Organizational structures intrinsic to nervous systems can be more precisely analyzed and compared with other logical structures once they are expressed in mathematical languages. A standard mathematical language for expressing organizational structure is that of groups. Groups are especially well suited to organizational structures involving multiple symmetries such as spatial structures.

The vestibular system is widely believed to mediate many neural functions involving spatial structure. The vestibular nuclei receive direct projections from the vestibular endorgans, the semicircular canals and the otolith organs. The near-orthogonal directions of the semicircular canals are embedded in the bone. However, those canal directions are external to the nervous system.

This study addresses the way the three-dimensional space of rotations is also embedded in the group structure of neural connectivity. Although we know a great deal about physical rotation, it is not clear that nervous systems organize rotations in the same way as physicists do. It would make sense for nervous systems to organize rotations in such a way as to provide physiologically relevant information about performing or compensating for rotations. The vestibular nuclei, which might be expected to display an organization that binds rotations into a rotation space, do not give a clear organization. This may be because of the multiplicity of spatial functions performed by the vestibular nuclei; rather than one spatial organization, the vestibular nuclei are likely to accommodate multiple, related spatial organizations.

This study evaluates one particular data set from the literature that specifies the organization of the disynaptic canal-neck projection; other projections and neuronal populations may have other intrinsic organizations. The data are evaluated directly for their symmetry group. In the symmetry group, the vertebrate requirement that physiology have a right and left is found to be satisfied in two ways: (i) by a hexagonal symmetry arising from the right-left doubling of front and back, (ii) along with separate organizations on the two sides that may be required to operate independently to some extent. The eight observed muscle innervation patterns from the data are the complete set of possible combinations of inhibitory/excitatory polarities from three canal pairs. These eight innervation patterns are organized as the vertices of a cube. The two types of side muscles provide the vertical direction. As the head rotates in physical space, the cube rotates in sensorimotor space.

Like the canal-neck projection, otolith projections and proprioceptive afferents contact both the vestibular nuclei and neck motoneurons. They may have a similar organization, perhaps with extensions of the same pattern. Otherwise, like a checkerboard superimposed over a paisley, they will form an overlapping organization with the disynaptic canal-neck projection. Further research is required to determine whether the sensorimotor spatial structure of the canal-neck projection is widespread in nervous systems or whether there are several complete structures that are fragmented and reintegrated.