The long-term goal of these studies is to understand the cerebellar processing of vestibular information and visual/vestibular interactions. The proposed studies are motivated by recent findings that the cerebellar nodulus/uvula (vermis lobules X and IX, NU), the vestibular (VN) and deep cerebellar (CN) nuclei appear to form an interconnected network that implements multisensory convergence necessary to resolve the gravito- inertial acceleration (GIA) ambiguity and distinguish gravity from translational accelerations. Theory has proposed that the GIA ambiguity is resolved by generating an internal estimate of gravity by appropriately 'combining'otolith, semicircular canal and visual sensory information, as well as prior knowledge about the statistics of linear accelerations experienced most commonly in everyday life (conceptualized as a Bayesian prior). Here we plan to probe the neural implementation of such an internal model in macaque NU Purkinje cells (aims 1 &2), as well as start a functional dissection of how these signals are generated within the VN-CN- NU network by recording from NU-projecting and NU-target neurons in the VN/CN (aim 3). We will use stimuli that create the illusion of translation, without actually delivering any translation stimulus to test the following hypotheses: (1) NU Purkinje cell activity represents the output of long-postulated internal models for gravity and translational acceleration;(2) Neural correlates of gravity and translational acceleration signals are both found in distinct populations of NU Purkinje cells;(3) Population activity of these two groups of NU Purkinje cells are complementary to each other, such that their net sum equals GIA, as predicted by theory;(4) Vision can be used, instead of canal-driven signals, to compute a more reliable estimate of gravity during steady-state rotation, thus preventing the generation of erroneous tilt and translation signals;and (5) Translation-selective and gravity-selective response properties are only found in orthodromically-activated NU-target neurons, but not antidromically-activated NU-projecting neurons, all of which are either canal-only (i.e., have no response to translation) or GIA-coding cells. Together, these experiments constitute fundamental studies necessary to establish the NU-VN-CN circuitry as key areas in inertial multisensory processing for self-motion perception and spatial orientation, critical for allocentric orientation and inertial navigation. A major innovation in the current studies is the use of un-natural stimuli that are known to induce tilt and translation illusions to challenge the system and unmask the underlying computations. These experiments will provide novel quantitative evidence for the neural correlates of long-postulated theoretical concepts, like internal models and Bayesian priors. Thus, this work has a significant broader impact for unveiling the intricate mysteries of the functional roles of the cerebellum and its circuitry.
Clinical and experimental lesions involving the vestibulo-cerebellum in the caudal vermis and its interconnections with brainstem nuclei lead to clinical nystagmus, reduced visual acuity, postural instability and loss of spatial orientation. Neurological correlates of central vestibular disorders are still a mystery, posing a major hurdle in defining effective therapeutic strategies. The experiments proposed here aim at filling a very notable gap in knowledge, important for understanding and treating both basic postural and reflexive deficits as well as cognitive deficits of spatial perception.
Showing the most recent 10 out of 64 publications
