Downbeat nystagmus (DBN), the most common form of acquired fixation nystagmus, is often caused by cerebellar degeneration, especially if the vestibulo-cerebellum is involved. The upward ocular drift in DBN has a spontaneous and a vertical gaze-evoked component. Since cerebellar involvement is suspected to be the underlying pathomechanism of DBN, we tested in 15 patients with DBN whether the application of the potassium-channel blocker 4-aminopyridine (4-AP), which increases the excitability of cerebellar Purkinje cells as shown in animal experiments, reduces the vertical ocular drift leading to nystagmus. Fifteen age-matched healthy subjects served as the control group. 4-AP may affect spontaneous drift or gaze-evoked drift by either enhancing visual fixation ability or restoring vision-independent gaze holding. We therefore recorded 3D slow-phase eye movements using search coils during attempted fixation in nine different eye positions and with or without a continuously visible target before and 45 min after ingestion of 10mg 4-AP. Since the effect of 4-AP may depend on the associated etiology, we divided our patients into three groups (cerebellar atrophy, n = 4; idiopathic DBN, n = 5; other etiology, n = 6). 4-AP decreased DBN during gaze straight ahead in 12 of 15 patients. Statistical analysis showed that improvement occurred predominantly in patients with cerebellar atrophy, in whom the drift was reduced from -4.99 +/- 1.07 deg/s (mean +/- SE) before treatment to -0.60 +/- 0.82 deg/s afterwards. Regression analysis of slow-phase velocity (SPV) in different eye positions revealed that vertical and horizontal gaze-evoked drift was significantly reduced independently of the patient group and caused perfect gaze holding on the average. Since the observed improvements were independent of target visibility, 4-AP improved fixation by restoring gaze-holding ability. All in all, the present study demonstrates that 4-AP has a differential effect on DBN: drift with gaze straight ahead was predominantly reduced in patients with cerebellar atrophy, but less so in the remaining patients; 4-AP on the average improved neural integrator function, i.e. gaze-evoked drift, regardless of etiology. Our results thus show that 4-AP was a successful treatment option in the majority of DBN patients, possibly by increasing Purkinje cell excitability in the cerebellar flocculi. It may work best when DBN is associated with cerebellar atrophy. Furthermore, 4-AP may be a promising treatment option for patients with a dominant gaze-evoked component of nystagmus, regardless of its etiology.