Objective To investigate the effect of electroacupuncture at different points on auditory brainstem response (ABR) in a rat model of sodium salicylate-induced tinnitus.Method Forty-one male Wistar rats were randomly allocated to saline control (saline), sodium salicylate model (SA), electroacupuncture at Tinggong+ Yifeng (EA), electroacupuncture at Waiguan+ Zhongzhu (AA) and electroacupuncture at Zusanli+ Sanyinjiao (LA) groups. The saline group consisted of five rats and each of the other groups, nine rats. The model was made by intraperitoneal injection of sodium salicylate 275 mg/kg. The saline control group was injected with a corresponding volume of saline. Various acupoint electroacupuncture groups were given electroacupuncture at bilateral Tinggong+ Yifeng, Waiguan+ Zhongzhu and Zusanli+ Sanyinjiao, respectively, at 30 min after model making. Electroacupuncture lasted 30 min. The ABRs were recorded before model making and once every one hour for five consecutive hours after model making. The stimulus sounds were short clicks and tone bursts of frequencies of 4, 8, 16 and 32 kHz. The ABR threshold was used as an assessment index.Results Under the condition of clicks, there was a statistically significant difference in the ABR threshold at one to five hours after model making between the SA, EA, AA or LA group and the saline group (P<0.05), at two to five hours after model making between the EA or AA group and the SA group (P<0.05) and at one hour after model making between the LA and SA groups (P<0.05). Under the conditions of 4, 8 and 16 kHz, there was a statistically significant difference in the ABR threshold at one to five hours after model making between the SA, EA, AA or LA group and the saline group (P<0.05). Under the condition of 32 kHz, there was a statistically significant difference in the ABR threshold at one to five hours after model making between the SA, AA or LA group and the saline group (P<0.05). Under the condition of 4 kHz, there was a statistically significant difference in the ABR threshold at two to five hours after model making between the EA and SA groups (P<0.05) and at four to five hours after model making between the AA and SA groups (P<0.05). Under the condition of 8 kHz, there was a statistically significant difference in the ABR threshold at two to five hours after model making between the EA or AA group and the SA group (P<0.05). Under the condition of 16 kHz, there was a statistically significant difference in the ABR threshold at two, four and five hours after model making between the EA and SA groups (P<0.05). Under the condition of 32 kHz, there was a statistically significant difference in the ABR threshold at one to five hours after model making between the EA and SA groups (P<0.05).Conclusion Electroacupuncture at both periauricular and forelimb points can improve the ABR threshold in sodium salicylate-treated rats. The effect of electroacupuncture at periauricular points is superior to that at forelimb points.