Affiliated Hospital of Harbin Medical University, China between June 2018 and October 2018. The Beliefs about Medications Questionnaire (BMQ) was used to assess a patient's beliefs about medication. The Brief Illness Perceptions Questionnaire (BIPQ) was used to rapidly determine the cognitive and emotional representation of ischemic stroke. Self-reported adherence was assessed using the Medication Adherence Report Scale (MARS). Logistic regression analysis, Pearson correlations, and mediation analysis were used to evaluate the association and mediating effects among medication beliefs, perception of illness, and medication adherence. Results: Overall, 220 (65.48%) participants were non-adherent to their ischemic stroke medications. Non-adherent patients had greater stroke severity (p = 0.031) compared to adherent patients. After adjusting for demographic characteristics, specific concern (odds ratio [OR]: 0.652, 95% confidence interval [CI]: 0.431 to 0.987, p-value [P] = 0.043), and the perception of illness (overall score) (OR: 0.964, 95% CI: 0.944 to 0.985, P = 0.001) were significantly associated with medication adherence in ischemic stroke patients. The mediation analysis showed the significant indirect effects of specific concern, general overuse, and general harm. It suggested that some impacts of medication beliefs have been mediated on medication adherence. Conclusion: Perceived concern about adverse effects of medicines and perception of illness have an influential impact on self-reported medication adherence in ischemic stroke patients. To enhance adherence, patients' beliefs about medication and perceptions of their disease should be reconsidered. Future work should investigate interventions to influence patient adherence by addressing concerns about their ischemic stroke medications and the perception of the disease.

No abstract

SVM (support vector machines) have become an increasingly popular tool for machine learning tasks involving classification, regression or novelty detection. In particular, they exhibit good generalization performance on many real issues and the approach is properly motivated theoretically. There are relatively a few free parameters to adjust and the architecture of the learning machine does not need to be found by experimentation. In this paper,survey of the key contents on this subject, focusing on the most well-known models based on kernel substitution, namely SVM, as well as the activated fields at present and the development tendency ,is presented. § 1 Introduction If you knock the word "SVM" in the SCI index tool on International network,you would take on thousands of records immediately. This shows its great effects on our world. SVM,namely,support vector machines have been successfully applied to a number of applications ranging from particle identification and text categorization to engine knock detection, bioinformatics and database marketing El-°]. The approach is systematic and properly motivated by statistical learning theory ES. Training involves the optimization of a convex cost function that there are no false local minima to complicate the learning process. This approach has many other benefits. For example ,the model constructed has an explicit dependence on the most informative patterns in data (support vectors). Hence the interpretation is straightforward and the data cleaning E8-~°3 could be implemented to improve performance. SVM are the most effective class of algorithms that use the idea of kernel substitution which we will broadly refer to a kernel method. Although SVM are researched deeply abroad and have obtained great success, however, few of people understand it in our country. The present review involves the main ideas of SVM algorithms for clustering and Received : 2002-04-12. MR Subject Classification .-03C43. SVM CLASSIFICATION 333reports the applications based on optimal character recognition. We do not attempt a full treatment of all available literature and fields,but we present a somewhat biased point of view illustrating the main ideas drawn mainly from the work of many authors and providing the best of our knowledge reference to related works for further reading. We hope that it nevertheless will be useful for the readers. It differs from other reviews,such as the ones in literatures E~1-133 ,mainly in the choice of the presented materials,that are allembracing whereas rough. We place more emphasis on the latest researching tendency,or challenge for users,and on connection to key techniques of SVM.We start by presenting some basic concepts and techniques of SVMs in § 1. Then we introduce the idea of kernel feature space and the original SVM approach, its implementation and some variants. The important properties of SVM will be devoted to question of model in § 2. Subsequently,we discuss the major difficulties of kernel-based methods learning in § 3. Current developments and o...

As an advanced measurement technique of non-radiant, non-intrusive, rapid response, and low cost, the electrical tomography (ET) technique has developed rapidly in recent decades. The ET imaging algorithm plays an important role in the ET imaging process. Linear back projection (LBP) is the most used ET algorithm due to its advantages of dynamic imaging process, real-time response, and easy realization. But the LBP algorithm is of low spatial resolution due to the natural ‘soft field’ effect and ‘ill-posed solution’ problems; thus its applicable ranges are greatly limited. In this paper, an original data decomposition method is proposed, and every ET measuring data are decomposed into two independent new data based on the positive and negative sensing areas of the measuring data. Consequently, the number of total measuring data is extended to twice as many as the number of the original data, thus effectively reducing the ‘ill-posed solution’. On the other hand, an index to measure the ‘soft field’ effect is proposed. The index shows that the decomposed data can distinguish between different contributions of various units (pixels) for any ET measuring data, and can efficiently reduce the ‘soft field’ effect of the ET imaging process. In light of the data decomposition method, a new linear back projection algorithm is proposed to improve the spatial resolution of the ET image. A series of simulations and experiments are applied to validate the proposed algorithm by the real-time performances and the progress of spatial resolutions.