In healthy humans, the cortical brain rhythm (or electroencephalogram, EEG), shows specific mu (∼8-12 Hz) and beta (∼16-24 Hz) band patterns in the cases of both real and imaginary motor movements. As cerebellar ataxia is associated with impairment of precise motor movement control as well as motor imagery, ataxia is an ideal model system in which to study the role of the cerebellocortical circuit in rhythm control. We hypothesize that the EEG characteristics of ataxic patients differ from those of controls during the performance of a Brain-Computer Interface (BCI) task. EEGs were recorded from four patients with cerebellar ataxia and six healthy controls. Subjects were cued to imagine relaxation or motor movement, while an EEG-based BCI translated motor intention into a visual feedback signal through real-time detection of motor imagery states. Ataxia and control subjects showed a similar distribution of mu power during cued relaxation. During cued motor imagery, however, the ataxia group showed significant spatial distribution of the response, while the control group showed the expected decrease in mu-band power (localized to the motor cortex). This pilot study suggests that impairment of the cerebellocortical control network is associated with spatial spreading of the normal event-related synchronization of motor cortical areas during relaxation. The mechanism of this association, whether degenerative or compensatory, bears further investigation. Use of BCI has important implications for our basic understanding of motor imagery and control, as well as clinical development of brain-computer interface as an assistive or rehabilitative technology.