PROGRAMME - SUMMER TERM 1995

  • Thursday 25 May at 4.30pm

    Dr Laurence Harris, York University, Toronto, Canada - "The coding of velocity and direction of self motion in response to visual and vestibular information"

    venue: Department of Psychology, Royal Holloway

    Abstract

    Rotational component (work done in collaboration with Agnieszka Kopinska, Peter Mente, Lori Lott)

    I made formal a hypothesis that has been implied in studies involving stimuli transduced by the semicircular canals, namely that the entire system, not just the end organ, operates using a channel coding system. My hypothesis was that both a) the orientat ion of the axis of rotation and b) the magnitude of rotation, are coded by the pattern of activity across a small number of channels, not the activity in any one channel. Since multi-sensory information, especially visual and vestibular information, contr ibutes to the self-rotation system, an extension of my hypothesis is that information pertinent to self motion from any source (visual, vestibular etc..) is coded in such a channel system.

    I described two experiments that address these hypotheses. In the first experiment I measured psychophysical thresholds to full-field motion compatible with that normally found during a head rotation. The distribution of sensitivities is compatible with a channel method of coding and is difficult to explain any other way. In the second experiment I measured the direction and velocity of feline eye movements in response to visual and vestibular rotation around different axes at the same time. This independ ence of stimuli was achieved by mounting the planetarium projector on the vestibular platform within a spherical screen. Eye movements were measured in response to instantaneous combinations of visual and vestibular stimuli and also to vestibular stimulat ion in the dark following prolonged exposure to unusual combinations. The cat was chosen because of its lack of smooth pursuit eye movements which simplifies the interpretation of the results. The results are compatible with a channel method of coding.

    Translational component (work done in collaboration with Michael Jenkin, Nicole Acouin, Emre Onin, Danny Kreichman, Dan Zikovitz, Kristiina McConville)

    The experiments described assessed the relative contributions of visual and vestibular information to the perception of linear displacement. Subjects sat on a chair that could move at a constant acceleration (0.07-0.7m/s/s). They wore a virtual reality he lmet (84°x65°) in which motion along a corridor (2m x 2.5m x 50m) was simulated. The walls of the corridor were textured with dynamically varying features in order to limit the subject's ability to track fixed points. Motion could be either (i) vestibular only (dark), (ii) visual motion only (arranged to match vestibular only) or (iii) a combination. Subjects indicated when they reached the position of a previously-presented, virtual, visual target (0.5-1.75m). An analysis of variance showed that perceptu al accuracy, defined as perceived distance/target distance, was affected by the type of sensory information provided to the subject. The greatest accuracy was achieved when both vestibular and visual information were provided. Curiously subjects were usually more accurate when they moved in the dark (vestibular only condition) than under the vision alone condition. Models of combination strategies were suggested. Cooperation between different sources of information is important in maintaining percept ual accuracy. Judgements based on a SINGLE source alone (eg. vision when simulating an immersive environment) may be inaccurate.


  • Monday 12 June at 4.30pm

    Dr Mike Harris, University of Birmingham - "Is the analysis of optic flow 1D or 2D?"

    venue: Department of Psychology, University of Reading

  • Monday 10 July at 4.30pm

    Dr Yuede Yang, Vanderbilt University, Nashville TN, USA - "Stereopsis and motion in the spatial frequency domain"

    venue: Department of Psychology, University of Surrey


    Return to Seminar Index

    Return to Foursites Homepage