Performance of memory task indicators for assessment of the risk of depressive disorders / 预防医学

ABSTRACT

Objective@#To evaluate the risk of depressive disorders using memory task indicators, so as to provide insights into clinical assessment of depressive disorders.@*Methods@#A total of 68 patients with depressive disorders undergoing treatments in the departments of psychiatrics and clinical psychology in a tertiary hospital during the period from January to September, 2021 were enrolled as the case group, while a total of 31 hospital employees, social volunteers and university students served as controls. The error rate and response time of classical memory task experiments were compared between the two groups, including implicit memory, short-term memory and working memory tasks. In addition, the predictive indicators of depressive disorders were identified using multivariable logistic regression analysis and receiver operative characteristics (ROC) curve.@*Results@#The case group included 29 men and 39 women and had a mean age of (24.12±7.40) years, including 46 subjects with an educational level higher than diploma. The control group included 15 men and 16 women and had a mean age of (26.45±6.65) years, including 23 subjects with an educational level higher than diploma. Multivariable logistic regression analysis showed significant associations of age of >18 years (OR=3.431, 95%CI 1.259-9.350), error rate of 2-back task (OR=1.056, 95%CI 1.016-1.097) and error rate of short-term memory tasks (OR=1.078, 95%CI 1.009-1.152) with the development of depressive disorders. ROC curve analysis showed that the error rate of 2-back tasks showed an area under the ROC curve (AUC) of 0.730 (95%CI 0.630-0.831), cutoff of 22.5%, sensitivity of 42.6% and specificity of 93.5% for prediction of the risk of depressive disorders, and the error rate of short-term memory tasks showed an AUC of 0.717 (95%CI 0.605-0.829), cutoff of 23.5%, sensitivity of 67.6% and specificity of 71.0% for prediction of the risk of depressive disorders. In addition, the combination of the error rate of 2-back tasks and the error rate of short-term memory tasks showed an AUC of 0.829 (95%CI 0.734-0.923), sensitivity of 75.0% and specificity of 80.6% for prediction of the risk of depressive disorders.@*Conclusion@#Short-term and working memory task indicators are feasible for assessment of the risk of depressive disorders.