Detection of rotating gravity signals

Research output: Contribution to journalArticlepeer-review


It is shown in the preceding paper that neurons with two-dimensional spatio-temporal properties to linear acceleration behave like one-dimensional rate sensors: they encode the component of angular velocity (associated with a rotating linear acceleration vector) that is normal to their response plane. During off-vertical axis rotation (OVAR) otolith-sensitive neurons are activated by the gravity vector as it rotates relative to the head. Unlike "one-dimensional" linear accelerometer neurons which exhibit equal response magnitudes for both directions of rotation, "two-dimensional" neurons can be shown to respond with unequal magnitudes to clockwise and counterclockwise off-vertical axis rotations. The magnitudes of the sinusoidal responses of these neurons is not only directionally selective but also proportional to rotational velocity. Thus, responses from such "two-dimensional" neurons may represent the first step in the computations necessary to generate the steady-state eye velocity during OVAR. An additional step involving a nonlinear operation is necessary to transform the sinusoidally modulated output of these neurons into a signal proportional to sustained eye velocity. Similarly to models of motion detection in the visual system, this transformation is proposed to be achieved through neuronal operations involving mathematical multiplication followed by a leaky integration by the velocity storage mechanism. The proposed model for the generation of maintained eye velocity during OVAR is based on anatomical and physiological properties of vestibular nuclei neurons and capable of predicting the experimentally observed steady-state characteristics of the eye velocity.

Original languageEnglish (US)
Pages (from-to)523-533
Number of pages11
JournalBiological cybernetics
Issue number6
StatePublished - Oct 1992

ASJC Scopus subject areas

  • Biotechnology
  • General Computer Science


Dive into the research topics of 'Detection of rotating gravity signals'. Together they form a unique fingerprint.

Cite this