目的:探讨血清蛋白因子水平与精神分裂症患者临床症状和认知障碍的相关性及对认知障碍缺陷程度的预测作用，为临床评估精神分裂症认知受损严重程度及预后提供辅助方法。方法:选择2017年9月至2019年4月在昆明医科大学第一附院确诊的71例精神分裂症患者作为患者组，以同一医院体检中心的65例健康志愿者作为对照组。使用ELISA方法检测患者组与对照组外周血肿瘤坏死因子α(tumor necrosis factor α，TNF-α)、脑源性神经营养因子(brain-derived neurotrophic factor，BDNF)、钙结合蛋白β(calcium-binding protein β，S100β)浓度。使用系统评估工具-MATRICS认知功能成套测验(MATRICS consensus cognitive battery，MCCB)进行认知功能评估。采用阳性和阴性症状量表(positive and negative syndrome scale，PANSS)评定患者临床症状。使用SPSS 22.0进行统计分析，组间比较采用独立样本n t检验，血清蛋白因子水平与认知功能和临床症状的关系采用Pearson相关分析。以血清蛋白因子浓度和PANSS总分作为自变量，MCCB各认知因子缺陷程度作为因变量，建立Bayes判别函数，对精神分裂症认知功能障碍严重程度进行客观的预测、评价和验证。n 结果:血清TNF-α[(63.2±25.2)pg/L，(31.4±14.3)pg/L]、S100β[(68.0±26.4)pg/L，(47.3±20.2)pg/L]浓度患者组高于对照组，BDNF浓度低于对照组[(2 517.8±1 140.2)pg/L，(5 202.2±447.2)pg/L]，均差异有统计学意义(均n P<0.05)。Bayes判别函数模型中认知障碍严重程度回顾性检验四个认知因子的正确判别率为：信息处理速度69.0%，词语学习63.4%,推理和问题解决76.1%，视觉学习73.2%。交叉检验正确判别率分别为：信息处理速度66.2%，词语学习60.6%,推理和问题解决73.2%，视觉学习66.2%。n 结论:精神分裂症血清蛋白因子TNF-α、BDNF、S100β水平和临床症状评分与认知障碍存在不同程度相关性，Bayes判别模型对精神分裂症认知功能障碍严重程度有较高的正确判别率，精神分裂症相关蛋白因子水平和临床症状评分对认知功能缺陷程度可能具有预测作用，为临床疗效评估及预后分析提供较客观的依据。“,”Objective:To explore the correlation between serum protein factor level and clinical symptoms and cognitive impairment in patients with schizophrenia and to predict the degree of cognitive impairment, so as to provide an auxiliary method for clinical evaluation of cognitive impairment severity and prognosis of schizophrenia.Methods:From September 2017 to April 2019, 71 schizophrenic patients diagnosed in the First Affiliated Hospital of Kunming Medical University were selected as the patient group, and 65 healthy volunteers from the physical examination center of the same hospital were selected as the control group.The concentrations of tumor necrosis factor α(TNF-α), brain-derived neurotrophic factor(BDNF) and calcium-binding protein β(S100β) in peripheral blood were detected by ELISA method.Cognitive function was evaluated by MATRICS consensus cognitive battery(MCCB) cognitive assessment.The clinical symptoms of patients were evaluated by positive and negative syndrome scale(PANSS) scale.SPSS 20.0 software was used for statistical analysis, independent sample n t-test was used for comparison between groups, and Pearson correlation analysis was used for the relationship between serum protein factor level and cognitive function and clinical symptoms.In order to objectively predict, evaluate and verify the severity of cognitive impairment in schizophrenia, Bayes discriminant function was established with serum protein factor concentration and PANSS total score as independent variables and the defect degree of cognitive factors in MCCB as dependent variables.n Results:The serum TNF-α((63.2±25.2)pg/Ln vs (31.4±14.3)pg/L) and S100β((68.0±26.4)pg/Ln vs (47.3±20.2)pg/L) concentrations in the patient group were higher than those in the control group.The concentration of serum BDNF in the patient group was lower than that in the control group ((2 517.8±1 140.2)pg/L n vs (5 202.2±447.2)pg/L), and the difference was statistically significant (n P=0.000). In the retrospective test of cognitive impairment severity in Bayes discriminant function model, the correct discrimination rates of four cognitive factors were speed of processing(SoP) 69.0%, Verbal learning(VeL) 63.4%, reasoning and problem solving(RPS) 76.1% and visual learning(ViL) 73.2%.The correct discrimination rates of cross-examination were SoP 66.2%, VeL 60.6%, RPS 73.2%, ViL 66.2.n Conclusion:The levels of serum protein factors TNF-α, BDNF and S100β and clinical symptom scores of schizophrenia have different degrees of correlation with the severity score of cognitive impairment.Bayes discriminant function model has higher correct discrimination rate for the severity of cognitive impairment of schizophrenia.It is found that the levels of schizophrenia-related protein factors and clinical symptom scores may have predictive effect on the severity of cognitive impairment, providing a more objective basis for the clinical efficacy evaluation of schizophrenia patients.