The human vestibular cortex: functional anatomy of OP2, its connectivity and the effect of vestibular disease

Abstract

Area OP2 in the posterior peri-sylvian cortex has been proposed to be the core human vestibular cortex. We investigated the functional anatomy of OP2 and adjacent areas (OP2+) using spatially constrained independent component analysis (ICA) of functional magnetic resonance imaging (fMRI) data from the Human Connectome Project. Ten ICA-derived subregions were identified. OP2+ responses to vestibular and visual motion were analyzed in 17 controls and 17 right-sided vestibular neuritis patients who had previously undergone caloric and optokinetic stimulation during fMRI. In controls, a posterior part of right OP2+ showed: (i) direction-selective responses to visual motion and (ii) activation during caloric stimulation that correlated positively with perceived self-motion, and negatively with visual dependence and peak slow-phase nystagmus velocity. Patients showed abnormal OP2+ activity, with an absence of visual or caloric activation of the healthy ear and no correlations with vertigo or visual dependence-despite normal slow-phase nystagmus responses to caloric stimulation. Activity in a lateral part of right OP2+ correlated with chronic visually induced dizziness in patients. In summary, distinct functional subregions of right OP2+ show strong connectivity to other vestibular areas and a profile of caloric and visual responses, suggesting a central role for vestibular function in health and disease.

Keywords: perception; vestibular cortex; vestibular neuritis; visual.

Fig. 1

Overview of methods. (A) Spatially constrained ICA of area OP2⁺ within Human Connectome Project task-free data produced spatial maps of subregions in each hemisphere. Between-subject reproducibility was maximized and reproducibility was further enhanced by ranking and averaging by spatial reproducibility. The whole-brain connectivity of each subregion was then estimated by dual regression. The regional anatomy of right OP2⁺ (in blue) is shown on an inflated brain relative to areas OP1, OP3, the retroinsular cortex (Ri), and the insular granular cortex (Ig); the images are adapted with permission from Indovina et al. (Indovina et al. 2020) (B) CALORIC and VISUAL tasks were performed. The trace on the left is an example of right-beating nystagmus during cool caloric irrigation of the left ear; the orange line depicts SPV, which is defined as the gradient of the slow phase of nystagmus; the adjacent trace is a typical profile of the evolution of SPV over time following caloric irrigation at time zero; the time periods (10–30 and 30–50 s) that define the caloric contrast are illustrated.

Fig. 2

OP2⁺ Subregions are connected to known vestibular areas. Blue-shaded cells indicate significant functional connectivity between OP2⁺ subregions and areas that consistently activated following vestibular stimulation as reported in zu Eulenburg et al’.s (zu Eulenburg et al. 2012) meta-analysis. Family-wise error rate corrected 1 minus P values are illustrated, as determined by nonparametric permutation testing of participant and subregion-specific whole-brain connectivity maps using randomize, within FSL (Winkler et al. 2014). A value of > 0.95 is considered significant. Cells with nonsignificant connectivity are in the background color.

Fig. 3

Subregions within OP2⁺ identified from ICA. (A) Right OP2⁺ subregion images (blue) overlaid on right OP2⁺ mask (red, from Jülich atlas). The core/central area of the OP2⁺ mask (P > 0.5 in Jülich atlas) is shown in yellow. Brain networks with significant functional connectivity to each OP2⁺ subregion are shown adjacent (significant voxels red-yellow, P < 0.05). (B) Left OP2⁺ subregions are shown in green with adjacent functional connectivity maps. R = right. L = left.

Fig. 4

Common patterns of vestibular connectivity from OP2⁺ subregions. (A) Spatial similarity matrix for the whole-brain connectivity of right OP2⁺ subregions as measured by the Dice coefficient. Red square highlights 3 subregions with similar networks (R1 [anterior OP2⁺], R2 [lateral OP2⁺], and R3 [posterior OP2⁺]). These 3 subregions are illustrated in blue, on a background of the OP2⁺ mask in red. The core/central area of the OP2⁺ mask (P > 0.5 in Jülich atlas) is shown in yellow. Adjacent is a map of common functional connectivity from the 3 subregions (yellow, P < 0.05). (B) Spatial similarity matrix, OP2⁺ subregions (in green) and common functional connectivity for left OP2⁺.

Fig. 5

Bilateral OP2⁺ activation during caloric irrigation. (A) Areas of significant activation during caloric irrigation in healthy controls (red-yellow). Activation is seen in right and left OP2⁺. Masks of right (blue) and left OP2⁺ (green) are shown, defined using the Jülich atlas. (B). Vestibular activation produced in OP2⁺ by acoustic, caloric, or galvanic stimuli (black, courtesy of Dr Christophe Lopez, Aix-Marseille Université (Marseille, France); Lopez et al. 2012).

Fig. 6

Caloric and visual motion responses in OP2⁺. (A) Caloric responses in healthy controls and patients with vestibular neuritis within right and left OP2⁺ subregions. (B) Visual motion responses in healthy controls and patients with vestibular neuritis. Larger bar charts show OP2⁺ subregion responses for motion selectivity (right > left visual motion). Smaller bar charts show specific responses to rightward and leftward motion (motion > static visual stimulation). Asterisk indicates significant activation relative to baseline, FDR-corrected.

Fig. 7

Correlation between subregion caloric response and vertigo, nystagmus SPV, visual dependence, and situational vertigo. (A) Scatter plots of vertigo and peak nystagmus SPV versus the caloric response in R3 [posterior OP2⁺] in healthy controls (HC) and patients with vestibular neuritis (VN); data points were obtained in each of 4 stimulus conditions. Inset bar charts show mean peak nystagmus SPV (in degrees per second) and mean vertigo; error bars show the 95% confidence intervals of the respective means. (B) Scatter plot of situational vertigo and R2 [lateral OP2⁺] caloric activation in VN; the line of best fit is for illustration only. (C) (i) Scatter plots of visual dependence and R3 [posterior OP2⁺] caloric response; (ii) illustration of rod and disc task from which visual dependence was calculated. The subjective visual vertical is measured when participants align a central rod overlying a static background (S); this is repeated during clockwise (M_C) and counterclockwise (M_CC) rotation of the visual background; visual dependence is calculated as shown. ^*P < 0.05 for caloric response predictor; #P < 0.05 for Spearman correlation.