Ø PABELLÓN POLIDEPORTIVO - Cubiertas - techos:

EYE MOVEMENTS IN VISUAL PERCEPTION

^ A cholinergic mechanism underlies persistent neural activity necessary for eye fixation J M Delgado-Garc|¨a (Division of Neurosciences, University Pablo de Olavide, Carretera de Utrera, km.1, S 41013 Seville, Spain; e-mail: [email protected])

Where and how does the brain produce the sustained neural activity necessary for eye positions of fixation? The generation of the neural activity responsible for eye fixation after spontaneous eye movements was studied in vitro and in vivo. Rat sagittal brainstem slices taking in the prepositus hypoglossi (PH) nucleus and the rostral paramedian pontine reticular formation (PPRF) were used for the intracellular recording of PH neurons and their synaptic activation from PPRF neurons. Single electrical pulses applied to the PPRF showed a monosynaptic projection on PH neurons. This synapse was found to be glutamatergic in nature, acting on AMPA-kainate receptors. Train stimulation (100 ms, 50 ^ 200 Hz) of the PPRF area evoked a sustained depolarisation of PH neurons exceeding (by hundreds of ms) stimulus duration. Both duration and amplitude of this sustained depolarisation were linearly related to train frequency. The train-evoked sustained depolarisation was demonstrated to be the result of the additional activation of cholinergic fibres projecting onto PH neurons, because it was prevented by slice superfusion with atropine and pirenzepine (two cholinergic antagonists), and mimicked by carbachol (a cholinergic agonist). Carbachol also evoked a depression of glutamate release by a presynaptic action on PPRF neuron terminals on PH neurons. As expected, microinjections of pirenzepine in the PH nucleus of alert behaving cats evoked an ipsilateral gaze-holding deficit consisting of an exponential-like, centri- petal eye movement following each saccade directed toward the injected side. These findings strongly suggest that persistent activity characteristic of PH neurons carrying eye-position signals is the result of the combined action of eye-velocity signals originated in PPRF neurons and the facilitative role of cholinergic terminals of reticular origin.

[Supported by grant BFI2000-00936 from Spanish MECD.]

^ The physiology and psychophysics of visual search in monkeys free to move their eyes M E Goldberg, A L Gee, A Ipata, J W Bisley (Center for Neurobiology and Behavior, Columbia University, 1051 Riverside Drive, Unit 87, New York, NY 10032, USA; e-mail: [email protected])

Most studies of eye movements in awake, behaving monkeys demand that the animal make specific eye movements. We have developed a new paradigm in which the monkey performs a visual search for an upright or inverted T among 7, 11, or 15 cross distractors, and reports the orientation of the distractor with a hand movement. The search array is radially symmetric around a fixation point, but once the array appears, the monkey is free to move its eyes. The monkey's performance in this task resembles that of humans in similar tasks (Treisman and Gelade, 1980 Cognitive Psychology 12 97 ^ 136): manual reaction time shows a set size effect for difficult searches (the crosses resemble the Ts) but not for easy searches (the T pops out). Saccades are made almost exclusively to objects in the array, and not to intermediate positions, but fewer than half of the initial saccades are made to the T. We recorded from neurons in the lateral intraparietal area (LIP) while the monkey performed the search. LIP neurons distinguish the saccade goal at an average of 86 ms after the appearance of the array. The time at which neurons distinguish saccade direction correlates with the monkey's saccadic reaction time, suggesting that most of the jitter in reaction time for free eye movements comes from the discrimination process reflected in LIP. They distinguish the T from a distractor on an average of 111 ms after the appearance of the array, suggesting that LIP has access to cognitive information about the target, independent of the saccade choice. These data show that LIP has access to three different signals: an undifferen- tiated visual signal reporting light in the RF; a cognitive visual signal; and a saccade-related signal. [Supported by the National Eye Institute, Whitehall Foundation, W M Keck Foundation, James S MacDonnell Foundation.]

^ Statistics of fixational eye movements and oculomotor control

R Engbert, K Mergenthalerô (Computational Neuroscience, Department of Psychology, University of Potsdam, POB 601553, D 14415 Potsdam, Germany; ô Promotionskolleg ``Computational Neuroscience of Behavioral and Cognitive Dynamics'', University of Potsdam, POB 601553, D 14415 Potsdam, Germany; e-mail: [email protected]) During visual fixation, our eyes perform miniature eye movements involuntarily and uncon- sciously. Using a random-walk analysis, we found a transition from persistent to anti-persistent

34 Symposia: Eye movements in visual perception Thursday

correlations as a function of the time scale considered (Engbert and Kliegl, 2004 Psychological Science 15 431 ^ 436). This finding suggests functional dissociations (i) of the role of fixational eye movements on short and long time scales, and (ii) between drift and microsaccades. Here we propose a mathematical model for the control of fixational eye movement based on the concept of time-delayed random-walks (Ohira and Milton, 1995 Physical Review E 52 3277 ^ 3280). On the basis of results obtained from numerical simulations we estimate time delays within the brainstem circuitry underlying the control of fixational eye movements and microsaccades.

[Supported by Deutsche Forschungsgemeinschaft (DFG, grant 955/3) and Promotionskolleg ``Computational Neuroscience of Behavioral and Cognitive Dynamics'' (University of Potsdam).] ^ Fixational eye movements and motion perception

I Murakami (Human and Information Science Laboratory, NTT Communication Science Laboratories, NTT Corporation, Department of Life Sciences, University of Tokyo, 3-8-1 Komaba, Meguro-ku, Tokyo 153-8902, Japan; e-mail: [email protected]) Small eye movements maintain visibility of static objects, and at the same time randomly oscil- late their retinal images. The visual system must compensate for such motions to yield a stable visual world. According to the theory of visual stabilisation based on retinal motion signals, objects are perceived to move only if their retinal images make spatially differential motions with respect to some baseline movement, probably due to eye movements. Several kinds of motion illusions favouring this theory are demonstrated: with noise adaptation or flicker stimulation, which is considered to decrease motion sensitivity in restricted regions, image motions due to eye movements are actually perceived as random `jitter'. This indicates that the same amplitudes of such image motions are normally invisible as all regions are sensitive enough to register uniform motions as being uniform. As such, image oscillations originating in fixational eye movements may go unnoticed perceptually, but this does not mean that the oscillations have been filtered out from the brain; they can still exist in early motion processing and can influence various aspects of motion-detection performance. Lower threshold of uniform motion, for example, has been found to correlate positively with fixation instability, indicating that image oscillations are, though unnoticed, working as a limiting factor of motion detection. Also, the compelling motion illusion that appears in static figures (Kitaoka and Ashida, 2002 Perception 31 Supplement, 162), sometimes referred to as the `rotating snake', also positively correlates with fixation instability, such that poorer fixation makes the illusion greater. As a possible account, an interaction between image oscillations due to small eye movements and a low-level motion detection circuit is argued. Finally, the dependence of motion detection on the occurrence of fixational eye move- ments is analysed in finer detail, with some mention of separate effects of the two subtypes: fixational saccades and fixational drifts.

FROM PERCEPTIVE FIELDS TO GESTALT. IN HONOUR OF LOTHAR SPILLMANN ^ In honour of Lothar Spillmannöfilling in, emptying out, adaptation, and aftereffects

S Anstis (Department of Psychology, University of California at San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0109, USA; e-mail: [email protected])

During prolonged strict fixation, visual properties appear to become weaker over time. Colours become less saturated, contrast is reduced, and (we find) wiggly lines appear to become straighter, and an irregular lattice of dots appears to become gradually more regular. Also, a peripherally viewed gray patch on a red surround, or embedded in twinkling dynamic noise, seems to dis- appear from view after some seconds. When the adapting field is replaced by a uniform gray test field, the patch now appears to be filled-in with the red colour, or with the dynamic texture, of the surround. The short-term visual plasticity produced by this tangled mass of adaptation, aftereffects, spatial and temporal induction, and filling-in is examined. Is the filling-in during the aftereffect analogous to filling-in of the natural blind spot? Lothar Spillmann has cast much light on these topics during his highly productive career, but I shall restore the status quo.

[Supported by a UCSD Senate Grant.]

^ Lightness, filling-in, and the fundamental role of context

M A Paradiso (Department of Neuroscience, Brown University, 192 Thayer Street, Providence, RI 02912, USA; e-mail: [email protected])

Visual perception is defined by the unique spatial interactions that distinguish it from the point-to- point precision of a photometer. The perceptual properties of spatial interactions and more generally the importance of visual context for neuronal responses and perception have been explored. Investigations into the spatiotemporal dynamics of lightness provide insight into under- lying mechanisms. For example, backward-masking and luminance-modulation experiments suggest that the representation of a uniformly luminous object develops first at the borders and the centre

Symposia: From perceptive fields to Gestalt. In honour of Lothar Spillmann 35 Thursday

fills-in. The temporal dynamics of lightness induction are also consistent with a filling-in process. There is a slow cutoff temporal frequency above which surround luminance modulation will not elicit perceptual induction of a central area. The larger the central area, the lower the cutoff frequency for induction, perhaps indicating that an edge-based process requires more time to `complete' the larger area. In recordings from primary visual cortex, it is found that neurons respond in a manner surprisingly consistent with lightness perception and the spatial and temporal proper- ties of induction. For example, the activity of V1 neurons can be modulated by light outside the receptive field and, as the modulation rate is increased, response modulation falls off more rapidly for large uniform areas than smaller areas. The conclusion drawn from these experiments is that lightness appears to be computed slowly based on edge and context information. A possible role for the spatial interactions is lightness constancy which is thought to depend on extensive spatial integration. V1 neurons are not only strongly context-dependent, but this dependence makes V1 lightness constant on average. The dependence of constancy on surround interactions underscores the fundamental role that context plays in perception. In more recent studies, further support has been found for the importance of context in experiments using natural-scene stimuli.

[Supported by the USA National Eye Institute and the National Science Foundation.] ^ Beyond a relay nucleus: New views on the human LGN

S Kastner (Department of Psychology, Princeton University, Princeton, NJ 08544, USA; e-mail: [email protected])

The LGN is the thalamic station in the projection of the visual pathway from retina to visual cortex and has been traditionally viewed as a gateway for sensory information. Its topographic organisation and neuronal response properties have been extensively studied in nonhuman pri- mates, but are poorly understood in humans. I report on a series of studies aimed at elucidating functional roles of the human LGN in perception and cognition using fMRI. Functional LGN topography was studied by presenting periodic flickering checkerboard stimuli that evoked a traveling wave of activity. The contralateral visual hemifield was found to be represented with the lower field in the medial-superior portion and the upper field in the lateral inferior portion of each LGN. The fovea was represented in posterior and superior portions, with increasing eccen- tricities represented more anteriorly. This topography is strikingly similar to that of the macaque. Selective attention has been shown to modulate neural activity in both extrastriate and striate cortex. The poorly understood role of earlier, subcortical structures in attentional processing was studied. Attention was found to modulate neural responses in the LGN in several ways: it enhanced neural responses to attended stimuli, attenuated responses to ignored stimuli and increased baseline activity in the absence of visual stimulation, suggesting a role as a gatekeeper in controlling attentional response gain. Most recently, the level at which competing inputs to the eyes, as perceived in binocular rivalry, can be resolved started to be investigated. Similar neural correlates of binocular rivalry were found in the LGN and V1, suggesting a mechanism by which LGN layers that process the input from one particular eye are selectively enhanced or suppressed.

^ From perceptive fields to Gestalt

L Spillmann (Brain Research Unit, University of Freiburg, Hansastrasse 9a, D 79104 Freiburg, Germany; e-mail: [email protected])

I discuss select studies on visual psychophysics and perception that were done in the Freiburg laboratories during the last 35 years. Many of these were inspired by single-cell neurophysiology in the cat. The aim was to correlate the phenomena and effects under consideration to the possibly underlying mechanisms from retina to cortex. To this extent, I deal with light and dark adapta- tion (photochromatic interval, rod monochromacy, Ganzfeld), colour vision (spectral sensitivity, latency differences, colour assimilation), perceptive field organisation (Hermann grid illusion, Westheimer paradigm, tilt effect), visual illusions (Ehrenstein illusion, neon colour, abutting grating illusion), and long-range interaction (phi-motion, factor of common fate, fading and filling-in). While some of these studies succeeded in linking perception to neuronal behaviour, others did not. The task of probing the human brain by using phenomena in search of mechanisms continues to be a challenge for the future.

[Supported by DFG grant SP 67/9-3/.]

36 Symposia: From perceptive fields to Gestalt. In honour of Lothar Spillmann Thursday

ORAL PRESENTATIONS

THEORY AND MODELS

^ A model of velocity aftereffects: Two temporal filters and four free parameters

S T Hammett, P G Thompsonô, R A Champion, A B Morland (Department of Psychology, Royal Holloway University of London, Egham TW20 0EX, UK; ô Department of Psychology, University of York, York YO10 5DD, UK; e-mail: [email protected])

The perceived speed of moving patterns changes over time. Adapting to a moving pattern leads to an exponential decrease in its perceived speed. However, under certain conditions, perceived speed increases after adaptation. The time course of these perceptual effects varies widely. We measured the perceived speed of 1 cycle degÿ1 _{sinusoidal patterns over a range of adaptation}

and test speeds (2 ^ 20 deg sÿ1_{) and at a variety of adaptation durations (0 ^ 64 s). The results}

indicate that adapting to slow speeds results in an increase in the perceived speed of faster images and a reduction in the perceived speed of images of the same or slower speeds. Adapting to high speeds led to an exponential reduction in the perceived speed of all subsequently presented images. Thus, any model of perceived speed must capture both increases and decreases in perceived speed contingent upon prevailing conditions. We have developed a model that comprises two temporally tuned filters whose sensitivities reduce exponentially as a function of time. Perceived speed is taken as the ratio of these filters' outputs. The model has four free parameters that determine the time constants of exponential decay and the asymptotic response attenuation for the temporal filters. The model assumes that the decay of each filter's sensitivity over time is proportional to the relative sensitivity of that filter to the adaptation frequency. The model captures both increases and decreases in perceived speed following adaptation, and describes our psychophysical data well, resolving around 96% of the variance. Moreover, the parameter estimates for the time constants of the underlying filters ( 8 s) are very close to physiological estimates of the time course of adaptation of direction-selective neurons in the mammalian visual system. We conclude that a physiologically plausible ratio model captures much of what is known of speed adaptation.

^ A neurocomputational model for describing and understanding the temporal dynamics of perisaccadic visual perception

F H Hamker, M Zirnsak, D Calow, M Lappe (Department of Psychology, Westfa«lische-Wilhelms-Universita«t, Fliednerstrasse 21, D 48149 Mu«nster, Germany; e-mail: [email protected])

Several experiments have shown that the plan of making an eye movement affects visual processing. Under certain conditions, briefly flashed stimuli are mislocalised towards the saccade target (Ross et al, 1997 Nature 386 598 ^ 601). This effect starts before the eyes move and is strongest around saccade onset. The spatial pattern of mislocalisation is asymmetric in space and depends on stimulus position (Kaiser and Lappe, 2004 Neuron 41 293 ^ 300). In V4, perisaccadic recep- tive field (RF) shifts have been reported (Tolias et al, 2004 Neuron 29 757 ^ 767), and several, primarily oculomotor related areas, show perisaccadic remapping (Kusunoki and Goldberg, 2003 Journal of Neurophysiology 89 1519 ^ 1527). However, neither the underlying RF processes, nor the phenomenon of the `compression' of visual space are well understood. We have developed a neurocomputational model of perisaccadic perception in which an oculomotor feedback signal is directed towards the saccade target and changes the gain of the cells in extrastriate visual areas. As a result, the cortical activity profile induced by a flashed dot is distorted towards the saccade target. The model can reproduce the temporal course and the 1-D spatial pattern of mislocalisation as measured by Morrone et al (1997 Journal of Neuroscience 17 7941 ^ 7953) and the 2-D mislocalisation data of Kaiser and Lappe (2004). It further inherently predicts RF dynamics. For the selected parameters, we observe a perisaccadic shrinkage and shift of RFs towards the saccade target as reported for V4. Our oculomotor-feedback hypothesis differs from remapping, since the RF shifts are directed towards the saccade target. Thus, we suggest a further universal mechanism that is likely to occur in intermediate areas within the visual hierarchy. The oculomotor-feedback hypothesis is the first integrative account for both electro- physiological measurements of receptive field shifts and for psychophysical observations of spatial compression.

Oral presentations: Theory and models 37

^ Brightness integration: Evidence for polarity-specific interactions between edge inducers T Vladusich, M P Lucassenô, F W Cornelissen (Laboratory of Experimental Ophthalmology and BCN Neuro-imaging Center, School of Behavioral and Cognitive Neurosciences, University Medical Centre Groningen, PO Box 30001, NL 9700 RB Groningen,The Netherlands; ô Department of Perception, Vision and Imaging Group, TNO Human Factors, NL 3769 ZG Soesterberg, The Netherlands; e-mail: [email protected])

We present a computational framework for analysing data on the spatial integration of surface brightness. Our framework builds on the hypothesis, originating in Retinex theory, that brightness is computed by integrating induction signals generated at edges (log luminance ratios) in a scene. The model of Rudd and Arrington (2001 Vision Research 41 3649 ^ 3662) generalises Retinex theory by characterising how neighbouring edges can interact to partially block the flow of induction signals from one another. We show that both the Rudd ^ Arrington model and Retinex theory are special cases of a broader class of models in which opposite-polarity edges are parsed into separate half-wave rectified channels before spatial integration. Each model incorporates different polarity-specific constraints on the interactions between neighbouring edges. We fit these models to psychophysical data on spatial brightness integration (Hong and Shevell, 2004 Visual Neuro- science 21 353 ^ 357; Vision Research 44 35 ^ 43), comparing performance using a statistical technique for quantifying goodness-of-fit relative to the number of model parameters. We find that a model which strongly impedes the flow of induction signals across neighbouring edges of the same polarity, but does not restrict, or weakly restricts, flow across edges of opposite polarity, is most likely to be correct. Our results are at odds with published variants of the filling-in theory of brightness perception, which predict either unrestricted flow across edges of the same polarity or no flow at all. The framework can also be used to quantitatively assess models of colour perception, where putative polarity-specific interactions can be defined in terms of cone-specific contrasts, as implied by Retinex theory, or cone-opponent contrasts.

[Supported by grant 051.02.080 of the Cognition program of the Netherlands Organization for Scientific Research (NWO).]

^ Can perception violate laws of physics?

R L Gregory (Department of Experimental Psychology, University of Bristol,