Blindness primarily induces structural alteration in the primary visual cortex (V1). Some studies have found that the early blind subjects had a thicker V1 compared to sighted controls, whereas late blind subjects showed no significant differences in the V1. This implies that the age of blindness onset may exert significant effects on the development of cortical thickness of the V1. However, no previous research used a trajectory of the age of blindness onset-related changes to investigate these effects. Here we explored this issue by mapping the cortical thickness trajectory of the V1 against the age of blindness onset using data from 99 blind individuals whose age of blindness onset ranged from birth to 34 years. We found that the cortical thickness of the V1 could be fitted well with a quadratic curve in both the left (F = 11.59, P = 3 × 10) and right hemispheres (F = 6.54, P = 2 × 10). Specifically, the cortical thickness of the V1 thinned rapidly during childhood and adolescence and did not change significantly thereafter. This trend was not observed in the primary auditory cortex (A1), primary motor cortex (M1), or primary somatosensory cortex (S1). These results provide evidence that an onset of blindness before adulthood significantly affects the cortical thickness of the V1 and suggest a critical period for cortical development of the human V1.
Connectivity-based parcellation using diffusion MRI has been extensively used to parcellate subcortical areas and the association cortex. Connectivity profiles are vital for connectivity-based parcellation. Two categories of connectivity profiles are generally utilized, including global connectivity profiles, in which the connectivity information is from the seed to the whole brain, and long connectivity profiles, in which the connectivity information is from the seed to other brain regions after excluding the seed. However, whether global or long connectivity profiles should be applied in parcellating the primary cortex utilizing connectivity-based parcellation is unclear. Many sources of evidence have indicated that the primary cerebral cortices are composed of structurally and functionally distinct subregions. Because the primary cerebral cortices are rich in local anatomic hierarchical connections and possess high degree of local functional connectivity profiles, we proposed that local connectivity profiles, that is the connectivity information within a seed region of interest, might be used for parcellating the primary cerebral cortices. In this study, the global, long, and local connectivity profiles were separately used to parcellate the bilateral M1, A1, S1, and V1. We found that results using the three profiles were all quite consistent with reported cytoarchitectonic evidence. More importantly, the results using local connectivity profiles showed less inter-subject variability than the results using the other two, a finding which suggests that local connectivity profiles are superior to global and long connectivity profiles for parcellating the primary cerebral cortices. This also implies that, depending on the characteristics of specific areas of the cerebral cortex, different connectivity profiles may need to be adopted to parcellate different areas.
Identifying drug-disease associations is integral to drug development. Computationally prioritizing candidate drug-disease associations has attracted growing attention due to its contribution to reducing the cost of laboratory screening. Drug-disease associations involve different association types, such as drug indications and drug side effects. However, the existing models for predicting drug-disease associations merely concentrate on independent tasks: recommending novel indications to benefit drug repositioning, predicting potential side effects to prevent drug-induced risk, or only determining the existence of drug-disease association. They ignore crucial prior knowledge of the correlations between different association types. Since the Comparative Toxicogenomics Database (CTD) annotates the drug-disease associations as therapeutic or marker/mechanism, we consider predicting the two types of association. To this end, we propose a collective matrix factorization-based multi-task learning method (CMFMTL) in this paper. CMFMTL handles the problem as multi-task learning where each task is to predict one type of association, and two tasks complement and improve each other by capturing the relatedness between them. First, drug-disease associations are represented as a bipartite network with two types of links representing therapeutic effects and non-therapeutic effects. Then, CMFMTL, respectively, approximates the association matrix regarding each link type by matrix tri-factorization, and shares the low-dimensional latent representations for drugs and diseases in the two related tasks for the goal of collective learning. Finally, CMFMTL puts the two tasks into a unified framework and an efficient algorithm is developed to solve our proposed optimization problem. In the computational experiments, CMFMTL outperforms several state-of-the-art methods both in the two tasks. Moreover, case studies show that CMFMTL helps to find out novel drug-disease associations that are not included in CTD, and simultaneously predicts their association types.
With the gradual increase of car ownership in China, license plate recognition plays an important role in intelligent vehicle management system. The existing vehicle number recognition algorithms have slow recognition speed and low accuracy, and are easy to be affected by the light, the position angle of the license plate and the relative fixed position of the camera. The Faster-RCNN based on deep learning locates the license plate, generates the license plate extraction frame and extracts the license plate; Use VGGnet network model to recognize characters, and finally complete the recognition of car license plate. Training and testing are carried out in a large number of data sets. The simulation results show that the network model combining fast RCNN and VGGnet can recognize the license plate with an accuracy of 99.2% in a complex environment, and the recognition accuracy is better than other algorithms.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.