Sequential graph convolutional network and DeepRNN based hybrid framework for epileptic seizure detection from EEG signal

Automated epileptic seizure detection from ectroencephalogram (EEG) signals has attracted significant attention in the recent health informatics field. The serious brain condition known as epilepsy, which is characterized by recurrent seizures, is typically described as a sudden change in behavior caused by a momentary shift in the excessive electrical discharges in a group of brain cells, and EEG signal is primarily used in most cases to identify seizure to revitalize the close loop brain. The development of various deep learning (DL) algorithms for epileptic seizure diagnosis has been driven by the EEG's non-invasiveness and capacity to provide repetitive patterns of seizure-related electrophysiological information. Existing DL models, especially in clinical contexts where irregular and unordered structures of physiological recordings make it difficult to think of them as a matrix; this has been a key disadvantage to producing a consistent and appropriate diagnosis outcome due to EEG's low amplitude and nonstationary nature. Graph neural networks have drawn significant improvement by exploiting implicit information that is present in a brain anatomical system, whereas inter-acting nodes are connected by edges whose weights can be determined by either temporal associations or anatomical connections. Considering all these aspects, a novel hybrid framework is proposed for epileptic seizure detection by combined with a sequential graph convolutional network (SGCN) and deep recurrent neural network (DeepRNN). Here, DepRNN is developed by fusing a gated recurrent unit (GRU) with a traditional RNN; its key benefit is that it solves the vanishing gradient problem and achieve this hybrid framework greater sophistication. The line length feature, auto-covariance, auto-correlation, and periodogram are applied as a feature from the raw EEG signal and then grouped the resulting matrix into time-frequency domain as inputs for the SGCN to use for seizure classification. This model extracts both spatial and temporal information, resulting in improved accuracy, precision, and recall for seizure detection. Extensive experiments conducted on the CHB-MIT and TUH datasets showed that the SGCN-DeepRNN model outperforms other deep learning models for seizure detection, achieving an accuracy of 99.007%, with high sensitivity and specificity.

A(n) (a)symmetric perspective towards task-technology-performance fit in mobile app industry

Purpose

The purpose of this paper is to examine (a)symmetric features of task-technology-performance characteristics that are most relevant to fit, satisfaction and continuance intention of using apps in mobile banking transactions.

Design/methodology/approach

Exploratory factor analysis was used with maximum likelihood extraction and Varimax rotation on a separate sample of 183 mobile banking apps users prior to the main data collection. The theoretical model was tested applying a factor-based structural equation modelling approach to a sample of 250 experienced mobile banking apps users.

Findings

The study unveiled that the task and performance characteristics are more relevant compared to technology characteristics when doing transactions via apps. In addition, the findings uncovered that user satisfaction and continuous intention to use apps stem from the degree of fit in online transactions. The findings of moderation analysis highlighted that users in the lower income group are more concerned about the performance characteristics of banking apps, and there are no differences across age and gender groups. Surprisingly, technology characteristic has a nonlinear nature and this study shows potential boundary conditions of technology characteristics in degree of fit, user satisfaction and continuance intention to use apps.

Practical implications

Findings from the conditional probabilistic queries reveal that with 83.3 per cent of probability, user satisfaction is high when using apps for banking transactions, if the levels of fit, task, performance and technology characteristics are high. Furthermore, with 72 per cent of probability, continuance intention to use apps is high, if the levels of performance and task characteristics are high.

Originality/value

Contributing to task-technology fit theory, this study shows that performance characteristics need to be aligned with task and technology characteristics in order to have better fit when using apps for online banking transactions.

Retraction RETRACTION of "Methylation of the RASSFIA promoter in breast cancer" by Y. Ji, H.H. Jin, M.D. Wang, W.X. Cao, J.L. Bao - Genet. Mol. Res. 15 (2): gmr.15028261 (2016) - DOI: 10.4238/gmr.15028261

The retracted article is: Ji Y, Jin HH, Wang MD, Cao WX, et al. (2016). Methylation of the RASSFIA promoter in breast cancer. Genet. Mol. Res. 15: gmr.15028261. There are significant parts of this article (particularly, in the discussion section) that are copied from "Methylation of HIN-1, RASSF1A, RIL and CDH13 in breast cancer is associated with clinical characteristics, but only RASSF1A methylation is associated with outcome", by Jia Xu, Priya B Shetty, Weiwei Feng, Carol Chenault, Robert C Bast Jr, Jean-Pierre J Issa, Susan G Hilsenbeck and Yinhua Yu, published in BMC Cancer 2012; 12: 243. DOI: 10.1186/1471-2407-12-243. The first paragraphs of both discussions are identical. This is concerning. The abstract and introduction sections have much of their text plagiarized. Overall, there is high plagiarism detected. The GMR editorial staff was alerted and after a thorough investigation, we have strong reason to believe that the peer review process was failure and, after review and contacting the authors, the editors of Genetics and Molecular Research decided to retract the article in accordance with the recommendations of the Committee on Publication Ethics (COPE). The authors and their institutions were advised of this serious breach of ethics.

Methylation of the RASSFIA promoter in breast cancer

Tumor suppressor genes are the key targets of hypermethylation in breast cancer and may therefore lead to malignancy by deregulation of cell growth and division. Our previous pilot study with pairs of malignant and normal breast tissues identified a correlation between RASSFIA gene methylation and breast cancer. To determine the relationship between RASSFIA methylation and breast cancer, we conducted a larger study. We took samples from 108 patients with breast cancer, 28 patients with benign breast tumors, and 33 subjects with normal breast tissues at the Second Affiliated Hospital of Nanjing Medical University at Wuxi between July 2013 and September 2015. We used the samples to investigate methylation levels of the RASSF1A gene for associations with breast cancer. Quantitative real-time polymerase chain reaction (PCR) and methylation-specific PCR were used to investigate the levels of RASSF1A mRNA expression and RASSF1A methylation, respectively. RASSFIA was not expressed in 22 of the 108 breast cancer tissue samples (20.37%), and there was no statistically significant difference (P > 0.05); however, RASSFIA expression was significantly lower than that in the normal breast tissue samples (P < 0.05). Moreover, the methylation rate of the RASSFIA gene promoter was significantly higher in the breast cancer tissues (64.81%) than in the normal breast tissues (18.18%) and benign breast tumors (17.86%) (P < 0.05). High methylation of the RASSF1A gene promoter was an important reason for its downregulation, and the gene played a critical regulated role in the incidence and development of breast cancer.