Cognitive impairment (CI) affects 40–65% of multiple sclerosis (MS) patients. This study attempted evaluating the effects of fingolimod and interferon beta-1b (IFN β-1b) on CI progression, magnetic resonance imaging (MRI) and clinical outcomes in relapsing–remitting MS (RRMS) patients over 18 months. The GOLDEN study was a pilot study including RRMS patients with CI randomised (2:1) to fingolimod (0.5 mg daily)/IFN β-1b (250 µg every other day). CI was assessed via Rao’s Brief Repeatable Battery and Delis–Kaplan Executive Function System test. MRI parameters, Expanded Disability Status Scale scores and relapses were measured. Overall, 157 patients were randomised, of whom 30 discontinued the study (fingolimod, 8.49%; IFN β-1b, 41.18%; p ≤ 0.0001). Patients randomised to fingolimod had more severe clinical and MRI disease characteristics at baseline compared with IFN β-1b. At Month (M) 18, both treatment groups showed improvements in all cognitive parameters. At M18, relapse rate, total number and volume of T2/T1 gadolinium-enhancing lesions were higher with IFN β-1b, as well as the percentage brain volume change during the study. Safety and tolerability of both treatments were similar to previous studies. Both treatments showed improvements in cognitive parameters. Fingolimod demonstrated significantly better effects on MRI parameters and relapse rate. Imbalance in baseline characteristics and the drop-out pattern may have favoured IFN β-1b. A longer duration trial may be needed to observe the complete expression of differential effects on CI scales reflecting the between-groups differences on MRI. Although limited in size, the GOLDEN study confirms the favourable benefit–risk profile of fingolimod reported in previous studies. © 2017, The Author(s).