The vestibular system subserves eye, head, and body coordination in space, postural control, and the perception of verticality and self-motion (1). Its involvement in “higher” cortical functions such as spatial memory, orientation, and navigation, as well as cognition is well recognized (2). In contrast to other sensory modalities, cortical vestibular areas are multisensory; they closely interact with visual and somatosensory input.

Spatial neglect is a heterogeneous disorder of spatial and non-spatial attention which leads to reduced awareness of stimuli in the contralesional hemispace (3, 4). It can affect the representation of one’s own body, the surroundings, or abstract space, such as imagined space, time lines, and number space (3). Imaging evidence suggests that spatial neglect should be regarded as a network disorder of spatial and non-spatial attentional systems (5, 6).

Typically a spatial hemineglect is caused by an ischemic stroke within the middle cerebral artery (MCA) territory. The lesioned areas in neglect patients are generally located in the perisylvian regions and the superior and middle temporal gyrus particularly within the right hemisphere, the inferior parietal lobule (IPL) and less frequently the inferior prefrontal cortex (4).

Several findings support the view that the vestibular system contributes to the manifestation of the neglect syndrome. For example, the causative lesions overlap with the vestibular cortical network, which characteristically shows a strong preponderance of the right hemisphere in right-handers (5, 7, 8). Further, vestibular stimulation can temporarily improve neglect symptoms; this was first demonstrated with caloric irrigation by Rubens (9). Thus, the integrity of cortical vestibular function seems to be important for maintaining maps for spatial orientation (10–12).

There is growing evidence that spatial orientation relies on vestibular signals (11, 13, 14), but their exact contribution as to the particular cortical structures and interconnections to other sensory cortical areas has yet to be demonstrated. It is unclear which areas of the vestibular cortical network are involved in the acute stage and how the spontaneous recovery is achieved within the multisensory attentional system.

The observation that vestibular stimulation, e.g., by calories, can transiently improve the disrupted awareness of visual stimuli in the contralateral neglected hemifield raises the question as to whether the spontaneous recovery over months is mainly based on a functional reorganization of the multisensory attentional network rather than a structural and functional restoration of the ischemic neuronal assemblies.

The aim of the current study was to evaluate, first, the acute beneficial effect of vestibular stimulation on neglect and, second, the spontaneous recovery from neglect within days to months. Methodologically we measured changes in resting-state functional connectivity (fc) of the core regions of the central vestibular and visual networks within the affected hemisphere and between the two hemispheres during the acute and chronic stages of the condition. The following questions were addressed in a patient with a typical spatial hemineglect during the acute stage compared to 6 months later when the patient had completely recovered: 1. is the improvement of spatial hemineglect after caloric stimulation in the acute stage associated with changes in fc within the cortical vestibular network, and if so, which areas are involved? 2. Does caloric stimulation lead to changes in cortical network interactions between vestibular, attentional, and/or visual networks? 3. Are these changes in the acute stage after vestibular stimulation similar to or different from the changes seen in fc after recovery of the initial deficit 6 months later? It is hypothesized that the changes in fc elicited by caloric vestibular stimulation in the acute stage involve connectivity patterns that later provide the basis for a robust amelioration in the chronic stage.

A 62-year-old, right-handed male patient presented to the emergency department of the University Hospital of the Ludwig-Maximilians University, Munich, Germany, with signs of confusion, mild left hemiparesis, and peri-personal visuospatial neglect of the left hemifield. The confusion and hemiparesis were transient, lasting only a few hours, whereas the hemispatial neglect syndrome persisted. The patient completed paper–pencil neglect tests (the Bells test, Albert’s test), which disclosed pathologic scores consistent with spatial neglect [for detailed description of the tests, see in Ref. (15)]. The patient omitted all targets on the left side in the Albert’s test and all but three objects on the left side in the Bells test. He also exhibited a slowing of saccades toward the left hemifield and reduced optokinetic nystagmus to the left. A detailed neuro-orthoptic examination revealed no further deficits. Magnetic resonance imaging (MRI) showed an infarction in the parietal part of the right MCA territory which affected the IPL with the angular gyrus bordering and partly including the intraparietal sulcus (IPS), V5, and the temporo-occipital junction, but sparing the posterior insula, parietal opercular cortex, and posterior parietal cortex (PPC) (Figure 1).

In the acute phase the patient showed visuospatial hemineglect to the left on the behavioral level (see above) with neglect test scores that significantly improved in Albert’s test (no omission) after caloric irrigation.

Up to now there have been only a few fc studies which focused on inter- and intrahemispheric resting-state networks in neglect patients. In one resting-state fMRI experiment a reduced activation in distinct visual (lateral occipital cortex) and parietal cortex areas of the right hemisphere was seen as specific correlates of neglect, suggesting a disturbed modulation of visual processing by the parietal cortex (20). Furthermore, in this study an asymmetry of parietal cortex activation (relative hyperactivation of the left) was described in neglect as well as in stroke control patients without neglect (20). Recent fMRI studies have shown decreased interhemispheric connectivity of homologous regions in the acute stage of neglect across multiple resting-state networks, intrahemispheric decreases of anticorrelation between default mode and ventral attention/motor networks. Increased interhemispheric fc after recovery as well as a relative increase in anticorrelation of the aforementioned networks after compensation was also found (21, 22).

Our study focusing on the vestibular and visual networks detected changes in fc which were elicited by caloric vestibular stimulation in the acute stage of neglect. These changes were largely similar to changes in fc found in the resting state after recovery. In the acute stage, caloric stimulation particularly strengthened the transhemispheric connections within the vestibular network (insula, OP2, GTS, hippocampus, ACC), but weakened the transhemispheric connections between visual cortical areas. After functional recovery from neglect the fc at rest was similar to that of the vestibular network, but it differed in the intra- and transhemispheric visual-vestibular, and visuo-visual connections. In the following, we discuss similarities and differences between the effects of the transient changes elicited by caloric irrigation in the acute stage and the stable reorganization after recovery.

The effect of caloric stimulation on fc of OP2 was associated with an improved performance in Albert’s test which lasted about 20 min. This is in line with reports of caloric vestibular stimulation in neglect patients (23). The pattern of increased fc corresponds to that of activations during stimulation in healthy volunteers reported in positron emission tomography/fMRI studies (7, 8). It is noteworthy that there was a time gap of 5 min between caloric stimulation and start of the imaging run due to the performance of the neglect test. This (prolonged) effect of caloric stimulation on changes in the resting-state connectivity has not been shown before.

While rostral and mesial temporal regions in both hemispheres, which are involved in spatial perception and navigation (STG, MTG, hippocampus), showed stronger connectivity to OP2 after caloric stimulation, such an effect was not observed for the parietal regions (PPC), which are involved in spatial orientation to the contralateral left hemispace (Figures 2A,C and 4B). A direct influence of vestibular signals on the left PPC in the sense of sensory substitution or multisensory integration via enhanced visual input in the acute phase seems, therefore, unlikely. In contrast, the lack of interaction of the vestibular cortex and the PPC bilaterally implies that this transient improvement relies on a different mechanism, possibly an increase of the sensorial weight of the vestibular input to multisensory integration for spatial attention. The bilaterally increased fc between the vestibular cortex, the (para-)hippocampus, and the ACC may provide the structural basis for such a mechanism.

The maps for spatial orientation and navigation are created in the mesial temporal lobe. In brief, this is achieved by integrating landmark-based visual cues of the external environment with information about the position and movement of eyes, head, and body; this information is mediated by specialized cell populations in the hippocampus and entorhinal cortex [for detailed reviews see Ref. (24, 25)]. Vestibular responses in these regions have been shown in rodents and humans (13, 26). Indeed, neglect has also been reported in patients with parahippocampal lesions (27). Along with connections via the PPC (which was functionally disconnected), a pattern involving medial prefrontal cortex and dorsal anterior cingulum has been described (28). Our patient also showed such a pattern of increased fc, suggesting that the fc possibly takes this route between these areas.

Furthermore, while the interhemispheric connectivity between vestibular areas was increased, we found that the interhemispheric connectivity between homologous regions of the VC was reduced after caloric irrigation in the acute phase. This might have led to a reduced interhemispheric inhibition of the right VC after vestibular stimulation. It had been hypothesized earlier in a model approach that a lesion-induced neglect reduces the interhemispheric crosstalk (29). The reduced interhemispheric connectivity of VC regions is in line with the data of recent fc studies (21, 22). However, it reflects only one component of the complex vestibular–visual interactions in which the increase of the interhemispheric connectivity between vestibular areas plays a prominent role.

Like the fc pattern in the acute phase after caloric stimulation, the interhemispheric connectivity of the vestibular areas (OP2, hippocampus, dACC) was enhanced in the resting state after recovery from neglect 6 months later. Moreover, in contrast to the connectivity pattern in the acute phase, recovery was accompanied by increased fc of OP2 to the PPC (7A) on both sides and to visual areas in the right hemisphere. Connectivity between PPC (hIP3) and VC areas was also increased between both hemispheres. Additionally, increased fc was found between the right visual area hOC3d and both hippocampi, possibly representing substitutional processes of visual–vestibular interactions.

This implies that the repair mechanisms—as seen in the resting state without stimulation—show within-months similarities to as well as differences from the short-term improvement of neglect following caloric irrigation. The observed increase in fc of the vestibular cortex OP2 to the PPC indicates that multisensory signal integration was intensified over 6 months (Figure 3). The finding that the fc to multisensory vestibular areas is decreased following caloric stimulation in the compensated phase and increased between visual area hOC3d and the hippocampi might be due to a retuning of multisensory neurons to respond preferably to visual stimuli (i.e., substitution). Compared with other sensory systems, a specific feature of the cortical vestibular network is the multimodal nature of neural responses (30–32). While the increased tuning of the multisensory brain areas to visual cues led to compensation of the deficit at rest, vestibular stimulation created a sensory mismatch in the compensated phase. Usually a mismatch is suppressed by inhibitory visual–vestibular interactions (33), but in this case the higher preference for visual stimuli in the compensated phase might have led to an alteration of visual–vestibular interaction.

The presence of edema in the acute phase as well as structural reorganization of the damaged tissue in the chronic phase might also have contributed to the increase in fc between visual and parietal cortex. However, the increase of fc in vestibular cortical areas following vestibular stimulation in the acute phase in comparison to the relative decrease following stimulation after recovery suggest functional rather than structural reorganization.

To conclude, caloric irrigation led to changes in fc that imply the existence of a beneficial (short-term) mechanism that relies on an enhanced vestibular input to a multisensory integration process. The process is mediated by increased fc of the vestibular cortex to mesial temporal lobe areas involved in visual–vestibular signal integration for spatial attention, perception, memory, and navigation. This led to reduced interhemispheric connectivity in both VC. Functional recovery of the hemineglect 6 months later was associated with (i) fc increases of the ipsilateral vestibular cortex to VC areas of the right hemisphere, (ii) enhanced interhemispheric PPC connectivity, and (iii) fc increases between VC areas and PPC within the ipsilateral hemisphere. These pattern changes might be based on the tuning and enhancing of the sensorial weight of vision in multisensory neuronal assemblies in the parietal and insular cortex.

The fc data cannot answer the question of whether these connections are excitatory or inhibitory. It has been shown for the visual system that most transcallosal networks are inhibitory (34, 35). Therefore, models were proposed which explained the deficits of neglect by interhemispheric inhibition (29, 34, 35). However, excitatory transcallosal connections have been described for MT/V5 after motion stimulation of patients with complete homonymous hemianopia (36) and in an fMRI study using coherent visual motion stimulation of the right or left visual hemifield in healthy volunteers (37). To date, transcallosal vestibular connections have been demonstrated in humans by tracking fibers with diffusion tensor imaging (38), but it is not known whether the fc is inhibitory, excitatory, or both.

A further limitation of our study is that the changes in fc have been observed in a single patient whose spatial hemineglect was perceptional and deficient in peri-personal space and space perception/navigation. It remains an open question whether the observed fc changes associated with vestibular stimulation and the clinical course of our neglect patient are typical for this syndrome or whether other patients and other neglect subtypes show different neuronal reorganizations associated with functional recovery.

The study was performed in accordance with the 1964 Declaration of Helsinki (last applicable revision 2008, Fortalzea) and was approved by the institutional review board of the University of Munich. The patient gave his written informed consent to participate in the study and to publication of this case report.

JC: drafting/revising the manuscript, study concept and design, acquisition of data, and analysis and interpretation of data. RB: drafting/revising the manuscript and analysis and interpretation of data. ME: drafting/revising the manuscript and acquisition, analysis and interpretation of data. TB: critical revision of the manuscript for important intellectual content, study supervision, and obtaining funding. MD: drafting/revising the manuscript, study concept and design, analysis and interpretation of data, study supervision, obtaining funding.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.