Functional MRI of galvanic vestibular stimulation.

International audienceThe cortical processing of vestibular information is not hierarchically organized as the processing of signals in the visual and auditory modalities. Anatomic and electrophysiological studies in the monkey revealed the existence of multiple interconnected areas in which vestibular signals converge with visual and/or somatosensory inputs. Although recent functional imaging studies using caloric vestibular stimulation (CVS) suggest that vestibular signals in the human cerebral cortex may be similarly distributed, some areas that apparently form essential constituents of the monkey cortical vestibular system have not yet been identified in humans. Galvanic vestibular stimulation (GVS) has been used for almost 200 years for the exploration of the vestibular system. By contrast with CVS, which mediates its effects mainly via the semicircular canals (SCC), GVS has been shown to act equally on SCC and otolith afferents. Because galvanic stimuli can be controlled precisely, GVS is suited ideally for the investigation of the vestibular cortex by means of functional imaging techniques. We studied the brain areas activated by sinusoidal GVS using functional magnetic resonance imaging (fMRI). An adapted set-up including LC filters tuned for resonance at the Larmor frequency protected the volunteers against burns through radio-frequency pickup by the stimulation electrodes. Control experiments ensured that potentially harmful effects or degradation of the functional images did not occur. Six male, right-handed volunteers participated in the study. In all of them, GVS induced clear perceptions of body movement and moderate cutaneous sensations at the electrode sites. Comparison with anatomic data on the primate cortical vestibular system and with imaging studies using somatosensory stimulation indicated that most activation foci could be related to the vestibular component of the stimulus. Activation appeared in the region of the temporo-parietal junction, the central sulcus, and the intraparietal sulcus. These areas may be analogous to areas PIVC, 3aV, and 2v, respectively, which form in the monkey brain, the "inner vestibular circle". Activation also occurred in premotor regions of the frontal lobe. Although undetected in previous imaging-studies using CVS, involvement of these areas could be predicted from anatomic data showing projections from the anterior ventral part of area 6 to the inner vestibular circle and the vestibular nuclei. Using a simple paradigm, we showed that GVS can be implemented safely in the fMRI environment. Manipulating stimulus waveforms and thus the GVS-induced subjective vestibular sensations in future imaging studies may yield further insights into the cortical processing of vestibular signals