ObjectiveTo observe the possible mechanism of Dihuang Yinzi Decoction （地黄饮子） for improving cognitive dysfunction in Alzheimer's disease （AD） from the perspective of retina. MethodsForty-five APP/PS1 mice （AD model mice） were randomly divided into model group， Dihuang Yinzi Decoction group， and memantine group， with 15 mice in each group， while 15 wild-type C57BL/6J mice from the same litter were used as blank group. Mice in Dihuang Yinzi Decoction group were given Dihuang Yinzi Decoction 30.03 g/（kg·d） by gavage， mice in the memantine group were given memantine hydrochloride 6.1 mg/（kg·d） by gavage， and mice in the blank group and the model group were given normal saline 2 ml/（kg·d） by gavage for 4 consecutive weeks. Fasting blood glucose was measured weekly. After 4 weeks of intervention， oral glucose tolerance test （OGTT） and insulin tolerance test （ITT） were performed； Morris water maze was used to detect the changes in spatial memory ability of mice； glucose oxidase method was used to detect retinal glucose content of mice； enzyme-linked immunosorbent assay （ELISA） was used to detect serum and retinal insulin content of mice， and Homeostatic Model Assessment of insulin resistance （HOMA-IR） was calculated. Hematoxylin-eosin （HE） staining was performed to observe the histopathological changes in the retina， and the retinal thickness and ganglion cell number were counted； protein immunoblotting was performed to detect the retinal pathway-associated proteins ［insulin receptor substrate 1 （IRS1）， phosphorylated insulin receptor substrate 1 （pIRS1）， phosphatidylinositol-3-hydroxykinase （PI3K）， protein kinase B （Akt1）， phosphorylated protein kinase B （pAkt1）］ expression； retinal glucose transporter protein 4 （GLUT4） expression was detected by immunohistochemistry. ResultsCompared with the blank group， fasting blood glucose of mice in the model group at weeks 1， 2， 3， and 4， blood glucose and area under the curve （AUC） at different time point of OGTT and ITT test， fasting serum insulin， and HOMA-IR increased （P＜0.05， P＜0.01）； in the Morris water maze experiment， the escape latency increased from day 3 to day 5， and the number of crossing platforms， the percentage of target quadrant distance， and the percentage of target quadrant time decreased （P＜0.05， P＜0.01）； the outer nuclear layer of the retina became sparse， thinner， and the number of ganglion cells decreases （P＜0.01）； the expression level of retinal glucose increased， while the expression levels of insulin， pIRS1/IRS1， PI3K/β-Actin， pAkt1/Akt1， and GLUT4 proteins decreased （P＜0.01）. Compared with the model group， fasting blood glucose at week 4， blood glucose at each time point of the OGTT and ITT tests AUC decreased （P＜0.05， P＜0.01）， and fasting serum insulin and HOMA-IR decreased （P＜0.05） in Dihuang Yinzi Decoction group； In the Morris water maze test， the escape latency shortened on day 4 and day 5， number of platform crossings， target quadrant distance as a proportion of total distance， and target quadrant movement time as a proportion of total time decreased （P＜0.05， P＜0.01）； retinal pathological changes alleviated， and retinal thickness and ganglion cell number increased （P＜0.01）； retinal glucose content decreased， and retinal pIRS1/IRS1， PI3K/β-Actin， and GLUT4 protein expression elevated （P＜0.05 or P＜0.01）. ConclusionsDihuang Yinzi Decoction can improve cognitive dysfunction of Alzheimer's disease， which may be related to regulating retinal insulin content and insulin signaling pathway.