Search

Predicting the progression from mild cognitive impairment (MCI) to Alzheimer's disease (AD) is critical for early intervention. Towards this end, various deep learning models have been applied in this domain, typically relying on structural magnetic resonance imaging (sMRI) data from a single time point whereas neglecting the dynamic changes in brain structure over time. Current longitudinal studies inadequately explore disease evolution dynamics and are burdened by high computational complexity. This paper introduces a novel lightweight 3D convolutional neural network specifically designed to capture the evolution of brain diseases for modeling the progression of MCI. First, a longitudinal lesion feature selection strategy is proposed to extract core features from temporal data, facilitating the detection of subtle differences in brain structure between two time points. Next, to refine the model for a more concentrated emphasis on lesion features, a disease trend attention mechanism is introduced to learn the dependencies between overall disease trends and local variation features. Finally, disease prediction visualization techniques are employed to improve the interpretability of the final predictions. Extensive experiments demonstrate that the proposed model achieves state-of-the-art performance in terms of area under the curve (AUC), accuracy, specificity, precision, and F1 score. This study confirms the efficacy of our early diagnostic method, utilizing only two follow-up sMRI scans to predict the disease status of MCI patients 24 months later with an AUC of 79.03%.

Collaboratively optimizing operational parameters through leveraging accumulated production experience is an innovative approach to reducing energy consumption in aluminum electrolysis cells (AECs). Due to the dynamic heterogeneity of various AECs, an explicit evolutionary multitasking (EMT) framework capable of incorporating different optimizers, has the potential to tackle this challenge effectively. However, there is a notable gap in theoretical research on multitasking collaborative evolutionary algorithms specifically applied to AECs. Meanwhile, existing explicit EMT algorithms often overlook the intertask correlation of feature information extracted in isolation from individual tasks. These issues significantly limit the development of synergistic effects in multitasking optimization for addressing parameter design in AECs. To address these limitations, this work proposes an explicit evolutionary framework with multitasking feature fusion (EMFF). This framework thoroughly considers the potential connections among feature information from different tasks. It achieves effective knowledge transfer by the design of a unique multitasking feature fusion mechanism, which enhances the information value of source tasks for target tasks. Furthermore, a transfer individual derivation (TID) strategy is introduced to ensure the rapid evolution of critical knowledge. Finally, comprehensive components and designed process are presented. Experimental results demonstrate EMFF's exceptional performance in various benchmark tests and real-world AEC parameter optimization cases.

There are plenty of risk factors associated with splenic hilar lymph node metastasis (SHLNM) in patients with advanced gastric cancer (AGC). Whereas, their main influencing factors have not reached a consensus yet. The aim of the study is to investigate the related clinicopathological factors influencing SHLNM in AGC. A retrospective study was performed to investigate 150 patients who underwent D2 curative partial or total gastrectomy for gastric carcinoma from January 2007 to November 2012. Clinicopathological factors were analyzed by univariate and multivariate analysis. A total of 10.7% (16/150) of the patients had SHLNM. The overall ratio of metastatic lymph node (positive lymph nodes/lymph nodes harvested) in the splenic hilum was 17.5% (38/217). Univariate analysis results showed SHLNM was related with depth of invasion, tumor grade, tumor size, tumor location and Bormann type, with significant difference (P<0.05); Multivariate analysis demonstrated that SHLNM was related with depth of invasion and tumor size, with significant difference (P<0.05). Consequently, depth of invasion, tumor grade, tumor size, tumor location and Bormann type were associated with SHLNM in AGC, meanwhile depth of invasion and tumor size are independent risk factors. Preoperative predicting risk factors of SHLNM greatly benefits making more rational surgical scheme of treating AGC.

With community character indices and principal component analysis, this paper studied the variation of insect community composition and structure in Chinese prickly ash plantations at four altitudes in Yunnan Province, and analyzed the temporal patterns of the variation and its affecting factors. The results showed that with the increase of altitude, the species diversity, abundance and evenness of the community decreased gradually, while the population number had a certain increase. Such a variation was more obvious with time. The factor dominating the variation was the gradual change of the species and individuals numbers of natural enemy sub-community to those of pest sub-community.