Multisensory processing

The Effects of Size, Duration, and Luminance of Visual Line on Apparent Vertical While the Head was Tilted

Atsuki Higashiyama

We determined orientation of a line that is seen to be vertical (ie., apparent vertical) while the head is tilted with the upright trunk. In this condition, it has been documented that apparent vertical is independent of head tilt (ie., orientation constancy) or is in a direction opposite to the head tilt (ie., the Müller effect). The most of previous studies regarding apparent vertical, however, have focused on the effect of head tilt but have been less attended to the parameters of the visual line by which apparent vertical is judged. In this study, we manipulated size (5.5 and 22 deg arc in visual angle), duration (0.1s, 3s, and no time limit), and luminance (0.026, 0.003, and 0.001cd/m2 against total darkness) of the line with the head being tilted within 30 deg arc. The main findings were: 1) the line of shorter duration or of lower luminance facilitated the Müller effect and 2) apparent vertical for the line of small size leaned toward the left of that for the line of large size. These results are not explained by the theory of taking head tilt into account but are explained by modifying the sensory-tonic field theory.

Auditory stimuli modulates visual time dilation illusion

Takuro Kayahara

Previous studies reported an illusion in which motion information of visual stimuli can lengthen the perceived duration of those [Brown, S.W., 1995, Percept. Psychophys., 57, 105-116], and suggested that the illusion is based not on retinotopic- but on spatiotopic-level visual processing. [Au, R.K.C et al., 2012, frontiers in Psychology, 3(58), 1-7] Meanwhile, other multimodal studies reported the dominance of auditory information in the time-domain perception. [e.g. Shams, L. et al., 2000, Nature, 408, 788] Thus, in this study, the author examined whether the accompaniment of auditory stimuli with the time dilation illusion can modulate lengthened duration of visual stimuli in the illusion, mainly with the condition of the auditory motion. (auditory motion either by inter-aural time difference (corresponds to retinal motion) or by real speaker array (corresponds to spatial motion), or static sound (without any motion)). The relation between across-modal object representation and the time perception will be discussed with obtained auditory modulation on the visual time dilation illusion. [Supported by JSPS (No.23500261)]

Similar systematic biases in visual and vestibular heading perception

Luigi Cuturi and Paul MacNeilage

Visual and vestibular sensory information is processed by the brain in order to perceive direction of linear self-motion (i.e. heading). Visually, translation in space results in a characteristic retinal flow pattern: its focus of expansion provides information about heading direction. In the vestibular system, otoliths signal changes in linear acceleration allowing the brain to identify heading direction. Here we investigate accuracy of visual-only or vestibular-only heading judgments. Optic flow stimuli presented in stereo are used to investigate visual heading perception, whereas linear self-motion translation on the earth-horizontal plane is delivered to elicit a vestibular response. Subjects are asked to indicate the heading angle of the experienced motion (visual or vestibular) by adjusting the angle of a visually presented arrow. Accuracy of both visual and vestibular judgments depends significantly on heading angle. Heading directions close to the obliques (i.e. 45° and 135°) are the most biased, showing significant overestimation for forward directions and significant underestimation for backward directions. Visual and vestibular biases are significantly positively correlated and do not significantly differ for forward movement directions. Similarity between visual and vestibular heading biases suggests a common spatial representation influences heading judgments regardless of the sensory modalities involved.

The central tendency of judgment: a consequence of Bayesian estimation?

Neil Roach, James Heron, Paul McGraw and David Whitaker

A common finding in reproduction and estimation experiments is that judgments are compressed around a central value - the magnitude of stimuli falling below the mean of the range is typically overestimated, whereas the magnitude of stimuli above the mean is underestimated. This pervasive bias was first noted in studies of temporal interval perception by Vierordt (1868, Der Zitsinn nach Veruchen, Tubingen, Laupp), but has since been documented across a range of different sensory judgments. Recently formulated Bayesian observer models suggest this bias arises from the optimal fusion of noisy stimulus measurements with acquired knowledge of the set of stimuli from which it is drawn. However, in a series of duration reproduction experiments, we demonstrate that interleaving stimuli sampled from distinct, non-overlapping 'short'and 'long'distributions produces compression of estimates around a single duration value, even if the distributions are widely separated and clearly demarcated by way of spatial location, sensory modality or temporal sequence. These results are at odds with the simulated biases of a Bayesian observer model with prior expectations matching either one of the stimulus distributions or their combination.

Audio-visual synchrony detection under scotopic conditions

John Cass, Kate Churruca, Erik Van der Burg and David Alais

In cluttered visual displays, abrupt, uniquely synchronised audio-visual events capture attention. This study compares the temporal determinants of this 'pip and pop'phenomenon under photopic and scotopic conditions. In conjunction with prolonged dark adaptation, scotopic conditions were achieved using goggle-mounted neutral density filters. The primary task involved identifying the orientation of a target singleton, horizontally or vertically presented among 5, 9 or 13 distractor lines tilted (±4°: photopic; ±10°: scotopic) from horizontal and vertical. Target and distractor elements were each surrounded by a luminance-defined annulus modulating at 0.78 Hz, each with a unique temporal phase. On half of the trials an amplitude-modulating 500Hz tone was synchronised with the modulation of the target annulus and was absent on the remaining trials. The modulation profile of luminance and tone was sinusoidal or square. For sinusoidally modulating trials, the presence of the tone had no effect, with highly serial search performance under both illumination conditions. For square-wave trials the presence of the tone dramatically improved search efficiency only under photopic conditions, with no tone-related improvement observed scotopically. This result persisted after advancing the tone in time to account for luminance-related lags in neural latency. Contrary to predictions based on the scotopic temporal channels literature, the effects of illumination on AV synchrony-driven visual search appear to result from a profound low-pass temporal filtering within the scotopic system.