Analyzing patient behavior as optimal adaptation: measuring and modeling perceptual consequences of vestibular lesions


Principle Investigator(s):

Dr. Paul MacNeilage

Dr.-Ing Stefan Glasauer



We propose that patient behavior should not be compared with normal behavior, but should instead be compared with behavior of an ideal observer model that accounts for the known anatomical and/or physiological impairments. Thus, patient behavior can be considered ideal or optimal given the known constraints imposed by the disease.


Project description(s):

 I.)    Compensation in central versus peripheral unilateral vestibular patients

We suggest that long-term recovery from unilateral vestibular loss represents optimal central adaptation to diminished unilateral peripheral function. Over the course of recovery, central mechanisms take into account the changed gain and variability of sensory signals originating from the lesioned side. This sub-project aims to carefully evaluate and refine a previously developed statistically optimal ideal observer model for patients with peripheral and central lesions. The different patterns of compensation in the two patient types will be characterized with a unified model. Successful execution of this sub-project will lead to improved understanding of central compensation for unilateral vestibular loss which should inform recommended clinical treatment and evaluation of these conditions.

 II.)   Tilt and translation estimation in patients with cerebellar lesions

The otoliths signal the sum of gravity and linear acceleration, such that the brain must combine these signals with information from the semicircular canals in order to reconstruct separate estimates of tilt (i.e. orientation relative to gravity) and translation (i.e. linear motion). Here we propose to evaluate whether tilt-translation processing in normal subjects and patients can be predicted by an ideal observer tilt-translation model. We will develop a protocol on the motion platform in which accurate performance relies heavily on correct resolution of the tilt-translation ambiguity, i.e. integration of canal and otolith signals. We expect to show relatively accurate performance in normal subjects, but hypothesize a different pattern of results in cerebellar patients because tilt-translation processing is thought to occur in the cerebellum.




III.) Visual and ocular stabilization during simulated vestibular loss

Acute vestibular loss will be simulated in normal subjects on the motion platform by tracking and cancelling the actively generated head-on-trunk movement such that the trunk moves under the head while the head remains stationary in space. The impact of this manipulation will be assessed using both perceptual and oculomotor measures. This project provides a means to evaluate the consequences of acute vestibular loss that can never be achieved in patients because patients can never be observed in the first moments following vestibular loss. Results will provide clues to the functioning of the system in healthy subjects and allow quantification of the immediate perceptual and motor consequences of acute vestibular loss.



  • Virtual reality (MCM-M)
  • Computational modeling (MCM-M)
  • Eye movement tracking (MCM-T)


Relevant publications:

Cousins S, Kaski D, Cutfield N, Seemungal BM, Golding JF, Gresty M, Glasauer S, Bronstein AM. (2013) Vestibular perception following acute unilateral vestibular lesions. PloS ONE  8(5):e61862.

MacNeilage PR, Ganesan N, Angelaki DE. (2008) Computational approaches to spatial orientation: from transfer functions to dynamic Bayesian inference. J Neurophysiol. Dec;100(6):2981-96.

Mergner T, Glasauer S. (1999) A simple model of vestibular canal-otolith signal fusion. Ann. NY Acad. Sci. 871, 430-434.

Glasauer S, Merfeld DM. (1997) Modeling three dimensional vestibular responses during complex motion stimulation. In: Three-Dimensional Kinematic Principles of Eye-, Head , and Limb Movements in Health and Disease (Fetter M, Haslwanter T, Misslisch H, Tweed D, eds.) Harwood Amsterdam, pp 387-398.

MacNeilage PR, Turner AH, Angelaki DE. (2010) Canal–otolith interactions and detection thresholds of linear and angular components during curved-path self-motion. J Neurophysiol 104: 765–773.

MacNeilage PR, Banks MS, Berger DR, Bülthoff HH. (2007) A Bayesian model of the disambiguation of gravitoinertial force by visual cues. Exp Brain Res. 2007 May;179(2):263-90.


MacNeilagePhoto2013small Glasauersw Cuturi Luigi Garzorz Isabelle -klein
Paul MacNeilage Stefan Glasauer  Luigi Cuturi  Isabelle Garzorz



Paul R. MacNeilage

Address: Feodor-Lynen-Str. 19, 81377 Munich, Germany

Tel: + 49 (0) 89 4400 7 7823

Fax: + 49 (0) 89 4400 7 4801