OBJECTIVE Subthalamic nucleus (STN) deep brain stimulation (DBS) for Parkinson's disease (PD) is traditionally performed under local anaesthetic (LA). STN visualization and routine validation of electrode location on stereotactic MRI may allow surgery under general anaesthesia (GA). This study compares the clinical outcome of MRI-guided STN DBS performed under LA or GA in a consecutive patient series. METHODS Unified Parkinson's Disease Rating Scale motor scores (UPDRS-III) in 14 GA patients (mean age 56.1 years, disease duration 13.8 years) were compared with those of 68 LA patients (mean age 57.5 years, disease duration 15.2 years). RESULTS Baseline UPDRS-III were worse in the GA group, both on medication (GA: 20.9 ± 10.8; LA: 13.2 ± 7.8, p < 0.01) and off medication (GA: 57.9 ± 16.6; LA: 48.2 ± 15.7, p < 0.05). On stimulation off medication motor scores significantly improved in both groups (GA: 27.3 ± 11.8, mean 12-month follow-up; LA: 23.7 ± 11.8, mean 14-month follow-up). The percentage improvement was similar in both groups (GA: 52.8%; LA: 50.8%, p = 0.96). Transient surgical complications occurred in 1 GA and 7 LA patients. CONCLUSIONS MRI-guided STN DBS under GA with routine stereotactic verification of lead location did not have a negative effect on efficacy or safety. Surgery under GA is a viable option in patients who would find it hard to tolerate awake surgery due to disease severity, comorbidities or anxiety.