The purpose of this study is to evaluate the frequency of central nervous system adverse events (CNS-AE) on dolutegravir (DTG) and non-DTG containing ART, and their reversibility, in the observational prospective SCOLTA cohort. Factors associated with CNS-AE were estimated using a Cox proportional-hazards model. 4939 people living with HIV (PLWH) were enrolled in DTG (n = 1179) and non-DTG (n = 3760) cohorts. Sixty-six SNC-AE leading to ART discontinuation were reported, 39/1179 (3.3%) in DTG and 27/3760 (0.7%) in non-DTG cohort. PLWH naïve to ART, with higher CD4 + T count and with psychiatric disorders were more likely to develop a CNS-AE. The risk was lower in non-DTG than DTG-cohort (aHR 0.33, 95% CI 0.19–0.55, p < 0.0001). One-year follow-up was available for 63/66 PLWH with CNS-AE. AE resolution was reported in 35/39 and 23/24 cases in DTG and non-DTG cohorts, respectively. The probability of AE reversibility was not different based on ART class, sex, ethnicity, CDC stage, or baseline psychiatric disorder. At the same time, a lower rate of event resolution was found in PLWH older than 50 years (p = 0.017). In conclusion, CNS-AE leading to ART discontinuation was more frequent in DTG than non-DTG treated PLWH. Most CNS-AE resolved after ART switch, similarly in both DTG and non-DTG cohorts.