FOURSITES FIFTH VISION RESEARCH SYMPOSIUM

 Thursday June 21st 12.30 p.m. Department of Human Sciences, Brunel University

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

http://www.brunel.ac.uk/campus/uxbridge/


12.30 - 1.00 p.m. Registration, Poster Session 1. Buffet lunch available.

1.00 - 1.30 p.m. A.T. Smith, A.L. Williams, K.D. Singh
(Royal Holloway, Liverpool University)
fMRI measures of receptive field  size in human visual cortex

1.30 - 2.00 p.m. M.B. Hoffmann, A.B. Morland, A.T. Moore, D.J. Tolhurst
(Royal Holloway, Addenbrookes Hospital, Cambridge University)
Organisation of the visual cortex in human albinos

2.00 - 3.00 p.m. KEYNOTE SPEAKER: Peter Lennie
(New York University)
Efficient Representation in Visual Cortex

3.00 - 4.00 p.m. Poster Session  2 + tea

4.00 - 4.30  p.m. Louise Alston and Michael J Wright
(Brunel University)
Discrimination and localisation of change in multiple targets
.
4.30 - 5.00 p.m. Simon J Watt, Mark F Bradshaw
(Surrey University)
Binocular cues, motion parallax and the control of natural prehensile movements

5.00 - 5.30 p.m. John Wann,
(Reading University)
Where do we look when we steer and does it matter ?

5.30 - 6.00 p.m. Johannes M. Zanker,  Nick Fisher,
(Royal Holloway).
On the Directional Tuning of the Barber Pole Illusion

6.00 - 7.00 p.m. Drinks and Poster session 3. Poster competition and prize.

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

We thank all the Foursites Departments 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



KEYNOTE SPEAKER: Professor Peter Lennie

Efficient Representation in Visual Cortex

Peter Lennie, New York University, Center for Neural Science, New York, NY  10003, USA

Questions about the efficiency of representation need to be asked in a particular context, because design to optimize performance in one domain might impair performance in another.  I shall explore efficient representation in the context of redundancy reduction. Visual signals are highly redundant, in both space and time.  Retinal mechanisms remove some of the redundancy in images, but more is removed by mechanisms in cortex.  I shall examine the operation of these cortical mechanisms, expressed in the behavior of single neurons that rapidly adjust their sensitivities to images that persist in space and time, while remaining very sensitive to image changes.  I shall then go on to consider why it might be important for the brain to have developed machinery that seems so well-suited to removing redundancy.


FMRI measures of receptive field  size in human visual cortex

A.T. Smith , A.L. Williams, KD Singh1

Dept of Psychology, Royal Holloway, University of London.
1 Liverpool University; present address: Aston University, Birmingham.
E-mail: a.t.smith@rhul.ac.uk

Functional magnetic resonance imaging (fMRI) was used to estimate the average receptive field (RF) sizes of neurons in each of several occipital visual areas of the human cerebral cortex. Imaging was performed on healthy subjects using a 1.5T system. Local variations in blood oxygenation (BOLD response) were measured using echoplanar imaging. Anatomical images were also acquired and were used to determine the anatomical localization of functional responses. The boundaries of the various visual areas were determined by standard retinotopic mapping procedures (exploiting stimulus-location-related variations in the temporal phase of the response to a rotating wedge stimulus) and were visualised on flattened representations of the occipital cortex. Estimates of RF size were derived from the temporal duration of the activation at each cortical location as a visual stimulus passed through the RFs represented at that location. The stimulus was a slowly expanding patterned ring which made several cycles of expansion over several minutes. This gave a periodic response at any given location in visual cortex and the duty cycle of this response (measured as the duty cycle of the best-fitting rectangular wave) should reflect average RF size at that location. In Experiment 1, the stimulus was a spatially broadband pattern (a high-contrast, 8Hz counterphasing checkerboard), chosen to drive as many different neurons as possible. The results show that RFs are smallest in the primary visual cortex (V1). They are larger in V2, larger again in V3/VP and largest of all in areas V3A and V4v. In all these areas, RFs increase in size with increasing stimulus eccentricity. These results are qualitatively in line with those obtained by others in macaques using neurophysiological methods. In Experiment 2, the stimulus was a ring of 2-d visual noise which was filtered with each of a range of one-octave bandpass spatial frequency filters. Again, contrast was high and was reversed at a rate of 8Hz. In V1, at least, such a stimulus should isolate neurons of a particular RF size (because size is closely related to spatial frequency sensitivity), irrespective of visual field location. The results show that in each visual area examined (V1,V2, V3/VP), mean RF size fell sharply as the filter centre frequency was increased, but was completely invariant with stimulus eccentricity (in contrast to Experiment 1), confirming this expectation. However, RF size still increased from V1 to V2 and from V2 to V3/VP, as in Experiment 1, despite the successful isolation of neurons sensitive to limited spatial frequency bands in all areas. This suggests that the reason why RFs in V2 are larger than those in V1 is that summation of V1 neurons with common spatial frequency sensitivities but slightly different spatial locations occurs, yielding V2 neurons that are still tuned for spatial frequency but have a higher ratio of RF size to optimum spatial frequency. Further summation across space appears to occur in V3/VP. RFs are not larger because neurons in V2/V3 are sensitive to lower spatial frequencies than V1, nor because of integration of V1 outputs across spatial frequencies.

Acknowledgements: Supported by the Wellcome Trust. Experiment 1 was performed in collaboration with Prof MW Greenlee, University of Oldenburg.

Organisation of the visual cortex in human albinos

M.B. Hoffmann1, A.B. Morland1, A.T. Moore2, D.J. Tolhurst3
1Royal Holloway, Univ. London; 2Addenbrooke’s Hospital, Cambridge; 3Univ. Cambridge, UK

Introduction: It is unknown how the visual cortex of human albinos is reorganised in response to the abnormal visual input caused by their enhanced crossing of the optic nerves. The conseqeunces of this misrouting was assessed with fMRI employing retinotopic mapping procedures. Methods: Two albinos with stable fixation and two control subjects underwent T2* MRI scanning of the occipital lobe during visual stimulation. In separate experiments we stimulated monocularly the nasal and temporal retina with a phase reversing chequerboard stimulus to map polar angle representations (Engel et al., 1997): Seven 36 s cycles of the stimulus stepping through the central 5 degrees of the visual field as a rotating wedge were presented. 128 x 128 voxel images (voxel size: 1.8 x 1.8 x 4 mm) were acquired in 8 planes every 3 s. fMRI signals were projected onto the flattened representation of T1 weighted images, Fourier analysed, and correlated with the stimulus fundamental frequency. The response phase indicates the preferred polar angle of a voxel. Results: We obtained highly correlated orderly maps for polar angle representations which enabled the identification of the early visual areas. The misrouting was clearly evident in the albino subjects: In the subjects with albinism the hemisphere contralateral to the stimulated eye was activated whereas it was the hemisphere contralateral to the stimulated visual field that was activated in the normal controls. In the two albinos tested the cortical retinotopic representations of the contralateral and of the ipsilateral visual field overlaid each other. Conclusions: Our measurements provide the first evidence for a  reorganization of human albino visual cortex following the “Albino Pattern” (Guillery 1986): In the albino tested the normal contralateral and the abnormal ipsilateral representations of the visual field in the early visual areas are non-exclusive.
References: Guillery RW (1986) TINS 9:364-367; Engel AS  et al. (1997) Cerebral Cortex 7:181-92
 

Discrimination and localisation of change in multiple targets

Louise Alston and Michael J Wright

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

Discrimination thresholds for spatial frequency or contrast in two-interval tasks (with an ISI) increase rapidly with the number of targets, suggesting that the problem of detecting a signal in noise contributes to phenomena such as “change blindness”. We have found that the steepness of the set-size effect depends on whether the stimuli are processed frame by frame (visual search), compared globally, or compared individually (Wright, et al 2001). The present experiments provide further evidence for global processing in discrimination of target change, combined with individual processing of items during localisation.

In visual search, Baldassi and Burr (2000) found that biasing distracter values in the same direction as the target enhanced discrimination, but weakened localisation because it relies on target contrast. Biasing them in the opposite direction weakened discrimination, but enhanced localisation. Using a two-frame discrimination of spatial frequency, our results clearly show the same pattern for change discrimination as Baldassi and Burr reported for visual search.

We define same-direction change bias as a change in the distracters in the same direction as the change in the target, and opposite-direction change bias as a change in the distracters opposite to the change in the target. In the first experiment we found that same-direction change bias improves discrimination and worsens identification of changes. Opposite distracter change bias improves localisation and worsens discrimination. This suggests global processing of stimuli in change discrimination and local processing in change localisation.

In the second experiment, we randomised the initial spatial frequency of the distracters, reducing the ability to identify the target within a single frame. Non-uniform distracters increased discrimination thresholds, and led to very high localisation thresholds. The effect of the distracter change bias on these increased thresholds was similar to that of the first experiment. In typical "change blindness" scenarios, observers are asked to identify and localise the change. We show here that localisation of change is very difficult without single frame feature contrast or focal attention. (Supported by a project grant from the Wellcome Trust).

References:
Baldassi S & Burr, DC (2000) Feature-based integration of orientation signals in visual search. Vision Research, 40, 1293-1300.
Wright, MJ, Alston L & Popple A (2001) Set-size effects for spatial frequency change and discrimination in multiple targets. Spatial Vision, in press.

Binocular cues, motion parallax and the control of natural prehensile movements

Simon J Watt

Department of Psychology, University of Surrey, Guildford GU2 7XH, UK

In order to perform accurate and efficient prehensile movements, vision is required to specify the extrinsic and intrinsic properties of the object to be grasped.  The primary sources of information about the three-dimensional properties of objects are binocular cues and motion parallax.  Here we determine the efficacy of each of these cues to control reaching and grasping movements.  In a between-groups design, subjects reached for and grasped real objects viewed (i) binocularly, (ii) monocularly with head motion, or (iii) binocularly with head motion.  The stimuli were presented in a carefully controlled scene from which almost all other visual cues to distance and size had been removed.  Analysis of peak wrist velocities and
peak grip apertures revealed that reaches executed with only binocular viewing showed normal scaling of the transport and grasp components with object distance and size, respectively.  When only information from motion parallax was available the transport component again scaled reliably with object distance.  In this condition, however, grip apertures were unrelated to changes in object size, suggesting that this information was unavailable for the control of prehension.  Performance in the combined binocular viewing and head motion condition was equivalent to that when only binocular information was available.  This suggests that the poor grasping performance observed in the monocular head motion condition was not
attributable to possible difficulties in reaching after making head movements.  There was no evidence of improved performance when both cues were available simultaneously.  These results suggest that binocular vision can potentially make an important contribution to the control of reaching and grasping movements.  However, even though equivalent information is potentially available from motion parallax, it may be of comparatively limited use in the control of grasping.
 
 

Where do we look when we steer and does it matter ?

John Wann,

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

Land & Lee (1994) documented that car drivers direct gaze to the tangent point of the curve when steering a bend.  Wann & Land (2000) described how fixating on an appropriate point on the road (tangent of the curve or the path) can simplify the task of steering, by altering the pattern of retinal flow, or by providing an alternative estimate of egocentric rotation.  The
first part of this study revisited Land & Lee, using head-mounted gaze tracking and simulated roadways presented on a 90deg FoV screen.  Gaze records were digitized and catalogued by researchers who were not aware of the proposals in Wann & Land.  The results suggest that for the majority of the time, during bend steering, gaze is directed to a point on the future
path (centre of the road) rather than the tangent point.  This experiment was then repeated with a divided attention task, of monitoring both the road and a speed indicator, and similar results were obtained.  Finally a series of trials were constructed where gaze location was cued by small transparent indicators that were placed within the scene to test whether stabilizing gaze or fixating on the anticipated path resulted in more accurate control.  The general pattern of results suggest that gaze
fixation is an important component of controlling steering, this in turn modifies the debate as to what properties of the visual flow field are most important for locomotor control.

Supported by the UK EPSRC

On the Directional Tuning of the Barber Pole Illusion

Johannes M. Zanker 1,2 , Nick Fisher 2

Department of Psychology, Royal Holloway University of London, Egham, England 1;
Centre for Visual Sciences, RSBS, The Australian National University, Canberra, Australia 2

In order to study the integration of local motion signals in the human visual system, we studied how stimulus geometry affects the Barber Pole Illusion, i.e. the tendency to perceive motion along the major axis of an elongated aperture. We measured directional tuning curves for the Barber Pole Illusion by varying the orientation of a grating presented behind a rectangular aperture of variable aspect ratio, and asking the subjects to report perceived direction and the strength of their motion percept. Our experiment reveals different strategies of the visual system to deal with the Barber Pole stimulus. Some individuals respond strongly to the unambiguous motion information at the boundaries, leading to multistable percepts and multimodal distributions of responses. Others tend to report intermediate directions, apparently being less influenced by the boundaries. In general, deviations from the motion direction perpendicular to grating orientation decrease (a) with aspect ratio approaching unity, i.e. square-shaped apertures, and (b) with gratings approaching parallel orientation to the shorter aperture boundary This pattern of results is discussed in the context of simple phenomenological models of motion integration. The best fit between model predictions and experimental data is achieved by the interaction between two stimulus parameters: (i) cycle ratio, which is the sinewave gratings equivalent of the terminator ratio for line gratings, describing the effects from the aperture boundaries, and (ii) the grating orientation, responsible for perpendicular motion components, which describes the effects from inside the aperture. It thus seems that the most simple cycle (or terminator) ratio explanation cannot fully account for the quantitative properties of the Barber Pole Illusion.


ABSTRACTS OF POSTERS:

Metamerism in Spatial Vision

L D Griffin

Medical Imaging Science IRG, King's College, London

Metamerism is a phenomena of spatial vision as it is of colour vision. In colour vision, distinct spectra can give rise to the same cone responses. In spatial vision distinct luminance patterns can cause the identical responses in visual neurons (in,
for example, V1). Thus, since the firing rates of visual neurons are insufficient to uniquely determine the luminance structure within the extent of their receptive fields, it is unclear how such rates should be interpreted by higher visual areas. Building on work of Koenderink, I have explored the possibility that one could associate with a particular pattern of firing activity the luminance pattern that is consistent with it and is as simple as possible. In particular I have explored the consequences of defining simple in terms of the L1, L2 and L-inf norms. All three norms can be understood as measures of contrast, the L-inf norm being contrast defined as the range of luminance values and the L2 norm being contrast defined as luminance variance.
Using convolution with derivatives of 2D gaussian kernels as models of the measurements performed by the receptive fields of visual neurons, I have developed algorithms that discover the image that is consistent with any given set of measurements while minimizing the L1, L2 or L-inf norm. I have also proved the nature of these minimizing images. In the case of the L-inf norm, Koenderink's guess is shown to be mainly correct - the images are binary-valued with a simple transition locus. However, Koenderink's assertion that the locus of change between the two values is a nth-order algebraic curve is incorrect as the locus can be only a subset of such a curve. In the case of the L2 norm, the minimizing image is a sum of gaussian derivatives of various orders. Finally, for the L1 norm the minimizing images consist of a finite number of isolated weighted delta functions. The results naturally give rise to a host of speculations about the operation of the Visual System.
 

Filtered Directional Information and Psychometric Random Dot Kinematogram Data.

Andy Daniell and George Mather

Department of Experimental Psychology, University of Sussex, Brighton, UK.
Email: andrewd@biols.susx.ac.uk

Abstract
Feature statistical models of Random Dot Kinematograms (RDKs) use a single weighted statistic of nearest
neighbour feature distance, in a stationary frame, as an analogue of Dmax. (Morgan 1992, Eagle and Rogers 1996).

We extend this method by i/ using the whole distribution of expected nearest neighbour distances, and ii/ computing Directional Information (DI), the expected percentage of cross-frame matches in the correct direction, at any displacement size. Thus, DI provides a candidate model that can be tested against psychometric RDK data. The effects on DI of a pre-filter can also be computed relatively cheaply. We present comparisons between published data and filtered DI functions, using pre-filters of various types.

References
Eagle, RA and Rogers, BJ. (1996). Motion detection is limited by element density not spatial frequency. Vision Research, 36, 545-558.
Morgan, MJ. (1992). Spatial filtering precedes motion detection. Nature. 355, 344-346.

Centre-surround interactions: Modulatory effects in human visual cortex measured using fMRI.

A.L. Williams, A.T. Smith, K.D. Singh1
Royal Holloway, University of London, Egham, Surrey, TW20 0EX, UK.
1Neurosciences Research Institute, Aston University, Birmingham, UK.

The responses of visual neurons and the sensitivity of human observers to localised visual stimuli are both profoundly influenced by the visual context. We provide fMRI evidence of contextual modulation in the human
brain.  Subjects viewed a circular grating patch which was continuously present. A surround grating was added in an on/off block design to reveal its effect on the central region. Activation was expressed as correlation-weighted variance and visualised on a flattened map of the gray matter.  Modulation was measured in a region of interest activated by the
target alone. The observed effects were predominantly suppressive, consistent with the suppressive effects found in single neurons and perception. Suppression was greatest when the surround and centre had the same orientation (random spatial phase) and was reduced or absent when it was orthogonal. When spatial phase was controlled, suppression was greatest
for in-phase stimuli and much reduced or reversed (facilitation) for opposite-phase stimuli. When the surround was restricted to two opposite quadrants of the annulus, similar results were obtained for collinear and flanking surround gratings. With eccentric stimulus presentation, suppression was reduced and facilitation was seen in some cases. The results confirm the existence of powerful and stimulus-specific contextual effects in human visual cortex.
 

Dopamine and the representation of the upper visual field: evidence from vertical bisection errors in unilateral Parkinson’s disease

E.A. Atkinson1, J.P. Harris1, A.C. Lee1,3, K. Nithi 2,4, and M.S. Fowler1

1Department of Psychology, University of Reading, Whiteknights, Reading RG6 6AL, UK
e-mail: j.p.harris@reading.ac.uk; tel: : +44 (0) 118 9318522; fax: +44 (0) 118 9316715
2Department of Neurology, Battle Hospital, Oxford Road, Reading RG30 1AG, UK
3Present address:  Department of Psychology, Bath Spa University College, Newton Park, Newton St. Loe, Bath BA2 9BN, UK
4 Present address: Department of Neurology, Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, UK

It has been suggested that dopamine is an important neurotransmitter in the brain mechanisms which represent the upper visual field. This idea was tested with a vertical line bisection task in unilateral Parkinson’s disease. Stimuli of a range of lengths were presented on a large screen in 3 positions (left, centre and right) and at two viewing distances (0.6m and 1.5m).  The patients, who were compared with a group of normal age-matched controls, comprised 20 sufferers from predominantly unilateral disease, 10 with more severe left-sided symptoms (LPD) and 10 with more severe right-sided symptoms (RPD).  The LPD group consistently set the bisecting cursor below the midpoint of the stimulus lines, and their bisection error became larger as the length of the line increased.  In contrast, the controls set the cursor above the midpoint of the line, an error which also increased with line length.  The settings of the RPD group lay between those of the other groups. The results suggest altitudinal neglect in left unilateral PD, and support the hypothesis of dopaminergic involvement in the coding of upper visual space, with the proviso that the perceptual component of this involves the right hemisphere in humans.
 

Evidence for spatio-temporal selectivity in attentional modulation of the motion aftereffect

M.S. Georgiades, J.P. Harris

Department of Psychology, University of Reading, Whiteknights, Reading RG6 6AL, UK
e-mail: j.p.harris@reading.ac.uk; tel: : +44 (0) 118 9318522; fax: +44 (0) 118 9316715

An ignored region of the visual field might be monitored by an intermittent full visual analysis or by a more continuous but restricted analysis. We investigated which type of process is more likely in early vision by studying the effects of diverting attention on adaptation to a range of spatial (0.5, 2, 4, and 6 c/deg) and temporal (1.5 and 10 Hz) frequencies. During adaptation, subjects either fixated an unchanging digit (normal attention), or named the sequence of changing digits which formed the fixation point (diverted). The test field was always a static version of the adapting field, and the strength of adaptation was measured through the duration of subsequent Motion Aftereffects (MAEs). When attention during adaptation was normal, MAE durations rose with spatial frequency for the 1.5 Hz stimuli, and declined with spatial frequency for the 10 Hz stimuli. When attention was diverted from the 10 Hz stimuli, MAE durations fell by a similar amount at all spatial frequencies. However, for the 1.5 Hz stimuli, the effects of diversion were very small at 0.5 c/deg, and rose progressively with spatial frequency, so that MAE reductions were largest at 6 c/deg. It appears that diversion hardly affects the encoding
of coarse, slow, stimuli, but attenuates the encoding of finer/and or faster stimuli. This is consistent with the idea that during diversion the visual system monitors the scene continuously, but over a restricted range of spatial and temporal scales.
 
 
 

Multisensory Interactions in Saccade Curvature

Robin Walker & Melanie Doyle.

Royal Holloway, University of London, Egham, Surrey, TW20 0EX, UK.

In previous studies of saccade curvature the influence of task-irrelevant, and multisensory, distractors on trajectories have been relatively ignored.  We examined the impact of irrelevant (experiment 1) and informative (experiment 2) visual, auditory and tactile distractors on the trajectories of saccades to visual targets.  Observers made reflexive and voluntary saccades to visual targets above and below fixation in the presence of visual, auditory and tactile distractors which were located to the left and right of fixation.  Target and distractor stimuli were provided by LEDs (visual), Piezo sounders (auditory) and linear actuators (tactile) attached to a vertical stimulus framework: the fixation stimulus (a cross within a diamond) was presented on a monitor directly behind the framework.  In experiment 1 target location was indicated by target onset (reflexive) or a central arrow cue at fixation (voluntary): distractor onset was simultaneous with fixation point change or target onset.  In experiment 2 the target for voluntary saccades was determined by the distractor location (e.g., if the distractor was on the right observers had to saccade to the upper target and vice versa).  In control trials observers made reflexive and voluntary saccades to the target alone.  We observed that irrelevant distractors could influence saccade trajectories.  Further, both voluntary and reflexive saccades showed evidence of curved trajectories but this effect was modified by distractor modality and relative onset of target and distractor stimuli.  (Funded by Wellcome Trust, UK.)
 

The role of binocular information in the planning and control of prehensile movements in middle childhood.

Kathleen M. Elliott, Mark F. Bradshaw, Simon J. Watt, Tanya Holden and Tricia M. Riddell1.

Department of Psychology, University of Surrey, Guildford GU2 7XH, UK
1Department of Psychology, University of Reading, Whiteknights, Reading RG6 6AL, UK

Binocular information is considered to be of primary importance for the efficient execution of prehensile movements and its removal has been shown to have significant effects on the kinematics of adult prehension (Servos et al, 1992, Jackson et al, 1997, Watt & Bradshaw, 2000).  The present study examined the role of binocular information in the control of natural prehensile movements in middle childhood. Children between the ages of 5 and 11 reached for and picked up objects of different sizes at different distances under a variety of viewing conditions.  This age group is of particular interest as
pointing tasks have shown that reliance on on-line feedback for visuo-motor control appears to change during this period in a non-monotonic fashion (Hay et al, 1991).  Experiment one compared the kinematics of reaches made under binocular and monocular viewing conditions.  In contrast to adult data no significant effects of removing binocular information were found.  Experiment two examined separately the role of visual information in the initial view and on-line control stages of the movements.  In the initial view comparison subjects were presented with either a binocular or monocular view prior to movement onset.  To examine the role of visual feedback subjects reached under either binocular, monocular or open-loop conditions following a binocular initial view.  No effects of initial view were found in any of the age groups.  Type of feedback had no significant effect on reaches made by 5 to 9 year old children.  Reach kinematics of 10 to 11 year old children were significantly affected by the complete removal of visual feedback. However, no advantage of binocular feedback was found, in comparison to monocular feedback.  Overall the children exhibited normal scaling of peak velocity and peak grip aperture for object distance and size in all viewing conditions and no evidence for a non-monotonic trend
was found.  In contrast to adult data it appears that children in middle childhood do not rely on binocular information for the
pre-planning and control of their prehensile movements.

Endogenous shifts of covert attention operate within multiple coordinate frames: evidence from a feature-priming task.

D. J. K. Barrett, M. Bradshaw, D. Rose.

Department of Psychology, University of Surrey, Guildford GU2 7XH, UK

There is strong evidence that the location of visual objects and events is represented in a number of different coordinate frames.  Spatial neglect, for example, has been shown to be manifest for objects relative to a body-centred (egocentric) coordinate such as ‘to the left of fixation’ or relative to an external landmark (allocentric) in the visual scene (i.e. Behrmann, 1999; Robinson, 1999).  Previous research has shown that facilitation in response to a visual transient (exogenous prime) is centred upon the egocentric coordinate of the priming event (Cohen, 1981, Barrett et al, in press).  In contrast, endogenous shifts of attention have been characterised as operating at a level of representation in which locations are encoded independently of their egocentric position in an allocentric coordinate frame (i.e. Maljkovic & Nakayama, 1996; Vecera & Farah, 1994).  In two experiments we investigate the relative contributions of egocentric and allocentric coordinate frames to facilitation in a feature-priming task.  Primes and targets were presented in four conditions designed to vary systematically their spatial relationship in egocentric and allocentric coordinates.  For discrimination tasks involving judgements of changes in the geometric properties of the primed targets, a significant effect of both egocentric and allocentric priming was found.  In contrast, discrimination tasks not involving the geometric properties of the primed target produced a significant effect of egocentric priming only. The results suggest that priming in response to endogenous primes occurs in multiple coordinate frames and that task demands can influence the relative contributions of egocentric and allocentric coordinates to target facilitation.
 

Attention affects depth as well as direction aftereffects.

D. Rose, M. F. Bradshaw, and P. B. Hibbard

Department of Psychology, University of Surrey, Guildford GU2 7XH, UK

The perceptions of depth from binocular disparity and motion parallax are typically considered to be low-level pre-attentive processes. In the motion domain, however, such a view was rejected by Chaudhuri (Nature, 344, 60-62, 1990) who showed that the motion aftereffect can be modulated by a secondary attentional task.

In the present experiment we investigate whether the duration of a depth aftereffect (Blakemore and Julesz, Science, 171, 286-288, 1971) is also affected by attention to a secondary task (character processing at parametrically varied rates). We also compare the relative effects of the secondary task on disparity and motion aftereffects. Subjects adapted to stationary and/or moving random-dot patterns forming one or more depth planes and estimated aftereffect duration on stationary test patterns.

We found that the depth aftereffect was affected by attetional manipulations. The motion aftereffect was affected to a similar
extent. Both  aftereffect durations varied monotonically with the difficulty of the secondary task. The results are discussed in the light of recent theories of visual attention.
 

A quantitative method for assessment of clock drawings by right hemisphere (CVA) patients

Paul Roden

Brunel University

Clock drawing is a traditional assessment in neurological settings for visuospatial neglect.  Assessment of clock drawing is usually by qualitiative methods.  Therefore, a quantitative approach for assessing clock drawing was devised, based on the method of Mijovic (1991).  Average deviation was computed by comparing the drawn angle to the actual angle of a circular clock, in order to test whether left side distortion was greater than right side distortion.  Right hemisphere lesions may produce symptoms of spatial distortion or bias in contralesional visual fields, thus it was predicted that left-side distortion would be greater than right side distortion.  While this was true for 41 R hemisphere patients compared to 30 age-matched controls, there was no relationship between left -side distortion in clock drawing and standard measures of spatial bias in line bisection and cancellation tasks. The absence of a meaningful correlation between assessments of neglect and deviation in clock drawing suggests that clock drawing draws on different cognitive skills than those usually impaired by right hemisphere lesions.

Neural Correlates of Letter-String Length and Lexicality Effects in visual word recogniton: A MEG Study

T. Wydell*, T. Vuorinen**, P. Helenius**, & R. Salmelin**

*Department of Human Sciences, Brunel University, Middlesex, UK
**Brain Research Unit, Low Temperature Laboratory, Helsinki University of Technology, Espoo, Finland

Cortical activation was recorded using MEG during a reading task in a regular orthography, Finnish, with long and short word/non-word stimuli. Length effects, a signature of sequential processing, were identified in two distinct cortical regions: (1) in the bilateral occipital lobes at 100ms as differences in strength of activation regardless of the lexical status of the stimuli, and (2) the left superior temporal cortex between 200 and 600ms as differences in the duration of activation.

The duration of long letter-strings was twice that of short letter-strings for both words and nonwords, showing a striking linearity. Overall the duration of nonwords was greater than that of words, thus showing a lexicality effect.

The early length effect was interpreted as low-level visual analysis common to all visual stimuli, while the late length effect was interpreted as on-line phonological processing. The late length effect, however, was also modulated by lexicality,  revealing lexical influence.

Variability in the Response to a Naturalistic Target in Adults and Infants

P.M. Riddell, A.M. Horwood, J.E. Turner, S.M. Houston

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

The visual targets for young infants differ considerably in their accommodative demands from the text or high contrast gratings used with older subjects. This study compares adult accommodation to both text and a naturalistic target (a 12x8 cm coloured clown picture) and adult to infant performance on this more naturalistic target.
   A remote haploscopic photorefractor was used to study accommodation and convergence binocularly and monocularly in 25 children and adults (9 – 44 years) and in 134 healthy infants (1-54 weeks). Targets were placed at 5 distances ranging from 0.25 to 2 m representing target demands of 4 - 0.5 D. Adults and children were tested with text and/or clown targets. Infants were tested only with the clown.
   For the text target, accommodation for unoccluded adults followed target demand with a small lag for targets at 4D. This did not significantly reduce on occlusion (p=0.65), although 8% of the sample ceased to accommodate. For the clown target, adults’ unoccluded accommodation was lower than for text, especially at higher target demands. This significantly reduced on occlusion (p=0.02) with 25% showing flat responses. By 3 months of age, infant responses for the clown target compare well with those of the adults, showing a similar, significant reduction in accommodative response on occlusion, with 30% showing flat responses.
  This study demonstrates that by 3 months of age infants’ accommodative performance is equivalent to that of uninstructed adults when using naturalistic targets. This is true for both the unoccluded and occluded conditions. In addition, the range of individual behaviours found in adults and infants is similar.

Supported by MRC grant GR9608874N
 

Role of the dorsal visual stream in subitization

Louise Alston1 & Glyn W. Humphreys

The School of Psychology,  University of Birmingham, Edgbaston, Birmingham B15 2TT
1Present address: Department of Human Sciences, Brunel University, Uxbridge, UB8 3PH, UK
 

The fast and accurate counting of only a small numbers of visually presented items during subitization has been explained in terms of both small numbers having familiar patterns and as a result of spatial attention being limited to only a few items. These two explanations differ in their predictions about the visual areas involved, with pattern-recognition relying more on the ventral stream and spatial attention on the dorsal stream. This study examines whether subitization of randomly configured items is carried out predominantly by the dorsal visual stream by comparing performance at enumerating items presented in the upper and lower visual field. There is evidence that the parietal system has better representation of the lower visual field (Previc, 1990) and this has been related to better lower visual field performance in attention-based tasks (He, Cavanagh and Intriligator, 1996). Our results did not show a lower visual field advantage when randomly configured static or moving items were presented alone, or when the items formed recognisable patterns. The visual field effect was confined to the condition in which the visual system was “stretched” by the inclusion of distracters. It is concluded that the dorsal and ventral routes operate in parallel, allowing fast performance via the ventral route when the stimuli form recognisable patterns and via the dorsal route for unrecognisable configurations of small numbers of items. Subitization via the dorsal route is consistent with the subitizing limit arising from a limited capacity of visual attention.
 

The Importance of Foveal Vision in Reading: A Case Study

N.J. Upton1, T.L. Hodgson2, G.T. Plant3, R.J.S. Wise4 and A.P. Leff4

1 Brunel University, Uxbridge, Middlesex
2 Imperial College School of Medicine, Division of Neuro Science and Psychological Medicine, Charing Cross Hospital, London
3 Department of Neuro-opthalmology, Moorfield's Eye Hospital
4 MRC Clinical Sciences Centre, Cyclotron Unit, Hammersmith Hospital

Abstract

We report the case of MN, who, due to a slow growing brain tumour, shows an unusual pattern of visual field defects.  MN was found to read significantly faster with his left eye than his right when reading single words and text.  A linear WLE was shown in the presence of a macular splitting hemianopia, but not a macular sparing hemianopia, with no evidence of pure alexia.  Infrared pupil tracking showed MN’s right eye to be impaired in its guidance due to the lack of foveal vision.  The results highlight the importance of foveal and parafoveal vision in reading.