FOURSITES THIRD VISION RESEARCH SYMPOSIUM


  • Tuesday June 29th 1.00 p.m. Department of Human Sciences, Brunel University

  • Psychology Labs, Second Floor, Gaskell Building (Arts & Soc. Sci), West Spur Road, Uxbridge Campus.

    1.00 - 1.30 p.m. Arrival, set up posters. Buffet lunch available.

    1.30 - 2.30 p.m. keynote speaker:

    Mark Georgeson (Birmingham University)

    The role of spatial and temporal features in motion analysis - it's a bit derivative

    2.30 - 3.00 p.m. Robin Walker, D. Maurer, S. Mannan, A. Pambakian and C. Kennard

    (Royal Holloway).

    Naso-temporal asymmetries in saccade latency in normal and hemianopic subjects

    3.00 - 4.00 Tea + posters

    4.00-4.30 p.m. John Harris, Alison Lee & Sally-Ann McDowell,(Reading University)

    Why does Parkinson’s disease affect the ability to walk through doorways?

    4.30 - 5.00 p.m. David Rose (Surrey University)

    Anisochronic processing of features

    5.00 - 5.30 p.m. Michael Wright, Alison Green and Steve Baker (Brunel University)

    How many Gabor stimuli does it take to produce change blindness?

    5.30 - 6.30 p.m. Drinks and posters. Poster competition and prize.

    7 p.m. Participants are invited to join the guest speaker for dinner in a restaurant.

    We thank the Scientific Affairs Board of the British Psychological Society and the Department of Human Sciences for financial support.

    Those wishing to attend should inform Prof. Michael Wright (michael.wright@brunel.ac.uk) as soon as possible. Note to contributors: poster dimensions - 2 x 94x64cm landscape panels per poster; one above, one below.


    ABSTRACTS OF TALKS


    (T1)

    The role of spatial and temporal filters in motion analysis - it's a bit derivative

    Mark Georgeson

    University of Birmingham

    Visual information is recovered from the retinal image by chains of relatively simple image processing operations. The entire process is dauntingly complex, however, because of the diversity of ways in which information can be combined and re-transformed from one step to the next. In this talk I focus on the process of extracting motion information from changing sequences of images. As in spatial vision, so in motion it proves to be useful to think of the initial stages as linear filtering operations that serve to compute spatial and temporal derivatives of image luminance.

    Our experimental analysis of an intriguing phenomenon (movement aftereffect with no adapting motion) specifically implies that temporal differentiation (rather than a delay line) forms a key part of the early motion sensor. Considering the motion sensor to be the sum of spatial and temporal derivative operators, we then derive the spatial scale of these operators from novel experiments on 'motion acuity'. The implied receptive fields are shorter and less orientation-specific than expected. Later stages of the motion energy model are addressed in experiments on direction discrimination for dynamic gratings that contain energy in both leftward and rightward directions. Finally, I present new evidence to support the idea that motion energy detection is complemented by "feature-tracking", especially when there are gaps of 40 msec or more between successive frames of animation.


    (T2)

    Naso-temporal asymmetries in saccade latency in normal and hemianopic subjects

    Walker R, Maurer D, Mannan S, Pambakian A and Kennard C.

    Department of Psychology, Royal Holloway, University of London, Egham, Surrey, TW20 0EX

    Visual distractors have been used as an implicit test of residual visual functioning (or 'blindsight') in patients with cortical blindness. We examined the influence of visual distractors on saccade latency in hemianopic and normal subjects.

    Eye movements were made to a target presented in one hemifield, under monocular viewing conditions, and on some trials a distractor appeared in the opposite hemifield. For the hemianopic subjects distractors always appeared in the functionally blind hemifield. The hemianopic subjects were not aware of the distractors. Visual distractors in the hemianopes blindfield did not influence saccade latency. By contrast, the latency of saccades made by normal subjects was increased under similar distractor conditions. Furthermore, a small naso-temporal asymmetry was observed with a greater interference effect for temporal field distractors.

    These results are inconsistent with the view that the small crossed projection from the nasal hemiretina to the midbrain may mediate an inhibitory effect in the absence of the geniculostriate pathway.


    (T3)

    Why does Parkinson's disease affect the ability to walk through doorways?

    John Harris, Alison Lee & Sally-Ann McDowell,

    Department of Psychology, University of Reading, Whiteknights, Reading RG6 6AL

    Although locomotion is often impaired in Parkinson's disease (PD), the impairment tends to be worse while navigating confined spaces such as doorways, when the sufferer may "freeze" and have difficulty in continuing to walk. Two possible components of this problem ("distractibility" and the distortion of perceived space) have been investigated using psychophysical and questionnaire techniques.

    Manual reaction times to the onset of a visual signal were measured with and without irrelevant visual stimuli (fields of dots) surrounding the signal. In the PD group, but not the controls, RTs were slowed by the irrelevant stimuli. The slowing occurred when the offset, as well as the onset, of the distractors occurred at about the same time as the visual signal. Such impairment of motor performance during locomotion might arise from the looming of the edges of a doorway in peripheral vision. A questionnaire survey suggests that distractibility is more commonly reported by patients with right-sided symptoms (and so left hemisphere damage), an intriguing suggestion which we have yet to test in the laboratory.

    PD patients with left-sided motor symptoms (and so right hemisphere damage) show left visual neglect on line bisection tasks and and tend to overestimate the size of the aperture needed to allow their bodies to pass through (in contrast to right-sided patients who are essentially normal on such tasks, despite their motor problems). Thus a second problem for patients may lie in judging the size and position of the doorway. The work suggests that two separable visual components contribute to the akinesia of Parkinson's disease, and that they may be associated with damage to individual hemispheres.


    (T4)

    Anisochronic processing of features

    David Rose

    Dept. Psychology, University of Surrey, Guildford, Surrey GU2 5XH, U.K.

    Recent research on the perceived simultaneity of changes in different features of a visual stimulus have indicated a dramatic lack of binding between the changes. Moutoussis and Zeki (1997) report that simultaneous changes in (any pairing of) colour, motion direction and orientation appear asynchronous, by up to 120 ms. Zeki and Bartels (1998) found that brightness changes are seen synchronously with colour changes, and depth synchronously with motion. The results have been interpreted to indicate independent generation of many micro-consciousnesses in separate areas of visual cortex (Zeki, 1998).

    Nishida and Johnston (1999) however were only able to find asynchronies if the stimulus changes occurred repetitively at a fast rate. Moreover Barbur, Wolf and Lennie (1998) found much smaller processing asynchronies with a reaction time technique. In fact, dissociations of processing time have been commonly reported in the literature, and selective attention and stimulus salience can also influence the speed of processing. Further data will be presented confirming the dependence of relative processing time on experimental parameters.

    The results have important implications for Zeki's and Lennie's theories that awareness arises in separate cortical areas, and Dennett's theory (1991) of asynchronous processing in the brain. Alternative interpretations based on theories of specific time-processing modules will be discussed.


    (T5)

    How many Gabor stimuli does it take to produce change blindness?

    Michael Wright, Alison Green and Steve Baker

    Department of Human Sciences, Brunel University, Uxbridge, UB8 3PH, U.K.

    In typical change blindness studies observers fail to notice major changes in a natural scene, such as disappearance or displacement of an object or person, when the change coincides with a visual transient such as a scene shift or blank ISI. Rather than starting with a complex scene, we started with possibly the simplest and best understood visual stimulus, a Gabor patch, which we presented in two frames with an ISI. The observer had to report when the spatial frequency had changed, which it did on 50% of trials, and the change-detection threshold was expressed as a Weber fraction. Although it is relatively easy to detect spatial frequency change in a single target, it becomes progressively more difficult when the changing target is one out of two, three and four, and we do not know in advance which target will change. However if the changing target in a set of four is pre-cued, it is just as easy to detect the change as for a single isolated target. This argues against a perceptual limitation such as crowding.

    We plotted the log of Weber fraction against the log. number of targets and the points approximated to a straight line. The slope of this line is known as the set size effect. We obtained set-size effects which are steeper than those which have been reported for visual search, and which are consistent with a division of attentional resources across the potential targets. Thus even moderate complexity of a stimulus in terms of number of possible targets can overload attentional processes suggesting a possible low-level mechanism for change blindness.

    Our answer to the question posed in the title is thus "a few", but our experiments suggest that the "blindness" in change blindness is not absolute: change detection should depend on the complexity of the scene (no. of objects) and the magnitude of the change in terms of simple stimulus dimensions such as spatial frequency.


    ABSTRACTS OF POSTERS


    (P1)

    Evidence from horizontal and vertical line-bisection tasks for visuospatial neglect in left hemi-Parkinson's disease

    Alison Lee & John Harris,

    Department of Psychology, University of Reading, Whiteknights, Reading RG6 6AL

    Previous investigations indicate that there may be a problem with the perception of space in Parkinson's Disease (PD). Such problems may contribute to the difficulties PD sufferers have in navigating around the world, and particularly to akinesia (the inability to initiate movement).

    The perception of peripersonal and extrapersonal space was examined in two line-bisection tasks, one vertical and the other horizontal. 20 PD patients were divided into two groups, those with predominantly left-sided symptoms (LPD), and those with predominantly right-sided symptoms (RPD). Their performance was compared to 10 age- and gender-matched controls. Stimuli were displayed on a large screen (2*2.4m), with the horizontal lines displayed 10cms above eye level. The stimuli varied in size (48mm to 480mm) and in location on the screen (left space, centre and right space) and were presented at two viewing distances (0.6m and 1.5m).

    Results indicate that the LPD group made significant rightward and downward mis-bisections. Errors increased with the length of the stimulus line. For the horizontal bisection, errors were greater when the stimulus was displayed in left and central space. The RPD group performed similarly to the controls, making smaller leftward and upward errors. (Similar mis-bisections in the normal population are often referred to as pseudo-neglect.) Unlike the LPD group, the errors of the RPD and control groups did not increase in proportion to the stimulus line length. The data indicate the presence of a small, but consistent visual neglect in left hemi-Parkinsonism, which may reflect the degree of right brain damage caused by the disease to the basal ganglia and associated cortex.


    (P2)

    Horizontal line bisection bias in personal space in Parkinson's disease

    Liz Atkinson, Alison Lee, John Harris

    Department of Psychology, University of Reading, Whiteknights, Reading RG6 6AL

    The line bisection task from the Behavioural Inattention Task battery was used to examine horizontal bisection bias in unilateral Parkinson's disease (PD). This traditional test for neglect in personal space was used in conjunction with tests for neglect of peripersonal and extrapersonal space in PD. The stimuli were three horizontal lines of 208 mm, presented 60 mm apart on a sheet of A4 paper in ‘landscape’ orientation. The top line was closest to the upper right quadrant, the middle line was centred on the page, and the lower line was closest to the lower left quadrant. The A4 sheet was placed central to the mid-sagittal line of the body, and the patient had unlimited time to bisect the lines using a pencil held in the preferred hand (for 95%, the right). There were three groups of subjects: 22 right-handed controls, 9 PD sufferers with predominantly left sided-symptoms (LPD) and 13 PD sufferers with predominantly right-sided symptoms (RPD).

    The control group made the characteristic leftward errors reported in other studies (pseudo-neglect), with their largest errors on the most central line. Both patient groups showed a tendency to bisect the right-hand line too far to the left, and the left-hand line too far to the right. This may reflect the order in which the lines were bisected, or their relationship to the edge of the paper, issues which we plan to investigate further. However, this was superimposed on a tendency in the LPD group to bisect all lines further to the right than the RPD group, which is similar to that found in peri- and extra-personal space. One lesson from the study is that, when measuring relatively weak neglect with clinical tests, the nature of the task is important.


    (P3)

    The role of Stochastic Resonance in the detection of signals in noise

    David R R Smith

    Department of Experimental Psychology, University of Sussex, Falmer, Brighton BN1 9QG, UK; fax: +44 1273 678 611; email: dsmith@biols.susx.ac.uk

    Stochastic resonance is a non-linear cooperative effect whereby the response of a non-linear system to a weak periodic stimulus is enhanced by the presence of an optimal level of input noise. In other words, adding noise to a system can help in the detection of an otherwise invisible signal. The effect of superimposing two-dimensional static Gaussian white noise (RMS contrast 0-0.39) upon the contrast detection thresholds (CDTs) of one-dimensional stationary gabor patches was measured.

    A two temporal interval 2AFC paradigm was used to measure performance in signal detection tasks. Noise was added to both temporal intervals. The signal was displayed in one interval. Observers reported in which temporal interval the signal was present. CDTs were obtained in the presence of noise for (a) foveally presented gabor signals possessing a range of spatial frequencies (1-8 c/deg), (b) a 2 c/deg gabor signal presented 2.8 deg above the point of fixation and (c) a foveally presented 2 c/deg gabor signal at two presentation durations (13 and 80 msec).

    Facilitation (lowered CDTs) occurs when noise (up to approximately 0.15 RMS) is added to the signal. Masking (elevated CDTs) occurs for greater noise amplitudes. Facilitation is most marked for 2 c/deg gabors (maximal facilitation for 0.08 RMS noise where the CDT is lowered by a factor of sqrt(2) compared to no noise baseline). Facilitation effect is robust across the tested retinal locations and length of presentation.


    (P4)

    Suppression of activity in human visual cortex by visual stimulation

    AT Smith, KD Singh, MW Greenlee

    Department of Psychology, Royal Holloway, University of London, Egham, Surrey, TW20 0EX

    Functional MRI methods were used to examine the activity generated in the visual cortex by simple grating patches (12 deg diam.) which were either drifting or counterphasing and were presented in a conventional on/off sequence (27s per phase). Imaging was performed on several healthy observers using a 1.5T Siemens scanner. Retinotopic mapping and flattening of the grey matter were also performed to allow the different visual areas (V1, V2 etc) to be distinguished.

    As expected, grating stimuli produced a swathe of activation in occipital cortex encompassing all retinotopic visual areas but restricted to the central portion of the field. The effects of spatial and temporal frequency were quantified; to be reported elsewhere. Beyond the central region of the visual field, large areas of negative activation were seen (i.e. the cortex was less active when the stimulus was present than when absent). This might be due either to (a) long-range inhibitory interactions between different portions of the visual field, or (b) attentional modulation (attention is focussed in the centre when a grating is present but is diffuse when it is absent, so that peripheral parts may receive more attention when the stimulus is absent).

    We performed additional studies in which attention was modulated by changing the colour of the fixation spot and requiring the subject to count the occurrences of a particular colour. When the task was performed throughout the run, or only when the grating was absent, the negative activation disappeared. When it was performed only when the grating was present, the effect was enhanced. These results suggest that attentional modulation, rather than long-range inhibition, is responsible. Conclusion: attentional modulation occurs in V1 and V2 (maybe V3) and it occurs even in the absence of visual stimulation.


    (P5)

    Temporal integration of binocular disparity depends on surface orientation

    Paul Hibbard and Mark Bradshaw

    Dept. Psychology, University of Surrey, Guildford, Surrey GU2 5XH, U.K.

    Purpose. The existence of a bandpass oriented filter in the cyclopean domain is supported by a range of psychophysical results (e.g. Tyler 1991, Binocular Vision D. Regan). Such a filter should be characterised by a single temporal impulse function irrespective of surface orientation. To test this hypothesis we compared the temporal integration period of disparity-defined corrugations oriented vertically and horizontally. In further experiments we measured the two-pulse integration time, and the trade off between presentation time and perceived depth, for disparity-defined corrugations in both orientations.

    Methods. Dynamic random dot stereograms subtending 5 degrees were presented at a frame rate of 120Hz. In experiment 1 disparity thresholds were measured for horizontal and vertical corrugations (0.2 to 0.8 cycles/deg) at different presentation times from 50 to 8000ms. In experiment 2 the temporal integration limits for horizontal or vertical corrugations were determined by varying the temporal separation of two signal frames (with a fixed disparity). In experiment 3 a depth matching paradigm was used to investigate the effect of presentation time on the perceived depth of horizontal and vertical corrugations.

    Results. Thresholds decreased as presentation time was increased up to 1000ms for horizontal gratings, but continued to improve over longer durations for vertical gratings. This difference was not observed in the results of the two pulse integration experiment. In the final experiment, matched depth was a simple function of the magnitude of disparity and presentation time, and the build up of perceived depth was similar for horizontal and vertical surfaces.

    Conclusions The temporal integration period differed for vertical and horizontal disparity-defined corrugations. This difference was only manifest in the disparity threshold-duration functions. This dependence of temporal integration of binocular disparity on surface orientation cannot be accounted for by a single, simple cyclopean mechanism.


    (P6)

    A RESTRICTED FIELD OF VIEW AFFECTS REACHING AND GRASPING

    S.J. Watt, M.F. Bradshaw and S.K. Rushton

    Dept. Psychology, University of Surrey, Guildford, Surrey GU2 5XH, U.K.

    Dolezal (1982, Academic Press) observed that wearing goggles which restricted field of view (FOV) caused familiar objects to appear both smaller and nearer. They also produced significant `under-reaches'. These observations have important ramifications for studies of perception and action which manipulate visual displays with a textureless or occluded surround. Consistent with this, Sivak and MacKenzie (1990) found that lack of peripheral information affected the transport component (but not the grasp) of reaching movements. However their subjects reached to objects located at only one distance and so the nature of the possible transformation of visual/haptic space could not be determined. We addressed the issue of restricted FOV further with 2 experiments.

    Experiment 1 examined the effects of extent of FOV on reaching. Subjects made open-loop reaches under binocular viewing to 2-D objects placed along the midline, under 5 FOV conditions (4°, 8°, 16°, 32° & 64°). Experiment 2 investigated the possible transformation of visual/haptic space. Subjects reached to three distances along the midline, with 8° and 64° FOV. Reaches were made under binocular and monocular viewing to establish whether the effects of reducing FOV arise from changes in monocular information, or from the degradation of binocular information. In experiment 1, reducing the FOV produced progressively greater under-reaches and slower peak wrist velocities.

    In experiment 2, the magnitude of the under-reaches increased with increasing object distance. There was no effect of binocular versus monocular viewing, suggesting that the effects of restricting FOV are based primarily on the degradation of monocular information. We conclude that the extent and nature of the visual surround must be taken into account in the interpretation of experiments in reaching and grasping.

    Supported by the BBSRC, the Royal Society and the Wellcome Trust


    (P7)

    PHOTOPIC AND SCOTOPIC PERFORMANCE OF DYSLEXICS AND NORMALS ON STEREO AND MOTION TASKS.

    D.J.K. Barrett, M.F. Bradshaw, P.B. Hibbard & J. Everatt

    Dept. Psychology, University of Surrey, Guildford, Surrey GU2 5XH, U.K.

    At least two anatomically distinct pathways project from the retina to the cortex; the parvocellular and magnocellular. It is thought that dyslexics may have a deficit in the M-pathway (Stein & Walsh, 1997) which could account for their impaired performance on certain psychophysical tasks (Everatt et al, 1998). However, the deficit may also arise from inappropriate interactions between the two pathways. To distinguish between these hypothesis, we compared the performance of normals and dyslexics on three visual tasks related to M and P function in photopic and scotopic conditions. The latter having been shown to diminish the operation of the P pathway (Purpural et al, 1988).

    Thresholds for dyslexics (N15) and controls (N15) were determined using a TUD staircase procedure in three tasks: (I) coherent motion (SNR), (ii) disparity threshold (corrugation phase detection) and (iii) biological motion (SNR). The stimuli comprised bright dots (13 cd/m2) or dim dots (0.05 cd/m2) in the photopic and scotopic conditions respectively, on a dark background (0.01 cd/m2). Before commencing the scotopic trials, subjects dark-adapted for 30mins.

    Disparity and biological motion thresholds did not differ significantly between the groups although both groups performed worse under scotopic conditions (p <0.01). Thresholds for coherent motion detection was 20% and 32% for controls and dyslexics respectively in photopic conditions whereas in scotopic conditions this difference was not observed. Our results suggest that when interpreting the performance of dyslexic and normal observers in psychophysical tasks the interaction of both M and P pathways must be considered.


    (P8)

    Illusory conjunctions in patients with right parietal lesions: evidence for attentional impairment producing binding errors.

    Tracey Brown and Michael Wright

    Department of Human Sciences, Brunel University, Uxbridge, UB8 3PH, Regional Rehabilitation Unit, Northwick Park Hospital, Harrow, Middx.

    Explanations of illusory conjunctions (IC's) are of two main types: there is an overload of attention or there is uncertainty about the location of features. Both theories were tested in a mixed design using 3 values of attention load and 2 of feature proximity. Participants were 3 patients with right parietal lesions (RP), 2 control patients with left hemisphere lesions (BDC) and 27 age-matched normal participants (NC, NC2).

    4 Objects of different shape (circle, square, triangle) and colour (red, blue, yellow) were presented for 300 msec, and the task was to say whether a red triangle was present (condition 1). The stimuli appeared in a vertical line, to the left or right of fixation, and were flanked by two letters. Attentional load was increased by requiring subjects additionally to name a letter presented at fixation (condition 2) or report a letter plus a tone (condition 3). An IC was defined as a false positive occurring when both a red object and a non-red triangle were present. There was a very low rate of false positives in the absence of red objects and triangles.

    All three Right Parietal (RP) patients showed many more IC's than left hemisphere controls (BDC) or 7 controls with intact brains (NC's), and the number of their IC's rose steeply with attentional load. However at this relatively long exposure time, both NC and BDC scores were subject to a floor effect, so another 20 normal control participants (NC2) were run at a briefer exposure of 200 msec. Analysis of variance on the data from RP and NC2 showed that the main effect of group was balanced out when the normals were tested at short exposure. There was a main effect of feature proximity (F=9.52; df=1,21; p<0.01) in that IC's were more common for adjacent than non-adjacent stimuli, in both RP and NC2 groups. There was also a main effect of attentional load in both RP and NC2 groups (F=6.51; df=2,42; p<0.005). The only difference between the RP and NC2 groups was the interaction between group and attention level (F=4.09; df=2,42; p<0.05) in that each patient in the RP group showed the same steep increase in IC's with increasing attention load, whereas this did not occur in the NC2 group.

    It is argued a) that the findings support the idea that both focal attention and proximity influence feature binding and b) that RP patients make more IC errors because they are impaired in distributing or shifting visual attention, rather than in encoding of location.


    (P9)

    Drawing the trees but not the wood: copy-drawing of Navon figures in constructional apraxia.

    Angela Merry and Michael Wright

    Department of Human Sciences, Brunel University, Uxbridge, UB8 3PH, Regional Rehabilitation Unit, Northwick Park Hospital, Harrow, Middx.

    24 patients undergoing rehabilitation following stroke were asked to copy a global letter (e.g. H) composed of local letters (e.g. T) (Navon, 1977; Cognitive Psychology, 9, 353-383). 11 patients (5 l.hem, 4 r.hem, 2 bilat.) showed constructional errors in drawing, as categorised by 6 independent raters. Dissociations were found, in that errors could occur independently in the global shape, in the local shape, or in the number of local elements. Errors were more common in the global shape or the number of local elements than in the local shape.

    The errors were analysed using an information- processing model of copy drawing (Roncato, et al. 1987; Cognitive Neuropsychology, 4, 113-129). It is necessary to modify this model to include a hierarchical (local/global) shape schema to account for the errors seen. Models of this kind provide a principled method for classifying the errors found in copy drawing, and are thus useful for the analysis of constructional apraxia.


    (P10)

    Vision and Action

    Stephen J, Anderson

    Department of Psychology, Royal Holloway, University of London, Egham, Surrey, TW20 0EX

    Vision and action are intimately linked: it has been argued that the perception of common graspable objects automatically potentiate motor response codes regardless of whether or not there is an intention to act (Tucker & Ellis, 1998). It is generally believed that objects potentiate components of the actions they afford. We examined this hypothesis using a Stimulus Response Compatibility (SRC) paradigm, which refers to the relative advantage of certain stimulus-response mappings over others (see Alluisi & Warm, 1990, for a review). The stimuli included everyday objects, which either afforded an action or not, and non-object patterns: each was represented as a line drawing on an Eizo display monitor and viewed binocularly at 0.5 m. Our measure was reaction time: subjects were instructed to press the 'right' response button with their right hand if the object was rotated clockwise and the 'left' button with their left hand if it was rotated anti-clockwise. The object was always centred on fixation. When analysing the results, attention was paid to the effect of two 'irrelevant' object properties: (1) whether the object afforded an action or not (e.g. an analogue clock was deemed not to afford an action whereas a pair of scissors and a wine glass were); and (2) whether the object was left-asymmetric, right-asymmetric or symmetric (e.g. a clock time of 3.15 was deemed right-asymmetric; a time of 9.45 was deemed left-asymmetric; a pair of scissors with handles to the right of fixation was deemed right-asymmetric; a wine glass was deemed symmetric).

    The pattern of results was near identical for all object types: left-hand responses were significantly faster for left-asymmetric objects than right-asymmetric objects, whereas right-hand responses were fastest for right-asymmetric objects. The left- and right-hand response times were not significantly different for symmetrical objects. Importantly, similar results were obtained for asymmetric and symmetric non-object patterns. These results provide evidence that the factor determining the potentiation of motor response codes is not the action afforded by an object, but rather the presence or absence of directed visual attention. Other results show that directed attention was only present with asymmetrical patterns. These results do not support the view that seen objects automatically potentiate components of the actions they afford. We suggest that some motor codes must be elicited prior to any stage of semantic processing, and that some structural features of the object (such as asymmetry) facilitate the activation of potential motor actions. This effect is believed to be mediated by the dorsal pathway.


    (P11)

    Tilt and size aftereffects in motion parallax, stereopsis and luminance.

    M.F. Bradshaw, P.B. Hibbard & M.A. Hogervorst

    Dept. Psychology, University of Surrey, Guildford, Surrey GU2 5XH, U.K.

    We have recently shown that motion parallax, like luminance and stereopsis, is tuned to particular corrugation frequencies (Hogervorst et al, ECVP, 1998). Here we investigate the nature of these mechanisms further by determining whether perceived (i) corrugation frequency and (ii) corrugation orientation are altered by adaptation to parallax defined corrugations of particular frequencies (spatial frequency shift AE) and orientations (tilt AE). Both phenomena exist in the luminance and stereopsis domains (Blakemore and Sutton, 1969, Science 166, 245-247; Tyler, 1975, Perception 4, 187-192). We also address whether adaptation in one domain (e.g. luminance) affects the perceived size or orientation defined in another domain (e.g. parallax).

    Observers adapted to two patterns (5 deg in diameter) arranged in a horizontal 'dumbbell' configuration. The 1cpd adapting patterns differed in either orientation (± 15, ±30, ±45 or ±60 degrees) or in corrugation frequency (±0.125, ±0.25, ± 0.5, ±0.75 octaves). A staircase procedure was employed to determine when the test surfaces (left and right parts of the dumbell) appeared to be of equal orientation or corrugation frequency. The initial period of adaptation was 1 mins followed by a test-adaptation cycle of 2 sec test and 6 sec adaptation. (No phase dependent aftereffects were produced.). Both size and orientation aftereffects were found within the motion parallax domain. The peak magnitude of the aftereffect (~ 4 degrees), the location of its peak (30 degrees) and its estimated bandwidth (25 degrees) were similar to those already established, and replicated here, in the cyclopean domain (Tyler 1975, op cit). The peak magnitude of the size aftereffect was ~12% in all three domains and occurred with adapting patterns which differed by 0.5 octaves. Cross adaptation of size and orientation was possible between parallax, stereo and luminance defined contours.

    The visual system contains parallax channels which are tuned to frequency of depth modulation and orientation. Cross adaptation of size and orientation is possible.


    (P12)

    The walker's direction affects the perception of biological motion

    R. van der Willigen, M.F. Bradshaw, R. Leach, P. B. Hibbard and S. K. Rushton.

    Dept. Psychology, University of Surrey, Guildford, Surrey GU2 5XH, U.K.

    Previous studies of biological motion have typically investigated spatio- temporal patterns of point-light displays which specify 'walkers' moving across the visual field orthogonal to the line of sight. However, this is a limited case as in everyday life we view humans walking in many different directions. In the present experiments we investigate whether the direction of walk affects the perception of biological motion.

    Real walks in four different directions (0, 30, 60 and 90ƒ to the image plane) were video recorded and converted into light displays containing 13 points (one on the head and two on the shoulders, elbows, wrists, hips, knees and ankles). In the natural, ecological case, walkers will change in both position and size by varying amounts depending on direction. To disaggregate these effects we created three further conditions in which global changes in position and/or size were eliminated. These were shown to 12 subjects in a 2AFC design where an interval containing 'the walker' plus noise had to be identified from an interval containing only noise. The SNR for this discrimination was determined for four types of walk where (i) the walker changed position and size appropriate to the direction of walk, (ii) the walker did not change position (iii) the walker did not change size, and (iv) the walker did not change position or size. In each case observers were more sensitive to walker motion along the line of sight and thresholds appeared greater in the condition where no position or size changes occurred. Surprisingly perhaps, discrimination performance was best in the condition for which neither global position nor size changes existed.

    Overall, observers showed best performance for walks along the line of sight. This agrees with Johansson's (1976) findings and the differential sensitivity may reflect the social significance in perceiving approaching figures (see also Oram and Perrett, 1996) which remains even when the global expansion component is taken out of the stimuli.


    (P13)

    Complete crossing of the visual pathways in a human albino: fMRI measurements.

    A.B. Morland

    Imperial College, London

    The normal path of the optic nerve from the eye involves the projection of temporal fibres from the eye to the same (ipsilateral) side of the brain, and the crossing of nasal fibres to the opposite (contralateral) side of the brain. The net result of this semidecussation is that information concerning one visual hemifield from each eye is convergent on the contralateral dorsal lateral geniculate nucleus (dLGN). From the dLGN the afferent fibres form the optic radiation and project to the primary visual cortex. So, the primary cortical visual representation in one hemisphere comprises a map of the contralateral visual hemifield. In the albino visual system, however, the majority of all fibres from each eye cross to contralateral brain at the optic chiasm. This abnormal descussation means that the information concerning the whole visual field from one eye projects to the contralateral visual cortex. The information concerning albino visual pathways has only been derived in non-human vertebrates. In this study, a human albino underwent functional magnetic imaging whilst viewing visual stimuli through either eye. The results indicate that there is a complete crossing of the central 5 deg of visual field in this human albino. Therefore, human albinos appear to have visual projections that are consistent with previous experiments on animals.