ABSTRACT
This article systematically reviews the research results related to the machine learning based suicide ideation prediction on social networking platforms, so as to provide references for group and individual suicide prediction. This article will address the current states (issues of algorithm accuracy and efficiency, privacy leakage and stigma) and limitations of machine learning based suicide prediction on different platforms (light blogging, acquaintance social platforms, forums, picture and video sharing applications and clinical databases).
ABSTRACT
Paradoxical insomnia is a common clinical phenomenon.Due to the lack of comprehensive understanding of its pathogenesis, there were wide variability in diagnostic criteria and treatment outcomes, thus interfering with the correct statistics of its epidemiological data.Through the summary of previous studies, combination with our own research on the pathogenesis of chronic insomnia, we propose that the differences in the subjective and objective assessment of paradoxical insomnia are related to the asynchronous inactivation of the thalamus and cortex during the sleep initiation phase and the cortical activation sequence and intensity disorders in each region after sleep.Furthermore, the diagnosis of paradoxical insomnia and clinical characteristics are described according to the related pathogenesis and a targeted treatment strategy is formulated.Paradoxical insomnia is a type of insomnia with specific pathophysiological mechanism.