AVA Annual Meeting 2006
Vision in Perception and Cognition
Tuesday, 4th April 2006
University of Bradford
John Stanley Bell Lecture Theatre
Abstracts for oral presentations (in session order):
Session 1
Colour and shape of faces and their influence on attributions
David Perrett (School of Psychology, University of St Andrews, St. Mary's College, St.
Andrews, KY16 9JP. e-mail: dp@st-andrews.ac.uk)
We are highly sensitive to the shape and colour cues to faces. These visual attributes control
our aesthetic judgments influence psychological attributions about face owners (do we trust
them, are they healthy). I will discus data and theory concerning the basis of such visual
influences focusing on underlying biology of face owners and perceivers.
Co-authors: Miriam Law Smith, Ben Jones, Lisa deBruine, Jamie Lawson, Michael Stirrat,
Vinet Coetzee, Peter Henzi.
Contrasting the disruptive effects of view changes in shape discrimination to the
disruptive effects of shape changes in view discrimination
Rebecca Lawson1 and Heinrich H. Bülthoff2 (1School of Psychology, University of Liverpool,
Liverpool, L69 7ZA; 2Max Planck Institut für biologische Kybernetik, Spemannstraße 38,
72076 Tübingen, Germany. e-mail: r.lawson@liverpool.ac.uk)
A series of three sequential picture-picture matching studies compared the effects of a view
change on our ability to detect a shape change (Experiments 1 and 2) and the effects of a
shape change on our ability to detect a view change (Experiment 3). Relative to no-change
conditions, both view changes (30° or 150° depth rotations) and shape changes (small or
large object morphing) increased both reaction times and error rates on match and mismatch
trials in each study. However, shape changes disrupted matching performance more than
view changes for the shape-change detection task ("did the first and second pictures show
the same shape?"). Conversely, view changes were more disruptive than shape changes
when the task was to detect view changes ("did the first and second pictures show an object
from the same view?"). Participants could thus often discriminate between the effects of
shape changes and view changes. The influence on performance of task-irrelevant
manipulations (view changes in the first two studies; shape changes in the final study) does
not support Stankiewicz's (2002; Journal of Experimental Psychology: Human Perception &
Performance, 28, 913-932) claim that information about viewpoint and about shape can be
estimated independently by human observers. However the greater effect of variation in the
AVA Annual Meeting 2006
task-relevant than the task-irrelevant dimension indicates that observers were moderately
successful at disregarding irrelevant changes.
Detection of convexity and concavity: no evidence for a concavity advantage
Marco Bertamini and Amy O'Neill (School of Psychology, University of Liverpool, Bedford
Street South, Liverpool L69 3BX. e-mail: M.Bertamini@liverpool.ac.uk)
We compared sensitivity to shape changes, and in particular detection of convexity and
concavity changes. The available evidence in the literature is inconclusive. We report three
experiments that used a temporal 2AFC change-detection task in which observers viewed
two polygonal contours. They were either identical, or one vertex had been removed (or
added). The polygonal stimuli were presented as random dot stereograms to ensure
unambiguous figure-ground segmentation. We report evidence that any change of sign of
curvature along the contour is salient, leading to high sensitivity to change (d') (experiment 1,
also in Bertamini & Farrant, 2005, Acta Psych, 120, 35-54). This in itself is not surprising.
However, in experiments 2 and 3 we failed to find any evidence that changes to concave
vertices, changes within concave regions, or the introduction of a new concave vertex are
more salient than, respectively, a change to a convex vertex, a change within a convex
region, or the introduction of a new convex vertex. We conclude that performance in
detection for shape changes is good when there is a change of sign, which in turn may signal
a change in part structure. Performance was also higher for changes nearer the centre of the
object. This factor can explain why a concavity advantage has been previously reported (cf.
Cohen, Barenholtz, Singh & Feldman, 2005, Journal of Vision, 5, 313-321).
Modelling the spatial tuning of the Hermann grid illusion
Michael J. Cox, Jose B. Ares-Gomez, Ian E. Pacey, Jim M. Gilchrist, Ganesh T. Mahalingam
(Vision Science Research Group, School of Life Sciences, University of Bradford, Bradford.
e-mail:M.Cox@bradford.ac.uk)
Purpose: To test the ability of a physiologically plausible model of the retinal ganglion cell
(RGC) receptive field (RF) to predict the spatial tuning properties of the Hermann Grid
Illusion (HGI). Methods: The spatial tuning of a single intersection HGI was measured
psychophysically in normal subjects using a nulling technique at different vertical cross line
luminances. We used a model based upon a standard RGC RF, balanced to produce zero
response under uniform illumination, to predict the response of the model cell to the
equivalent range of stimulus conditions when placed in the 'street' of the stimulus cross, or its
intersection. We determined the equivalent of the nulling luminance required to balance
these responses and minimise the HGI. Results: The model and the psychophysics
demonstrated broad spatial tuning with similarly shaped tuning profiles and similar strengths
of illusion. The limb width at the peak of the model tuning function was around twice the RGC
RF centre size, whilst psychophysically the peak of the spatial tuning function was at least
twice the size of RF centres expected from physiological evidence. In the model and
psychophysically the strength of the illusion varied with the luminance of the vertical line of
the cross when expressed as a Michelson contrast, but not when expressed as a luminance.
Conclusions: The shape, width, height and position of the spatial tuning function of the HGI
can be well modelled by a single-size RGC RF based model. The broad tuning of these
-2-
AVA Annual Meeting 2006
functions does not appear to require a broad range of different cell sizes either in the retina
or later in the visual pathway.
Fused video assessment using scanpaths
Timothy D. Dixon1, Stavri G. Nikolov2, Jian Li1, John Lewis2, Eduardo Fernandez Canga2,
Jan M. Noyes1, Tom Troscianko1, Dave R. Bull2, C. Nishan Canagarajah2 (1Department of
Experimental Psychology, 8 Woodland Rd, Bristol, BS8 1TN; 2Centre for Communications
Research, Merchant Venturer’s Build, Woodland Road, Bristol BS8 1UB. e-mail:
Timothy.Dixon@bris.ac.uk)
There are many methods of combining two or more images of differing modalities, such as
infrared and visible light radiation. Recent literature has addressed the problem of how to
assess fused images using objective human methods (Dixon et al, 2005, Percep 34, Supp,
142; accepted, ACM Trans Applied Percep). As these have consistently shown that
participants perform differently in objective and subjective assessment tasks, it is essential
that a reliable assessment method is used. The current study extends this work, considering
scanpaths of individuals involved in a tracking task. Participants were shown two video
sequences of human walking down a path amongst foliage, each containing the original
visible light and infrared input images, as well as fused ‘averaged’, discrete wavelet
transformed (DWT), and dual-tree complex wavelet transformed (DT-CWT) videos. Each
participant was shown the sequences on three separate occasions, and was asked to track
visually the figure, as well as performing a key-press reaction task, whilst an eye-tracker
recorded fixation data. The two sequences were similar in content, although the second had
lower luminance levels. Accuracy scores were obtained by calculating the amount of time
participants looked at the target as a ratio of the total time they looked at the screen. Results
showed that the average and DWT methods resulted in better performance on the high
luminance sequence, whilst the visible light was poorer than all others on the low luminance
sequence. Secondary task results were also analysed, although only small variations were
found in these data. The current findings suggest that this new method of fused image
assessment could have wide-reaching further potential for differentiating between fusion
methods.
Session 2
Aversion to contemporary art
Dominic Fernandez, Arnold J. Wilkins and Debbie Ayles (Department of Psychology,
University of Essex, Colchester CO4 3SQ. e-mail: dfernaa@essex.ac.uk)
Discomfort when viewing square-wave luminance gratings can be predicted from their spatial
frequency (Wilkins et al., 1984, Brain, 107, 989-1017). Gratings with a spatial frequencies
close 3 cycles per degree tend to evoke greatest discomfort and the largest number of visual
illusions/distortions. We asked people to rate the discomfort they experienced when viewing
images of non-representational art in order to examine the effect of the luminance energy at
different spatial frequencies in more complex images. To do this we calculated the radial
Fourier amplitude spectrum of image luminance. Images rated as more comfortable showed
-3-
AVA Annual Meeting 2006
a regression of luminance amplitude against spatial frequency similar to that found for natural
images (Field, D.J., 1987, Journal of the Optical Society of America A 4 2379-2394),
whereas more uncomfortable images had greater proportion of energy at spatial frequencies
close to 3 cycles per degree. Luminance amplitude at this spatial frequency explained
between 7% and 51% of the variance in the discomfort reported when viewing art. The more
uncomfortable images were rated as having lower artistic merit. The Fourier amplitude
spectra of photographs of urban and rural scenes, which had been classified as pleasant or
unpleasant, were also examined and exhibited a similar pattern, with the energy at 3 cycles
per degree accounting for 6% of the variance. This suggests that predictions of subjective
discomfort when viewing complex and real world images can be based on the distribution of
luminance energy across spatial frequencies.
Dependencies between gradient directions at multiple locations are determined by the
power spectra of the image
Alex J. Nasrallah and Lewis D. Griffin (Department of Computer Science, UCL, Malet Place,
London, WC1E 6BT. e-mail: a.nasrallah@cs.ucl.ac.uk)
A well-known statistical regularity of natural image ensembles is that their average power
spectra exhibit a power-law dependency on the modulus of spatial frequency. This result
does not describe the correlation between gradient directions but instead describes the
correlation between pairs of pixel intensities. In this study, we have used informationtheoretic measures to compute the amount of dependency, which exists between two
gradient directions at separate locations; we classify this result as 2-point statistics. This is
then extended to measure the dependencies of gradient directions at three separate
locations which we classify as 3-point statistics. To assess the influence of the power
spectrum on the interactions of gradient directions, we collected statistics from four different
image classes: A - natural images, B - phase-randomized natural images, C - whitened
natural images and D - Gaussian noise images. The image classes A and B had the same
power spectra, as did C and D.
The main result was that the dependencies between gradient directions at multiple locations
are determined by the power spectra of images. This is based on image classes A and B, as
well as C and D, having the same amount of 2-point and 3-point gradient direction
dependencies. Further, we have studied other image classes with different forms of powerspectra, and these experiments did not invalidate the main result.
Visual search with random stimuli
William McIlhagga (Department of Optometry, University of Bradford, Bradford, BD7 1DP. email: W.H.Mcilhagga@Bradford.ac.uk)
Visual search experiments are usually conducted with "designed" stimuli, which possess,
lack, or conjoin one or more obvious features chosen by the experimenter. In this study,
visual search was conducted with randomly generated stimuli, to see what - if any - new
results might be uncovered using them. Target stimuli were generated by randomly colouring
half the squares in a 5-by-5 grid black, the other half white. Each random target was paired
with a distractor, obtained by flipping 1,2, or 3 of the target's white squares for black, and
-4-
AVA Annual Meeting 2006
black squares for white. Stimuli were displayed on a white background. As usual, set size
was the dependent variable.
Using these stimuli, search times when the target was present varied from 2 to 50 ms per
distractor. The search times were usually unrelated to the number of squares flipped
between target and distractor. They were also unrelated to the discrimination threshold for
target vs. distractor obtained in a separate experiment. The search rates when the target was
absent were on average 3 times slower than when the target was present, rather than the
more usual 2 times slower. In particular, some searches that were essentially "parallel" when
the target was present, became "serial" searches (~30ms/item) when the target was absent.
This result is discussed in terms of memoryless search and a tradeoff between miss rate and
search effort.
The perception of chromatic stimuli in the peripheral human retina
Declan J. McKeefry1, Neil R. A. Parry2 and Ian J. Murray3 (1Department of Optometry,
University of Bradford, Bradford BD7 1DP; 2Vision Science Centre, Manchester Royal Eye
Hospital, Manchester; 3Faculty of Life Sciences, University of Manchester, Manchester. email: D.McKeefry@Bradford.ac.uk)
Perceived shifts in hue and saturation that occur in chromatic stimuli with increasing retinal
eccentricity were measured in 9 human observers using a colour matching paradigm for a
range of colour stimuli which spanned colour space. The results indicate that hue and
saturation changes are dissociable from one another in the respect that: i) they exhibit
different patterns of variation across colour space, and ii) they possess different
dependencies on stimulus size, perceived shifts in hue being relatively immune to increases
in stimulus size whilst the desaturation effects are eliminated by increases in size of
peripherally presented stimuli. This dissociation implies that different physiological
mechanisms are involved in the generation of perceived changes in the hue and saturation of
chromatic stimuli that occur in the retinal periphery. Furthermore, when we model the
magnitude of activation produced in the L/M and S- cone opponent systems by the peripheral
chromatic stimuli we find that the variations in perceived saturation seem to be largely
mediated by changes in output from the L/M opponent system, with the S-cone opponent
system affected to a much lesser degree.
Perception of colour contrast stimuli in the presence of scattering
Maris Ozolinsh1, Gatis Ikaunieks1 , Sergejs Fomins1, Michèle Colomb2 and Jussi Parkkinen3
(1Dept.of Optometry, University of Latvia, Latvia; 2Laboratoire Régional des Ponts et
Chaussées de Clermont-Ferrand, France; 3Color Research Laboratory, University of
Joensuu, Finland. e-mail: ozoma@latnet.lv)
Visual acuity and contrast sensitivity were studied in real fog conditions (fog chamber in
Clermont-Ferrand) and in the presence of light scattering induced by light scattering eye
occluders. Blue (shortest wavelength) light is scattered in fog to the greatest extent, causing
deterioration of vision quality especially for the monochromatic blue stimuli. However, for
colour stimuli on a white background, visual acuity in fog for blue Landolt-C optotypes was
higher than for red and green optotypes. The luminance of colour Landolt-C optotypes
presented on a screen was chosen corresponding to the blue, green and red colour
-5-
AVA Annual Meeting 2006
contributions in achromatic white stimuli. That results in the greatest luminance contrast for
the white-blue stimuli, thus improving their acuity. Besides such blue stimuli on the white
background have no spatial modulation of the blue component of screen emission. It follows
that scattering which has the greatest effect on the blue component of screen luminance has
the least effect on the perception of white-blue stimuli comparing to white-red and, especially,
to white-green stimuli. Visual search experiments were carried out with simultaneous eye
saccade detection. Red, green and blue Landolt-C stimuli were blurred using a Gaussian
filter to simulate fog, and were shown together with distractors. Studies revealed two
different search strategies: (1) for low scattering - long saccades with short total search
times; (2) for high scattering, shorter saccades and long search times. Results of all
experiments show better recognition of white-blue comparing to white-green colour contrast
stimuli in the presence of light scattering.
Session 3
Models of binocular contrast discrimination predict binocular contrast matching
Daniel H. Baker, Tim S. Meese and Mark A. Georgeson (School of Life & Health Sciences,
Aston University, Birmingham B4 7ET. e-mail: bakerdh@aston.ac.uk)
Our recent work on binocular combination does not support a scheme in which the nonlinear
contrast response of each eye is summed before contrast gain control (Legge, 1984, Vis
Res, 24, 385-394). Much more successful were two new models in which the initial contrast
response was almost linear (Meese, Georgeson & Baker, J Vis, submitted). Here we extend
that work by: (i) exploring the two-dimensional stimulus space (defined by left- and right-eye
contrasts) more thoroughly, and (ii) performing contrast discrimination and contrast matching
tasks for the same stimuli. Twenty five base-stimuli (patches of sine-wave grating) were
defined by the factorial combination of 5 contrasts for the left eye (0.3-32%) with five
contrasts for the right eye (0.3-32%). Stimuli in the two eyes were otherwise identical (1
c/deg horizontal gratings, 200ms duration). In a 2AFC discrimination task, the base-stimuli
were masks (pedestals) where the contrast increment was presented to one eye only. In the
matching task, the base-stimuli were standards to which observers matched the contrast of
either a monocular or binocular test grating. In both models discrimination depends on the
local gradient of the observer's internal contrast-response function, while matching equates
the magnitude (rather than gradient) of response to the test and standard. Both models
produced very good fits to the discrimination data. With no remaining free parameters, both
also made excellent predictions for the matching data. These results do not distinguish
between the models but, crucially, they show that performance measures and perception
(contrast discrimination vs contrast matching) can be understood in the same theoretical
framework.
Binocular combination: the summation of signals in separate ON and OFF channels
Mark A. Georgeson and Tim S. Meese (School of Life & Health Sciences, Aston University,
Birmingham B4 7ET, UK. e-mail: m.a.georgeson@aston.ac.uk)
How does the brain combine visual signals from the two eyes? We quantified binocular
summation as the improvement in contrast sensitivity observed for two eyes compared with
-6-
AVA Annual Meeting 2006
one, and asked whether the summation process preserves the sign of signals in each eye. If
it did so, then gratings out-of-phase in the 2 eyes might cancel each other and show lower
sensitivity than one eye. Contrast sensitivity was measured for horizontal, flickering gratings
(0.25 or 1 c/deg, 1 to 30 Hz, 0.5s duration), using a 2AFC staircase method. Gratings inphase for the 2 eyes showed sensitivity that was up to 1.9 times better than monocular
viewing, suggesting nearly linear summation of contrasts in the 2 eyes. The binocular
advantage was greater at 1 c/deg than 0.25 c/deg, and decreased to about 1.5 at high
temporal frequencies. Dichoptic, anti-phase gratings showed a very small binocular
advantage, a factor of 1.1 to 1.2, but with no evidence of cancellation. We present a
nonlinear filtering model that accounts well for these effects. Its main components are an
early linear temporal filter, followed by half-wave rectification that creates separate ON and
OFF channels for luminance increments and decrements. Binocular summation occurs
separately within each channel, thus explaining the phase-specific binocular advantage. The
model accounts well for earlier results with brief flashes (eg. Green & Blake, 1981, Vis Res,
21, 365-372 ) and nicely predicts the finding that dichoptic anti-phase flicker is seen as
frequency-doubled (Cavonius et al, 1992, Ophthal. Physiol. Opt, 12, 153-156).
The effect of induced motion on pointing in depth
Julie M Harris1 and Katie J German2 (1School of Psychology, University of St. Andrews, St.
Mary’s College, South St., St. Andrews, KY16 9JP, Scotland, UK; 2Psychology, School of
Biology, University of Newcastle upon Tyne, Henry Wellcome Building, Framlington Place,
Newcastle NE2 4HH, UK. E-mail: jh81@st-andrews.ac.uk)
An object appears to move if, although stationary, it is flanked by another moving object or
frame (termed induced motion, IM). As well as affecting relative motion perception, IM can
bias an observer's ability to point to the absolute locations of objects. It has been suggested
that IM may cause pointing errors due to mislocalisation of the body midline. Here we
studied observer's responses to objects that underwent IM in depth where there were no
changes in ego-centric direction. Observers viewed a stationary target dot flanked vertically
by a pair of inducing dots oscillating in depth (9.6 arcmin/s on each retina, amplitude 7.7
arcmin), at a viewing distance of 50cm. The CRT was located on a raised platform so that
observers could point to locations beneath it. Observers pointed (open-loop) to directly below
the start-location, then the end-location of the motion. We computed the IM as the difference
between start and end locations. This was compared with the difference between locations
pointed to for a target undergoing real motion in depth. We compared the proportion of IM
with that found in an IM nulling task, where observers were asked to adjust a target's motion
until it appeared stationary. There was a strong effect of IM on the pointed-to locations in
depth, similar to that for the IM-nulling perceptual task. Thus, IM can impact on open-loop
pointing for motion in depth. IM effects on pointing cannot be fully accounted for by
mislocalisation of the body midline.
How we judge the perceived depth of a dot cloud
David Keeble1, Julie Harris2 and Ian Pacey1 (1Department of Optometry, University of
Bradford, Bradford, BD7 1DP; 2School of Psychology, University of St Andrews, St Andrews,
KY16 9JU. e-mail: D.R.T.Keeble@Bradford.ac.uk)
-7-
AVA Annual Meeting 2006
Some visual properties such as position and texture orientation (Whitaker, et al, 1996, Vis.
Res., 36, 2957-2970) seem to be encoded as the centroid (ie mean) of the distribution of the
relevant variable. Although the perception of stereoscopic depth has been extensively
investigated little is known about how separate estimates of disparity are combined to
produce an overall perceived depth. We employed base-in prisms mounted on a trial frame
and half-images presented on a computer screen, to produce stereoscopic images of dot
clouds (eg as in Harris & Parker, 1995, Nature, 374, 808-811) with skewed (ie asymmetrical)
distributions of disparity. Subjects were to judge in a 2AFC task whether the dot cloud was in
front of or behind an adjacent flat plane of dots. Psychometric functions were generated and
the point of subjective equality was calculated. Three subjects were employed. For dot
clouds of thickness up to about 6 arc min disparity the perceived depth was close to the
centroid for all three subjects, regardless of skew. For thicker clouds, (we tested thicknesses
up to 25 arc min), the perceived depth was slightly closer to the subject than the centroid,
regardless of skew. The human visual system integrates depth information veridically for thin
surfaces. For thick surfaces, the dots closest to the observer have a higher weighting,
possibly due to their greater salience.
Spatial uncertainty governs the extent of illusory interaction between 1st- and 2ndorder vision
David Whitaker1, James Heron1 and Paul V. McGraw2 (1Department of Optometry, University
of Bradford, Bradford BD7 1DP; 2Department of Psychology, University of Nottingham,
Nottingham NG7 2RD. e-mail: d.j.whitaker@Bradford.ac.uk)
Illusory interactions between 1st-order (luminance-defined) and 2nd-order (texture-defined)
vision have been reported in the domains of orientation, position and motion. Here we
investigate one type of such illusion within a Bayesian framework that has not only proven
successful in explaining the perceptual combination of visual cues, but also the combination
of stimuli from different senses (e.g. ‘sound and vision’ or ‘touch and vision’). In short, the
perceptual outcome is considered to be a weighted combination of the sensory inputs, with
the weights being determined by the relative reliability of each component.
We adopt the ‘Stationary Moving Gabor’ illusion introduced by DeValois &DeValois (Vision
Research, 31, 1619-1626, 1991) in which a sinusoidal luminance grating (1st-order) drifting
within a static contrast-defined Gaussian envelope (2nd-order) causes the perceived position
of the envelope to be offset in the direction of motion. A vernier task was used to quantify the
extent of the illusion, and several manipulations of the stimuli were used to vary the spatial
uncertainty associated with the task. These included the spread (size) and spatial profile of
the envelope and the separation of the elements in the vernier stimulus. The 1st-order
characteristics were held constant.
Results clearly demonstrate that the magnitude of the illusion increases in line with the
spatial uncertainty of the 2nd-order envelope. As judgment of its spatial position is made less
reliable, so it becomes increasingly susceptible to an illusory effect.
-8-
AVA Annual Meeting 2006
Abstracts for Posters (in alphabetical order):
Motion-induced localization bias in an action task
Bettina Friedrich1, Franck Caniard2, Ian M Thornton3, Astros Chatziastros2 and Pascal
Mamassian4 (1Department of Psychology, University of Glasgow, 58 Hillhead Street,
Glasgow G12 8QB; 2Max Planck Institute for Biological Cybernetics, Spemannstraße 38,
72076 Tübingen, Germany; 3Department of Psychology, University of Wales Swansea,
Singleton Park, Swansea, SA2 8PP; 4CNRS & Université Paris 5, France. e-mail:
Bettina@psy.gla.ac.uk)
DeValois and DeValois (Vis Research, 31, 1619-1626) have shown that a moving carrier
behind a stationary window can cause a perceptual misplacement of this envelope in
direction of motion. The authors also found that the bias increased with increasing carrier
speed and eccentrcity. Yamagishi et al. (2001, Proceedings of the Royal Society, 268, 973977) showed that this effect can also be found in visuo-motor tasks. To see whether
variables such as eccentricity and grating speed increase the motion-induced perceptual shift
of a motion field also in an action task, a motor-control experiment was created in which
these variables were manipulated (eccentricity values: 0 deg, 8.4 deg and 16.8; speed
values: 1.78 deg/sec, 4.45 deg/sec and 7.1 deg/sec). Participants had to keep a downwardsliding path aligned with a motion field (stationary Gaussian and horizontally moving carrier)
by manipulating the path with a joystick. The perceptual bias can be measured by comparing
the average difference between correct and actual path position. Both speed and eccentricty
had a significant impact on the bias size. Similarly to the recognition task, the bias size
increased with increasing carrier speed. Contrary to DeValois and DeValois’ finding, here the
perceptual shift decreased with increasing eccentricity. There was no interaction of the
variables. If we assume an ecological reason for the existence of a motion-induced bias, it
might be plausible to see why the bias is smaller in an unnatural task such as actively
manipulating an object that is in an eccentric position in the visual field (hence the decrease
of bias magnitude in the periphery). Contrary to this, recognition tasks carried out in the
periphery of the visual field are far more common and therefore might “benefit” from the
existence of a motion-induced localization bias. As expected, task difficulty increased with
increasing speed and eccentricity. It seems interesting to further compare action and
perception tasks in terms of factors influencing the localization bias in these different task
types.
Attending to more than one sense - on the dissociation between reaction time and
perceived temporal order
James Hanson1, James Heron1, Paul V. McGraw2 and David Whitaker1 (1Department of
Optometry, University of Bradford, Bradford BD7 1DP; 2Department of Psychology,
University of Nottingham, Nottingham NG7 2RD. e-mail: j.v.m.hanson@Bradford.ac.uk)
Simple logic predicts that sensory stimuli to which we exhibit faster reaction times would be
perceived to lead other stimuli in a temporal order judgment task (‘which one came first?’).
This logic has received little empirical support. There is general agreement that reaction
times to auditory stimuli are considerably shorter than to visual. Despite this, temporal order
judgments involving the two types of stimuli usually require a small temporal lead of the
-9-
AVA Annual Meeting 2006
auditory stimulus in order to be perceived as simultaneous. We replicate these findings, but
argue that whilst simple reaction times are measured to one sensory modality in isolation,
temporal order judgments, by definition, require two modalities to be attended
simultaneously.
We therefore measure reaction times to auditory and visual stimuli randomly interleaved
within the same experimental run. This requires subjects to simultaneously monitor both
senses in an attempt to optimize their reaction times. Results show that reaction times to
both senses are significantly increased relative to uni-modal estimates, but auditory reaction
times suffer more (20-30% increase) than visual (6-10% increase). Nevertheless, whilst this
goes some way towards accounting for the dissociation between reaction time and temporal
order judgments, there still remains an unexplained visual lead in the perceptual judgment of
which sense appeared first.
Famous face recognition in the absence of individual features
Lisa Hill and Mark Scase (Division of Psychology, De Montfort University, The Gateway,
Leicester, LE1 9BH. e-mail: lhill@dmu.ac.uk)
The perceptual operations, which enable face recognition, require a synthesis of precision
and synchronicity. Insight into this process may be provided by exploring the timings and
sequence for which individual features are perceived. Previous studies have masked off
individual features along with the surrounding area (Lewis and Edmonds, 2003, Perception,
32, 903-920) and explored the feature by feature composition of faces from memory (Ellis,
Shepherd & Davies, 1975, Brit J Psychol, 66, 29-37). We investigated the effects of removing
individual features from famous faces, as opposed to masking areas of the face, in order to
examine whether the absence of any feature in particular, significantly affected reaction
times for face recognition. Participants viewed faces either without eyes, nose, mouth or
external features (hair and ears). Images were presented for 160ms and reaction times for
face identification were recorded. A significant increase in reaction times was observed
where faces were viewed without hair and ears compared to control conditions; implying that
during the initial stages of face perception, external features (lower spatial frequencies) may
act as the primary catalyst towards face recognition, before the visual system is able to
process the internal features of the face (higher spatial frequencies). This result has potential
implications for the revision of a pre-established test for prosopagnosia: the Benton Facial
Recognition Test (Benton et al, 1983, Contributions to neuropsychological assessment. NY:
OUP) as during testing, other cues such as hairline and clothing may be available to
participants in order to aid recognition, failing to provide an accurate account of actual face
recognition abilities (Duchaine & Weidenfeld, 2002, Neuropsychologia, 41, 713-720). Further,
these results help us understand the prominence of each feature in the face recognition
process, and the points at which perceptual abilities become fragmented
- 10 -