Objective To observe the retinal toxicity of repeated intravitreal injection with bevacizumab (Avastin) in diabetic rats. Methods Forty male Sprague Dawley (SD) rats were randomly divided into normal group (Group A, 10 rats) and diabetes mellitus group (30 rats). The rats in diabetes mellitus group were induced with streptozotocin injection for diabetic retinopathy model. And then randomly divided into diabetic retinopathy (DR) group (Group B,10 rats), the rats were not intervened; the left eyes of the other 20 rats were intravitreal injected with bevacizumab 3 mu;l (25 mg/ml) for 3 times as experimental group (Group C) ; the right eyes of the 20 rats were not intervened as experimental control group (Group D). 20 days after last intravitreal injection, retinal function was measured by Flicker Electroretinogram (F-ERG);retinal vascular pattern was determined by fluorescence microscopy of ethidium bromide (EB) stained retinal flat mounts; retinal morphological changes were determined by light microscope on hematoxylin-eosin (HE) stained sections;Thy-1 and VEGF expression was measured by immunohistochemistry staining. Results F-ERG showed that, the differences of a-and b-waves, the b-wave amplitude and the Ops-wave amplitude in the implicit time between group A, B, C and D were significant (F=33.165, 36.162,19.955, 23.243; P＝0.000); the differences of a-wave amplitude between group A,B,C and D was not significant (F=0.097,P＝0.961). Retinal blood vessel pattern was normal in Group A; retinal vascular vessels were tortuous and irregularly expanded in Group B; retinal vascular vessels of Group C were regular and thinner than Group A; microaneurysm were showed in Group D. Light microscope displayed that the layers of the rat retina of Group A were regular, the retinal architectures of Group B were irregular, the retinal layers were regular in Group C, the retinal layers were irregular in Group D. Immunohistochemistry staining discovered that Thy-1 and VEGF were mainly expressed in ganglion cell layer（GCL）. Conclusion Repeated intravitreal injection of bevacizumab is toxic to retina of diabetes mellitus rats.
In clinical, manually scoring by technician is the major method for sleep arousal detection. This method is time-consuming and subjective. This study aimed to achieve an end-to-end sleep-arousal events detection by constructing a convolutional neural network based on multi-scale convolutional layers and self-attention mechanism, and using 1 min single-channel electroencephalogram (EEG) signals as its input. Compared with the performance of the baseline model, the results of the proposed method showed that the mean area under the precision-recall curve and area under the receiver operating characteristic were both improved by 7%. Furthermore, we also compared the effects of single modality and multi-modality on the performance of the proposed model. The results revealed the power of single-channel EEG signals in automatic sleep arousal detection. However, the simple combination of multi-modality signals may be counterproductive to the improvement of model performance. Finally, we also explored the scalability of the proposed model and transferred the model into the automated sleep staging task in the same dataset. The average accuracy of 73% also suggested the power of the proposed method in task transferring. This study provides a potential solution for the development of portable sleep monitoring and paves a way for the automatic sleep data analysis using the transfer learning method.