Gait disturbances are frequent and disabling in advanced Parkinson's disease. These symptoms respond poorly to usual medical and surgical treatments but were reported to be improved by stimulation of the pedunculopontine nucleus. We studied the effects of stimulating the pedunculopontine nucleus area in six patients with severe freezing of gait, unresponsive to levodopa and subthalamic nucleus stimulation. Electrodes were implanted bilaterally in the pedunculopontine nucleus area. Electrode placement was checked by postoperative magnetic resonance imaging. The primary outcome measures were a composite gait score, freezing of gait questionnaire score and duration of freezing episodes occurring during a walking protocol at baseline and one-year follow-up. A double-blind cross-over study was carried out from months 4 to 6 after surgery with or without pedunculopontine nucleus area stimulation. At one-year follow-up, the duration of freezing episodes under off-drug condition improved, as well as falls related to freezing. The other primary outcome measures did not significantly change, nor did the results during the double-blind evaluation. Individual results showed major improvement of all gait measures in one patient, moderate improvement of some tests in four patients and global worsening in one patient. Stimulation frequency ranged between 15 and 25 Hz. Oscillopsia and limb myoclonus could hinder voltage increase. No serious adverse events occurred. Although freezing of gait can be improved by low-frequency electrical stimulation of the pedunculopontine nucleus area in some patients with Parkinson's disease our overall results are disappointing compared to the high levels of expectation raised by previous open label studies. Further controlled studies are needed to determine whether optimization of patient selection, targeting and setting of stimulation parameters might improve the outcome to a point that could transform this experimental approach to a treatment with a reasonable risk-benefit ratio.