General Information
Registration / Login
Abstract Submission
Travel Information
Poster Sessions

 APCV 2010 Poster Sessions
Friday, July 23 / Saturday, July 24 / Sunday, July 25

Friday, July 23

Morning Poster Session: 11:00-12:00 (Odd numbers present)

Afternoon Poster Session: 15:40-16:30 (Even numbers present)

Color in Content

Please tick ▼ for abstract content

1‧Mikako Kuroki
How is the induced color determined in the watercolor configuration?

When a dark contour is flanked on the inside by a lighter chromatic contour, the lighter color will spread over the entire enclosed area. This spreading of the inner color is known as the watercolor effect. However, it has been also demonstrated that a color different from the inner color could spread with the same stimulus configuration; e.g., when the outer color is magenta and the inner color is red, yellowish color could spread under certain conditions. In this study, we aimed to clarify how the color induced in the watercolor configuration is determined. We examined the effects of relative luminance between the inner and outer contours in Experiment 1 and the influences of the outer color on the induced color in Experiment 2. Results showed that, when the luminance of the outer contour was higher, the induced color became different form the inner color and was affected by complementary color of the outer color. These results can be accounted for by assuming that the induced color in the watercolor configuration is determined by a mixture of the inner color and complementary color of the outer color and the strength of each color depends on relative luminance of two contours.

2‧Han Nim Cha
The effect of binocular disparity and phase transformation of dotted lines on Water Color Illusion

Two experiments were conducted to investigate the impact of collinearity of inside-dotted-line and outside-dotted-line and binocular disparity on Water Color Illusion(WCI). The stimuli was the modification of figure used by Pinna, Brelstaff and Spillmann(2001); the figure was four- diamond shape and was constructed of dotted line, the outside-dotted-line color was purple and the inside-dotted-line was yellow. In the experiment 1, the effect of collinearity was tested in the outside-dotted-line and inside-dotted-line. In the experiment 2, the impact of dotted lines that were put in different depth were tested. The result showed that illusion was persisted when collinearity of outside-dotted-line and inside-dotted line was changed. Despite change of collinearity, results was convinced WCI and color assimilation like Neon Color Spreading appeared through by different mechanism. And illusion was persisted when components of stimuli were placed on the different plane. This result was supported the study of Pinna, Brelstaff and Spillmann(2001). With this, could confirmed that the WCI was affected by high-levels information processing stages were after combination of binocular information.

3‧Na Ri Shin
The effect of color and interval and binocular disparity information on Water Color Illusion

Three experiments were performed to investigate the impact of colors, interval of zigzag lines and binocular disparity on Water Color Illusion(WCI). The stimuli was transformation of figure used by Pinna and Tanca(2008); figure was constructed of the bars. That was formed double-zigzag lines shape. Black bars were located outside in the large zigzag line and red bar was located inside in the small zigzag line then that color was reversed. In experiment 1, the impact of changed colors at the outside of small zigzag line was tested. Color was changed four conditions; red, green, yellow, gray. In experiment 2, the effect of interval between large zigzag line and small zigzag line was tested. In experiment 3, the effect of binocular disparity was tested. In this case inside bar of large zigzag line was red and outside bar of small zigzag line was green. The results showed that illusion was reduced when induced bar was different color. The illusion was cleared when the interval of lines was narrow. The illusion was weaken when condition of binocular disparity regardless of depth of stimuli. That was suggested that color was filled from the outside to the inside by the Feature Contour System.

4‧Hiroto Kimura
Detectability of color modulation on isoluminant apparent motion stimuli

A previous study has reported that detectability of luminance modulation on multiple targets was considerably decreased when they were perceived as a single moving object. In this study we measured discrimination sensitivity to color modulation on apparent motion stimuli to investigate the influence of apparent motion on color perception. The observer was instructed to discriminate alternating two colors of different discs presented sequentially at concyclic eight locations on a black background. Strength of apparent motion perception varied with SOA (Stimulus Onset Asynchrony) between disc presentations. In the results, the color discrimination sensitivity was lower in the SOA corresponding to the strongest motion perception than in the other SOAs. These results may be consequence of integration of disc colors along apparent motion trajectory. In addition, to investigate the effect of motion perception on color modulation detectability more clearly we conducted similar experiments using a colored background isoluminant with the discs, where motion perception was weaker than on the black background. We compare the results between black and isoluminant backgrounds, and discuss the relationship between motion perception and color modulation detectability.

5‧Masataka Sawayama
Common-fate grouping affects brightness perception on the articulated surround.

Brightness of a target depends upon the luminance of a surrounding field. Moreover, the effect of the surrounding luminance can be enhanced by articulating the surround (i.e., adding small patches of different luminances to the uniform field), even if the spatially-averaged luminance is kept constant. This study further investigated this articulation effect by asking whether spatial interaction found between the target and surrounding field is governed only by retinal proximity or also affected by spatial organization such as common-fate grouping. The target in the experiment was a moving patch located at the center of an articulated surround. Either small patches or uniform background in the surround was moved in the same speed, direction, and timing with the target. The target was easily grouped with the patches in the former case, whereas with the background in the latter case, although retinal proximity was kept nearly constant. Results showed clear influences of common-fate grouping on the articulation effect; the effect was strong when the target was grouped with the patches, but reduced when grouped with the background. These findings suggest that brightness on the articulated field could be computed even after retinal elements were spatially organized according to some grouping principles.

6‧Kazuya Inamoto
The effect of background color on color matching to skin color under sinusoidal luminance modulation

In this paper, we investigated the influence of background color on color appearance when a central stimulus was skin color under sinusoidal luminance modulation.The color appearance was measured by color matching method.The skin color and the amount of luminance were obtained as the average of human face images presented on a CRT monitor.We used four kinds of central stimuli.They were the skin color square under sinusoidal luminance modulation and the same skin color under constant luminance for control (the highest-,lowest- and averaged-luminance of the sinusoidal luminance modulation stimulus).One of the central stimuli was surrounded by one background color (gray,red,yellow,green,blue,or purple).In one session, all combinations of the central stimuli and background colors were presented and matched to a color patch with gray background.Four color-normal subjects participated in.The results indicated following three points.First, when the background color is gray, subjects can match the luminance to the averaged-luminance. Second, when the background color is red,yellow,green,blue or purple, the influence of the background colors on the matching under sinusoidal luminance modulation is less than that on the matching under the constant luminance.Third, when the background is not gray, the skin colors under sinusoidal luminance modulation are perceived in higher luminance than the averaged-luminance.

7‧ Sze-Wing Lee
Impact of red color on mood and task performance on Chinese in Hong Kong

Red color means differently in diverse cultural contexts, and our mood is moderated by the object or physical space that attaches to color. This study studies the impact of red color on Chinese people in Hong Kong in an object-based context.
32 Chinese participants listed first ten objects that came to their mind with color red, and their associated moods. Culture-specific items (e.g. red packets) never mentioned in previously studies were noted. We selected the most frequently mentioned positive and negative objects to conduct a word categorization task. In each trial, an object-word pair, either congruent or incongruent in mood, was printed in either red or black shown for 2sec. Participants judged whether the word contained positive or negative mood. We predicted observers would benefit from the task-irrelevant but red-associated object if it was presented in red and with congruent associated mood. No significant difference in response time was found in either printed color or mood condition. However, error rate analysis showed an advantage in mood congruent condition. Our results suggested that mood effect associated with color requires more sensitive measurement (e.g. error rate), and response time may not be a good choice for future similar studies.

8‧Chia-Ching Wu
The effect of the chromaticity of image elements on symmetry detection

We investigated the effect of chromatic content of image elements on symmetry detection threshold with a noise masking paradigm. In a temporal 2AFC trial, the random dot noise mask was presented in both intervals. The target was randomly presented in one interval while a random dot control was presented in the other. The task of an observer was to determine which interval contained the vertical symmetric target. The image elements varied both in chromaticity and luminance. In Experiment 1, the target density threshold was measured at various target-mask chromaticity and luminance combinations. In Experiment 2, the chromaticity varied among image elements in both the target and the mask. The number of color types (red, green, blue and yellow) ranged from 1 to 4 in a pattern.
The symmetry detection threshold was highest when the target and the mask were of the same hue and luminance. The threshold decreased as the difference in chromaticity and luminance between the target and the mask increased. This suggests that the symmetry detector is color turned. The symmetry detection threshold decreased with the number of colors. This suggests a probability summation among independent symmetry detectors in each color channel.

Retinal Functions

Please tick ▼ for abstract content

9‧Ya-Ting Yang
Retinal network responses upon subretinal electrical stimulation

The development of the efficient subretinal implants requires deeper understandings of how the retinal network responds to the electrical stimulation. In the present study, we characterized the responses of retinal ganglion cells (RGCs) upon subretinal electrical stimulation with a silicon-based microphotodiode array in the normal adult New Zealand White rabbits. Various pharmacological reagents were also used in conjunction with the electrical stimulation to dissect the upstream components of these RGC responses. In contrast to previous studies, we found that the RGC responses were initially increased as the stimulation strength increased, but then decreased when the stimulation exceeded an optimal range. During the repeated electrical stimulation, the RGC responses were significantly suppressed when the temporal frequency was above 10-20 Hz. Importantly, we noticed that the ON- and OFF-alpha RGCs showed different response thresholds and latencies, and sometimes recurring bursts of spikes after the initial responses, a suggestion of synaptic inhibition and excitation oscillation. All RGC responses upon electrical stimulation were abolished by CNQX, and picrotoxin can significantly decrease the spike latency and increase the firing rate. Furthermore, the responses of ON-alpha RGCs were inhibited by APB, an indication of direct photoreceptor activation. These results provide insights into the artificial retina design.

10‧Yukio Shimoda
Dopaminergic retinal nurons in various kinds of animals revealed by glyoxylic fluorescent technique

The glyoxylic acid fluorescence method was developed by Lindvall and Björklund (1974) to find monoamine neuron systems in the brain. We have applied this method on the retinas of various kinds of animals to investigate the morphology of dopaminergic cells. We could distinguish the dopaminergic cells from other monoaminergic cell, such as serotonin and glutamate cells by fluorescent color. We have also performed immunohistochemical studied using anti-tyrosine hydroxylase antibody and Alexa Fluor 546 to confirm the results of the glyoxylic acid fluorescence method. The results obtained by both methods are almost identical in every animal, fish, frog, chicken and mouse. The cell bodies of dopaminergic cell were seen in the inner nuclear layer and they extended many dendrites both in the outer- and inner plexiform layers. They have very wide dendritic field approximately 0.4mm in diameter. Distances between the neighboring cells are around 160μm. Therefore, density of the dopaminergic dells is roughly estimated as 40 cells/mm2. This value is very few as compared to other retinal neurons, for instance photoreceptors are about 10000 and the horizontal cells are 4000 cells/mm2 each. Intriguingly, these cell counts are almost identical among the animals which have been used in these experiments.

11‧ Makoto Kaneda
Two choline transport pathways in cholinergic amacrine cells in the mouse retina

Ion channels of P2X2-purinoceptors have a permeability to large cations. In the present study, we examined whether choline can permeate the ion channels of P2X2-purinoceptors in the cholinergic amacrine cells of the mouse retina by the patch clamp technique. An application of ATP activated choline current in the OFF-cholinergic amacrine cells. Concentration response curve of ATP-induced choline current was fitted with a Hill coefficient of 1.93 and EC50 of 46 microM. Reversal potential of ATP-induced choline current shifted 25 mV per 10 times change of extracellular choline concentration. Choline current was activated by an application of ATP-gamma-S but not by alpha,beta-methylene ATP or benzoyl, benzoyl ATP. In the presence of pyridoxalphosphate-6-azophenyl-2’,4’-sulfonic acid, ATP did not induce any choline current. These characteristics of ATP-induced choline current well corresponded to the characteristics of the current mediated by P2X2-purinoceptors. The immunoreactivity for choline transporter (high affinity) was stronger in the ON-cholinergic amacrine cells than in the OFF-cholinergic amacrine cells. Our data raise a possibility that choline transport mechanism in cholinergic amacrine cells is different between ON-and OFF-pathway. Ion channels coupled with P2X2-purinoceptors work as an additional pathway of choline transport especially in the OFF-cholinergic amacrine cells of the mouse retina.

12‧Wen-Hsuan Huang
Responses of the ganglion cells upon light and electrical stimulations in the rat retina

To achieve an efficient retinal prosthesis, it is important for the retinal ganglion cells (RGCs) to generate meaningful spike patterns upon electrical stimulation that are comparable to the normal light stimulation. While many studies have investigated the responses of RGCs upon electrical stimulation, fewer studies have compared the spiking activities upon light and electrical stimulations in the same cells. In the present study, we used normal Sprague-Dawley rats to characterize RGC responses upon both light and electrical stimulations with a silicon-based microphotodiode array. The rat retina was placed onto the chip with its photoreceptor side down, and the stimulating current was elicited by activating the solar cell with a 532 nm laser light source. The charge density dependent response and the frequency dependent pair-pulse suppression were examined. We found that the light intensity dependent responses of ON, OFF, and ON-OFF types of RGCs were similar to their charge density dependent responses. However, the frequency dependent pair-pulse suppression was more severe upon electrical stimulation than upon light stimulation. This suggests that RGCs may not achieve the same temporal precision in subretinal electrical stimulation as in its normal light stimulation. This study thus points out some limitations in developing subretinal prosthesis.

13‧ Hung-Ya Tu
Expression of Connexin 36 during postnatal development of the rabbit retina

Connexin 36 (Cx36) gap junction channels play an important role in the rod pathway by directly connecting AII amacrine cells to ON cone bipolar cells and other AII amacrine cells in the mammalian retina. The expression of Cx36 in cone photoreceptors and some ganglion cells also plays significant roles in retinal circuitry. Given the importance of Cx36 in the adult retinas, it is of great interest to characterize the temporal and spatial patterns of Cx36 expression in the developing retina, and to examine its functional significance regarding various developmental events.
The retinas from New Zealand White rabbits of different developmental stages were used for immunohistochemistry and Western blots. The antibody against Cx35/36 was used to examine the postnatal expression patterns of Cx36. We found that the Cx36 puncta were faintly labeled both in the IPL and the OPL after birth. The increase of Cx36 expression remained slowly during the early postnatal stage, and then became faster around P10, reaching the adult level after P22. A similar trend was observed from the quantification of Cx36 protein levels in Western blots. These results suggest that the formation of Cx36 gap junction is correlated with synaptogenesis around the second postnatal week.

14‧Hsing-Yen Huang
Arl6ip1 functions in differentiation, mitosis, and ER-stress mediated apoptosis during retinogenesis of zebrafish embryos

We found that arl6ip1 was a maternal expression gene which distributes around the whole embryo before 24-hpf, but after 48-hpf, arl6ip1 was exclusively expressed in brain, retina, heart and kidney. When Arl6ip1 translation was blocked by injecting arl6ip1-morphlino, many organs and tissues developed abnormally, and defective embryos could not survive beyond 5-dpf. Next, we focused on the retinal development to discuss the functional role(s) of Arl6ip1 that because small eyes were easily observed in all phenotypic defects. We observed the failure of cell differentiation, the reduction in cell number, bigger cells, and the imperfect progression of cell cycle in the defective retina. Moreover, mosaic analysis demonstrated Arl6ip1 is required for retinal differentiation in a non-cell autonomous manner by cell transplantation. In addition, we observed the swelling of ER, the swelling of Golgi apparatus, paled and engorged mitochondria, and bigger nucleus in the defective retina. ER-stress responded marker, chop, was highly expressed in the defective retina that indicating cells undergoing ER stress, which led, in turn, the cell death were observed at later development stage. These evidences suggested that Arl6ip1 secures cell-cycle progress and maintains the structural integrity of ER and mitochondria in the retinae.

Eyemovement & Gaze I

Please tick ▼ for abstract content

15‧ Kuan-Hui Li
The Development and the Characteristics of Oculomotor Inhibition In the Preschool Children

The development of executive function is very rapid in preschool children. One of vital ability in executive function is to prevent execution of a behavior in circumstances where to do so may be inappropriate. Only several studies have investigated the performance of oculomotor inhibition in preschool children (e.g. 4 to 6 yrs). However, the characteristics of the developmental curves in the preschoolers are still unclear. In this study, we used a well-established Pro-/Anti-saccade paradigm (Juan et al., 2004; 2008; Liu et al., 2009) to gauge the function of the oculomotor inhibition and the developmental curves in the preschoolers. The results indicated the performance of inhibitory had gradually but significantly improved across the stages of the preschool ages. The children mainly had difficulties when they were required to antisaccade into a position where there was no object. When an extra object was placed in that position, their performance of antisaccades improved. This pattern of results indicates that the competition between target and distractor may benefit the antisaccade latencies and accuracy. This also suggests preschoolers may develop more slowly in the spatial-based attention than in the object-based attention.

16‧Chang-Mao Chao
Predictability of saccadic behaviors is modified by transcranial magnetic stimulation over human posterior parietal cortex

Predictability in the visual environment provides a powerful cue for efficient processing of scenes and objects. Recently, studies have suggested that the directionality and magnitude of saccade curvature can be informative as to how the visual system processes predictive information. The present study aimed to investigate the role of the right posterior parietal cortex (rPPC) in shaping saccade curvatures in the context of predictive and non-predictive cues. We employed an orienting paradigm with the target predictability and delivered transcranial magnetic stimulation (TMS) over rPPC. Participants were presented with either an informative or un-informative cue to upcoming target locations. Our results showed that rPPC TMS generally increased saccade latency and error rates. Intriguingly, rPPC TMS increased curvature away from the distractor only when the target location was unpredictable and decreased saccadic errors attracted by the distractor but no effects in predictable target location. These results dissociate the strong contingency between saccade latency, curvature and also indicate that rPPC plays an important role in allocating and suppressing attention to distractors when the target demands visual disambiguation. Furthermore, the study suggests that rPPC is critically involved in determining saccade curvature and the generation of saccadic behaviors under conditions of differing target predictability.

17‧ Chia-Lun Liu
Frontal eye fields and the location probability effects: a repetitive Transcranial Magnetic Stimulation study.

Utilization of the regularity that imbedded in the environment allows the organism to preserve more cognitive resources for monitor other peril. Our recent study demonstrates that the location probability of a saccade modulates the magnitude of antisaccade cost (Liu et al., 2009). This highlights that the probabilistic information can override the attentional effect like the antisaccade cost. The neural substrates of the probability effect were first associated with the neural activities in the superior colliculus (SC). It is plausible that the modulation can be descended from higher oculomotor regions like frontal eye field (FEF) and supplementary eye field (SEF), given their immense reciprocal connections. We investigate this issue with an antisaccade task combined with the manipulation of the location probability. rTMS was applied over FEF and SEF, respectively, to probe the functional roles of these area in the involvement of location probability. The results revealed that FEF TMS prolonged the overall saccade reaction times, moreover, the interference is more pronounced in high probability locations. In contrast, none of the effects was observed found in the SEF TMS condition. This pattern of results confirms the critical role of FEF in initiating a saccade and in modulation of the location probability effects.

18‧Hui-Yan Chiau
The roles of frontal eye field and supplementary eye field in trial type probability – a transcranial direct current stimulation study

The effects of prior probabilistic information on our behaviors are influential. However, the characteristics of its dynamic interaction with our flexible control over behaviors are still unclear. The antisaccade task, where eye movements are made away from a target, has been used to investigate the flexibility of cognitive control of behavior. In this study, the trials for pro- and antisaccade were interleaved within one block and their relative probabilities were systematically manipulated across blocks. The saccade latencies were decreased with higher trial type probabilities. The transcranial direct current stimulation (tDCS) was applied to investigate functional roles of frontal eye field (FEF) and supplementary eye field (SEF). We found that anodal tDCS over SEF, the accuracy of saccadic responses declined when the likelihood of pro- and antisaccade was equal. Furthermore, the elongated saccade latencies were consistently observed for almost all conditions after anodal tDCS over FEF. The results suggested the supervisory role of SEF on saccades that may provide the signals to bias the preference for preparation of a saccade according to prior probabilistic information. These SEF signals may descend to area, which is directly responsible for saccade, such as FEF, and consequently speeded the response of trials with higher probability.

19‧Yasumasa Ogata
Influence of a visual target presented at detection sub-threshold on saccadic induction

Saccades are unconsciously induced by objects in the natural scene. It would be common that a saccade is triggered toward an object even if we are not aware of the object before the saccade occurs. In this study we examined whether a target could induce saccades when it was set at the sub-threshold level in detection. We used the metacontrast masking paradigm to make the stimulus being below detection threshold. The saccade latency was measured at various SOAs between the target and the mask onsets. A trial consisted of two sequences of the target-mask presentation. Either in the first or in the second sequences no target, but only the mask, was presented. The observer made a saccade toward any stimulus that first appeared. He also answered the sequence that contained the target in the 2AFC manner. The results showed that the saccade latency was shorter at a certain SOA than the total time (T) of the latency for the no-target (only-mask) presentation plus the SOA value. At this SOA the target detection was below threshold. This indicates that the saccade was not determined solely by the mask, but influenced by the target that was not detected.

20‧ Da-Lun Tang
How Balanced Configuration influence the visual scanpath and processing

Balance is considered as one of the most essential elements of aesthetics. However, few definite evidences were offered from objective or empirical aspects. In order to exclude the influence of confounding variables and to follow a strict
statistical process, this study tried to develop two kinds of objective indicators with mathematical parameters called “uniform index” and “progressing index” to re-verify the influence of balance configuration during aesthetic evaluation. In this study, the manipulation of balance configuration was conducted by two experts in visual art. Twenty well-balanced pictures and twenty ill-balanced pictures were used as experimental stimuli. The participants were randomly assigned to perform the memory test or the balance judgment task after viewing each picture in one
session. Eye movements were recorded using eye-tracking system when participants were performing the tasks. The following results were obtained: (1) The relationship between the balanced configuration and the uniform index
achieved significant level totally, no matter what task was performed. (2) The relationship between the balance configuration and the progressing index achieved significant level only in the balance judgment task. (3) The balanced configuration did affect participants' memory performance significantly.

21‧ Wang-Meng Hsun
Plausibility of a lie detection procedure based on eye tracking patterns

Previous studies of lie detection tests have been mostly based on liar’s emotional responses. They assumed that tension or stress induced by lying affects responses of the sympathetic nervous system. Increasing numbers of recent studies took interests in the visual cognitive processes involved in lying that sheds light on the design of practical lie detection procedures. Current study tested a hypothesis derived from the relationship between working memory and an eye movement behavioral model. Participants were randomly assigned to lying or truth telling conditions. They drew a number between one and six and then answered number-relevant and control questions. The target number that they drew earlier was displayed along with three distracter numerals briefly after presenting the question. Eye tracking pattern showed that participants in the lying condition fixated less than in the truth telling condition if the question was number relevant. For control questions, lying participants fixated more on the target number than truth telling participants. These findings suggested that liars may deliberately avoid fixating the to-be-lied target. The implication of these eye tracking patterns for lie detection tests was also discussed.

22‧ Hsueh-Cheng Wang
The role of semantic transparency in the processing of two-character Chinese words

The study proposes a computational method to provide quantitative measures of semantic transparency of Chinese words and English compound words using Latent Semantic Analysis, and we found that LSA can distinguish transparent and opaque constituents. In an eye-movement experiment, two-character words with opaque-transparent (OT), transparent-opaque (TO), high-frequent opaque-opaque (OO1), and low-frequent opaque-opaque (OO2) combinations were paired with matched transparent-transparent (TT) words, and each target word was embedded in an individual sentence. Different eye movement measures (first fixation duration, gaze duration, and total time) were examined for both whole-word and individual-character level. The results support dual-route model that the meaning of two-character Chinese words is mainly from whole word, but character-level processing was involved during Chinese reading.

23‧Masae Yokota
The relation between perceptual filling-in facilitation and eye movement

When a small area is presented in peripheral vision, it becomes invisible and invaded by surrounding texture within a few seconds. This visual illusion is called perceptual filling-in. In our preliminary study (Yokota, eMBEC 2005), we found that incomplete fixation distributes filling-in time. Furthermore, that we can see nothing by restraining eye movement artificially is well known. Therefore, we can consider that filling-in time is influenced by eye movement. Although it has been recently reported that eye movement influences the filling-in occurrence (Martinez-Conde, Neuron 2006), the relation between eye movement and the filling-in time has rarely been reported.
For this study, we measured the filling-in time for three subjects, for four surrounding textures, with simultaneous recording of eye movement. The results show that the filling-in time correlates the standard deviation of the power of the eye distance from the fixation point. Furthermore, we found relatively strong correlation between the filling-in time and the power of high frequency component 50-200 (Hz) in the eye movement, though the correlation of the power of low frequency component 10-50 (Hz) is not so high. Thus we suppose that filling-in is inhibited by small involuntary eye movement.

Attention I

Please tick ▼ for abstract content

24‧Jinglin Li
Expectation overcomes the impairments induced by a large task-irrelevant salient line in visual search

Our previous study found that a task-irrelevant large salient line impaired visual search (Jingling, 2010). In this study, we further investigated how expectation modulates this impairment. The search display was a lattice of regularly spaced short horizontal bars. The target, either left-tilted or right-tilted, was located on one of the bars in the middle of the display. The salient line was one of the vertical columns with all the bars rotating 90 degrees, forming a line of orthogonal bars to the texture. In Experiment 1, the location of the target and the salient line varied independently from trial to trial. Therefore, the salient line did not predict target location. We replicated our previous findings in that a target presented directly on the salient line was discriminated less quickly and less reliably. In Experiment 2, the target was on the salient line for 60 % of the trials, and the impairments vanished. In Experiment 3, 100 % of relevance was used and the salient line finally facilitated visual search. We argue that high expectation was required to overcome the low-level perceptual grouping effects in our display.

25‧Kuan-Ming Chen
Asymmetric cross-modal effect on time perception depends on stimulus duration

Chen and Yeh (2009) observed asymmetric cross-modal effects in time perception: A concurrent sound expanded perceived visual duration but not the reverse. We examined the cross-modal effects across different stimulus durations to test whether modal variability is constant (Gibbon et al., 1984). A two-interval-forced-choice paradigm with the method of constant-stimuli was used. Participants indicated the longer duration after viewing two successive intervals with different durations. One of the intervals served as a standard while the other as a comparison. The comparison was either in the same modality as the standard (visual or auditory) or with the other modality added, and participants judged the duration based on one modality while ignoring the other. To exclude the possibility of biased preference on bi-modal comparison, in one experiment participants indicated the shorter duration. Results showed that sound expanded perceived visual duration and this cross-modal effect did not result from biased preference for bi-modal stimuli. However, this auditory bias on visual duration judgment declined with longer duration whereas no visual bias on auditory duration judgment was found across all durations. These results support the view that relative variability between audiovisual modalities changes with the mean stimulus duration (Wearden et al., 1998).

26‧Shih-Yu Lo
Effect of sound on visual persistence

A briefly presented visual stimulus remains visible for 150 ms after offset, and this visual persistence should have led to blurry vision. Previous research has identified the factors reducing such visual persistence, for example, eye movements and object updating. Beyond the within-modal interaction between visual processes, we hypothesize that auditory stimuli may modulate visual persistence. Based on the previous finding that a change in visual size influences visual persistence, we additionally presented sounds to see whether the sound change in pitch can modulate visual persistence. Participants were required to report whether they saw one or two discs in the final frame, after viewing a disc moving continuously in a semicircle trajectory. If visual persistence remains, participants should perceive two discs even when there was only one disc because visual persistence of the last second frame should make it visible and co-exist with the last disc. Two experiments showed that visual persistence was reduced when the sound change in pitch occurred 186 ms prior to the change in visual size. However, this did not occur if pitch change occurred concurrently with the size change. These results have potential applications on display industry to solve motion-blur problems in LCD displays.

27‧ Ming-Chou Ho
Object-based attention: Spread, scanning and shift

Sensory enhancement, scanning prioritization, and attentional shift are often used to account for object-based attention (OBA). The present study reports on two series of four experiments using a modified flanker paradigm and on both reaction time and data-limited procedures to examine the possibility of (1) sensory enhancement within objects in the experimental conditions used to support the strong scanning prioritization; and (2) the weak versions accounting for OBA. In the target positional certainty condition, a target is always in the center of the central rectangle with the flankers in the same or in the different rectangles. In the uninformative and informative cue conditions, a central cue is either uninformative about the target location (50%) or it indicates the rectangle where the target is most likely (80%) or less likely (20%) to appear. The flankers and target in the cue condition never appear in the same rectangle. Findings based on the data-limited procedure suggest the sensory enhancement within the attended object in target positional certainty. Further, the current study provides evidence for the weak versions, because sensory enhancement and scanning prioritization are observed together in 80% of the cue validity condition and neither of these two accounts is absent.

28‧Fuminori Ono
Attention can distort visual space backwards

A brief visual cue that attracts attention repels the perceived position of a subsequent target from the focus of attention (attentional repulsion effect). In the present study, we presented the visual cue “after” the presentation of a visual target and found that the perceived location of the preceding target was shifted toward the location of the following cue (attentional attraction effect). The attraction effect decreased when a masking pattern was presented between the target and the cue but the repulsion effect did not. The last experiment showed that the attraction effect due to following cue was countered by the repulsion effect due to preceding cue. The present findings suggest that both preceding and following contexts modulate the perceived location of a briefly presented stimulus.

29‧ Sung-En Chien
Feature-location binding when tracking moving objects do not affect distribution of attention within objects

Tracking multiple objects in a dynamic environment is one of important abilities of our vision. It has been proposed that ‘object files’ keep track of items of current interest and accumulate information about them as they move and change (Cavanagh & Alvarez, 2005). In the literature of visual attention, it is known that the distribution of attention within objects affected by objects’ features (shape and color). The purpose of this study was to examine whether memory for binding objects’ feature (colors and spatiotemporal information) affects how visual attention spreads within objects. The experiment combines an irregularity detection task and a probe detection task in order to test if participants fail to bind objects’ feature and spatiotemporal information in memory (i.e., fail to detect irregularity), whether visual attention spreads within objects would be affected? The results showed that distribution of visual attention within objects was not affected by unsuccessful feature-location binding, indicating that feature encoding is also retained when tracking objects, but separated from location information.

30‧ Wei-Lun Chou
Subliminal spatial cues capture attention and cause reversed object effects

Mulckhuyse, Talsma, and Theeuwes (2007) showed that a subliminal spatial cue can capture attention to its location. According to the spreading hypothesis of object-based attention (e.g., Richard, Lee, & Vecera, 2008), a subliminal cue that can successfully capture attention to a location within an object should also cause attention to spread throughout the whole cued object and led to the same-object advantage (i.e., faster response to a target within an cued object than within a non-cued object). By adopting the two-rectangle method (Egly, Driver, & Rafal, 1994) and using an effective subliminal cue, we tested whether the same-object advantage can be obtained as with the classic suprathreshold cue. Opposite result patterns were found for subliminal and suprathreshold cues. A subliminal cue presented 100 ms before the target facilitated the cued location and led to different-object advantage, whereas when it was presented 1000 ms before the target it inhibited the cued location and led to same-object advantage. The use of a suprathreshold cue replicated the conventional location and object effects. This finding was inconsistent with the spreading hypothesis. A subliminal cue strengthens between-object links, which are coded primarily within the magnocellular to dorsal pathway that governs the visual guidance of action.

31‧Satoko Ohtsuka
Interacting properties of spatial and non-spatial attention revealed by cueing paradigm

The literature has shown the property of spatial attention. For instance, its facilitatory effect is followed by inhibitory effect as is a typical result of cueing paradigm. The property of non-spatial attention and its relationship with spatial attention, on the other hand, is to be studied. The present study aimed to study the interrelationship of two aspects of attention. In Posner type cueing paradigm, the cue and target was determined by a combination of position (left/right) and color (blue/red), so the trials were divided into position-valid and -invalid conditions, and into color-valid and -invalid conditions. The SOAs were 300 and 1,200 ms. Observers were required to respond to the target position or color in separated experiments. RT data in position discrimination experiment showed the facilitatory and inhibitory effects on position-validity analysis whereas did no effect on color-validity analysis. The pattern of data was definitely different in color discrimination experiment: No facilitatory effect was shown on position-validity analysis whereas facilitatory and inhibitory effects were on color-validity analysis. These suggest a mutual relationship between spatial and non-spatial deployment of attention. On these evidences a property of attention resource would be discussed.

Face: Other Race Effect

Please tick ▼ for abstract content

32‧Hsin-Yueh Hsu
Exploring the Other-Race-Effect in 6-, and 9-month-old Taiwanese infants and adults

The present study investigated whether the other-race effect (ORE) exist in Taiwanese infants aged between 6 and 9 months when the visual system is still maturing. The faces of three ethnic groups (Asian, Caucasian, African) and of three levels of difficulty: Easy (change identity), Median (change one feature and one spacing), Hard (change one spacing) were included. Experiment 1 adopted the visual-paired-comparison (VPC) task to assess 6-month-old infant’s discriminability for the familiar/novel faces. Each infant completed a total of 18 trials (3 ethnic groups  3 levels of difficulty  2 reversed locations) of three ethnic blocks in randomized order. Looking preference (fixation time) was recorded by two observers. Experiment 2 further tested 9-month-old infants’ discriminability with the same condition. Experiment 3 was an adult control experiment that offers a performance reference point. Our results showed that 6-month olds can discriminate “easy” faces of all three ethnic groups and “median” Asian faces. 9-month olds can further discriminate “hard” Asian faces and “median” Caucasian faces. Taken together, these findings suggest a mixture of general improvement in face discrimination task as well as the other-race effect. The role of visual experience seems to be facilitation rather than merely maintenance.

33‧Jie-Ju Lin
The effect of spatial frequency on the Own Race Advantage in face recognition

People are better in recognition of own-race than other-race faces, which is known as ORA (Own Race Advantage). Past studies have identified components and configurations of the faces to be factors influencing facial recognitions. In previous study, Goffauz et al., (2005) indicated changes in components lead to better recognitions in high spatial frequency (SF) faces while changes in configurations lead to better performances in low SF faces. To date, the issue of how SF modulates ORA is still unclear. In experiment 1, memory recognition task, we manipulated five levels of SF (from lowest to highest), own- and other-race effects. The results showed that ORA was present in middle-ranged SF. Experiment 2 was the training-with-feedback discrimination task. We did two groups of low and high spatial frequencies and examined the effect of familiarly on ORA. Results indicated that under low SF, same-race performance is better than that of other-race faces, but the ORA appear in certain (optimal) SF. Under high SF, component clues are important in identifying other-race facial characteristics. Under slightly high SF, we find other-race advantage. Taken together, our findings suggest a dynamic recognition mechanism in human facial recognition system.

34‧Tsung-Ting Wang
Influence of life experience on other race effect

It is well known that we usually do not recognize faces of other-race as well as those of own-race. Such other-race effect, or ORE, has received extensive support. One fMRI study suggested that people's Fusiform face area (FFA) respond more activated to same- than other-race faces. We expect that life experience will affect this result: Caucasian participants who have stayed in Asian countries, their sensitivity for Asian faces will be modulated by their years of stay in that country. For example, the longer the Caucasian subjects live in Taiwan, the more activity of their middle Fusiform gyrus (mFG) will be toward Asian faces. Subjects are all Caucasians, with varying years of stay in Taiwan. They underwent fMRI while viewing pictures of Caucasian faces, Asian faces, and objects. Across these 14 subjects, we found their mFG activation of “Asian_faces minus Objects” are significantly correlated with to their duration of residency in Taiwan, supporting the experience modulation hypothesis. Besides, there is another mFG region whose “Asian_faces minus Caucasian_faces” seems marginally correlated with years of stay. With further evidence bridging the neural substrates and the behavior with fMRI, we again exemplify the adult brain plasticity.


36‧Myung Chan Lim
Ethnic diversity_gender perception

Two experiments were conducted to examine the Impact of facial features(e.g., eyes, nose and mouth) and expressions(e.g., neutral, happy, angry, sad and surprised) across different ethnic groups(d.g., caucasian, asian, & black) on perception of gender and facial expression. A series of pictures with 3 different facial features and 5 different expressions across 3 ethnic groups were presented to the participants and their perception of gender(Experiment 1) and facial expression(Experiment 2) was examined. The results showed that perception of gender and facial expression was depending on in order of eye, mouth, and nose. The interaction effect between ethnicity facial expressions was found; gender perception was most accurate for happy in Asian, angry in Black, and sad in Caucasian. In perception of facial expressions, differences across ethnic groups were found only for perception of male faces; Black and Asian males were perceived more accurately than Caucasian males. These results suggested the generality of gender perception even when only parts of facial features were available, and facial features specific to an ethnic groups could be helpful for gender perception.

37‧Cheuk-Fai Chung
Identification Accuracy and Confidence Reliability in Cross-racial Lineup Identification

Much previous research suggests that when testing eyewitness memory in legal settings, sequential lineups, where suspects are displayed one-by-one, are more effective at promoting correct identifications and reducing false identifications than simultaneous lineups, where suspects are all displayed together. However, most previous studies used only own-race photographs. The present research systematically tests and compares the identification accuracy and confidence reliability of these two types of lineups across own-race and cross-race lineups. Both Chinese and Caucasian participants were presented with target photographs and were later asked to identify them from target-present and target-absent lineups. The results found a consistent own-race advantage only in sequential lineups, but not in simultaneous lineups. In addition, there was a mild positive correlation between accuracy and confidence. A subsequent experiment showed that the own-race advantage for sequential lineups disappeared when faces were inverted, consistent with evidence that inversion disrupts holistic processing. These results have clear implications for the use of lineups to test eyewitness memory in legal settings, and suggest that lineup design should be based upon the similarity in race between the witness and the suspect.


Saturday, July 24

Morning Poster Session: 10:40-11:40 (Odd numbers present)

Afternoon Poster Session: 15:25-16:15 (Even numbers present)

Color Vision I

Please tick ▼ for abstract content

1‧ Chu-Heng Lee
Taiwan's traditional Koji colors of the visual cognition research

“Taiwan Koji Pottery” has the multi-color characteristic, these cultural relic could concrete show the colors utilization and presentation in the early Taiwan. It is an indispensable subject to study Taiwan color culture. The study finished the program “Index and Filter of Colors and Shades” by using the linear function, and then it can investigates the cognition of the color appearance expression from the populace by this tool. The final result developed the accordant color expression to conform the Taiwan color name. This research ran through the study of tradition color culture, the color investigation with statistics and the sign of color standardization system. On this way, it shows the culture of tradition color . On the other hand, it also establishes feasible expression foundation, simultaneously returns to the history of color to discuss the development of colors by the literature. All of studies anticipated that the findings may promote the application value for Taiwan color culture.

2‧Kahiro Matsudaira
Derivation of color confusion lines and copunctal point for dichromat observers from color discrimination thresholds

Our final objective is to develop a method of finding a copunctal point in the display RGB color space so that any sets of confusion colors can be identified in the device dependent color space. In the current experiment we tried to find a copunctal point from the discrimination thresholds around reference colors.
A reference and a test square patches were presented side by side on a liquid crystal display. The subject’s task is to set the color of the test where the color difference between the test and the reference is just noticeable to him/her. In a single trial, the color of the test was only adjustable along one of 18 directions around the reference color which was held constant meanwhile. Thus eighteen threshold colors were obtained around each reference. The copunctal point was determined where the squared sum of the distance of the threshold colors to the confusion line drawn from it to each reference colors is at the minimum. Copunctal points were obtained by pseudo-dichromats (color normal observers with a dichromat simulator) and compared in different color spaces such as CIE xy, CIE u’v’, and the device dependent rg chromaticity diagram.

3‧Kaori Ogawa
Color discrimination on various test of color deficiency

Many studies have shown the red-green discrimination threshold of people with color deficiency was high. It has not, however, been investigated the influence of a test color on the threshold. In this study, we measured color contrast threshold on five test colors (grey, red, green, blue and yellow) in a LMS color space. Stimulus was presented in a square array of four 1°squares with a temporal Gaussian modulation. One protanomalus, three deuteranomalos and four color normal subjects were tested. The thresholds of eight directions were obtained on L+M (luminance), L-2M and S-(L+M) plane, respectively. We also measured individual LM ratio at equiluminant setting by a flicker photometry. Results showed both protanomalous and deuteranomalous had large color contrast thresholds in the L-2M direction for all test colors. Only on the red background, the color contrast thresholds in the L+M direction of deuteranomalous were larger than other subjects. In flicker photometry, each group of protanomalous, deuteranomalous and normal subjects showed the different trend of equiluminant LM ratio, respectively. The LM ratio setting of protanomalous was the largest, implying the contribution of luminance component to the result of color contrast thresholds. (discuss relations between the color discrimination threshold and the equiluminant LM ratio.)

4‧Hirotoshi Nishita
Characteristics of color categories of dichromats

Dichromats cannot discriminate some colors because they lack a type of cones. However it was reported that they could categorize colors by using color-names as trichromats did. We reported the color memory-matching experiments designed to explore the higher color-vision mechanism of dichromats (Uchikawa and Nishita, ICVS, Braga, 2009). In the experiments a test color sample was presented for 5sec, and then after 30sec the observer started selecting a color sample from a set of 1406 simulated OSA UCS that matched the test color sample in his memory. 100 test colors were used in the experiments. The results showed that color samples selected by color memory were restricted in the categorical color-naming regions for trichromats, but this tendency was not clearly shown for dichromats. In the present study we further analyzed the data in order to clarify color categories of dichromats. It was found that dichromats’ matching points tended to concentrate on a single point in the OSA-g direction whereas trichrimat’s matching points spread on the j-g plane. These results suggest that dichromats’ color categories are less than those of trichromat and exist not along the r-g direction but just along the y-b direction.

5‧ Katsuaki Sakata
Effects of chromatic adaptation on visual search for orientation

The present study demonstrates the effects of chromatic adaptation on visual search tasks for right and left oriented Gabor patch stimuli defined in colour or luminance. Observers, who were confirmed their colour vision by Ishihara pseudo-isochromatic plates, searched for the different colour and orientation patch with an eye after chromatic adaptation by the other eye. The detection performance of chromatic adaptation condition always resulted in lower than non-adaptation condition. We conclude that the visibility of stimuli colour in a search task failed to draw on chromatic information from the target. The effect could be explained by the cortical mechanism of chromatic adaptation.

6‧Jinhui Yuan
Understanding cone spectral selectivities from information maximization and color constancy

Lewis and Zhaoping (2006) proposed to understand cone spectral selectivities by the need to maximize the amount of information extracted about the object surface spectra from natural scenes under a fixed illumination. The predictions from this proposal are consistent with the spectral selectivities of the S and M cones, but require the L cone to prefer much longer wavelength than observed, if one ignores the consequent penalties such as increased chromatic aberration and increased noise level caused by this preference shift toward longer wavelength. In this work, we propose that color constancy under variable illumination is an additional constraint on the cone spectral selectivities. In particular, to discount the variable illumination, the spectral selectivities should make von Kries coefficient law hold as much as possible for the ensemble of surface reflectance and the ensemble of illuminations in natural scenes. Using information theoretical terms, we formulate the combination of two constraints, information maximization and color constancy, as an optimization problem. Our proposal can better predict the cone spectral selectivities than the previous one, it particularly constrains the preferred wavelength of the L cones away from the long wavelength end of the visible light spectrum.

7‧Ayae Tajima
Color Constancy in Photographs and in Cartoon images

Color constancy is a mechanism required to the stable recognition of object color. It has been suggested that recognizing a room as a three dimensional space is important for color constancy, and that in a two dimensional photograph is lower (Mizokami et al, 2004). We predicted that color constancy in a cartoon image might be even lower than that in a photograph because of less spatial information (such as shade and depth) and weak 3D recognition in it. We examined color constancy in three conditions: A real room, photographs of it and cartoon images made from these photographs. We built a booth arranged like a natural room containing color patches, and took photographs in the room under two illuminants: fluorescent lamps with correlated color temperature 5000K and 2700K, respectively. A subject observed an image on a CRT display in a room illuminated by 5000K illumination, then answered the colors of stimuli in the image using the elementary color naming method. The result showed that color constancy in the real room condition was the highest among all conditions and that in the cartoon condition was the lowest. However, the difference between photograph and cartoon images would depend on the hue of stimulus.

8‧ Ryoichi Ishibashi
Colorimetry-free color management system for display based on color constancy

Many of color management systems (CMS) are to achieve colorimetrically equal color. However, in metameric color match, equal colorimetric values assure equal color appearance only to the standard observer but not to every observer in practical situations. Another and more serious problem is the effect of color adaptation to the illuminant. Even though spectral composition from the display is held constant, color adaptation to the illuminant may change the color appearance. In order to cope with observer and illuminant variabilities, we developed a new CMS based on color constancy for reflecting surfaces. The idea is to derive a conversion matrix for RGB values from one to another environments through a visual or subjective color match with reflecting color chips under the illuminant. Another advantage of the new CMS is a procedure where no colorimetric measurement is required. A conversion matrix between two environments is calculated by applying a pseudo-inverse matrix to an equation consisting of pairs of RGB values from visual color match. In the experiment, the conversion matrices were derived using various sets of color chips and their performances were compared. As a result, 5 or more color chips of uniformly distributed hues give better and stable performance.

9‧ Manabu Akimoto
Measurement of luminance and chromaticity distribution by a digital camera

A digital camera may be a useful instrument when you want to obtain chromaticity and luminance distribution from the natural scene. One of the problems is its small dynamic range. The objective of the study is to develop formulae to calculate CIEXYZ from RGB pixel values of several images taken with different exposures. In calibration, the Macbeth color chart under different illuminant levels were taken by a digital camera and measured by a color meter CS-100A. The luminance of color charts ranged from 0.59cd/m2 to 6140cd/m2. At each illuminance level, several images were taken with different exposure times from 0.00625sec to 10sec. From the collection of RGB and XYZ, the multiple regressions for the logarithms of each trisutimulus value were obtained like log10X= XR+ XG+ XB+ X( t). The coefficients , , are constant while the intersect was a function of exposure  t. The range, within which the multiple regressions are predictive, was also determined as functions of exposure  t. The error between the actual measurement values and the predictived values were 8% in X, 6% in Y, and 5% in Z. Therefore further refinement is needed for a digital camera as color meter.


Please tick ▼ for abstract content

10‧Yu-Chieh Chang
May threatening faces catch your eyes?

Based on the evolutionary view, people can quickly detect dangerous signals and respond unconsciously. Further, according to the lateralization hypothesis, right brain hemisphere is specialized for processing emotional information. In order to explore current issue, the influence of threatening faces on goal-driven saccadic eye movement was investigated.
For the purpose of generating goal-driven saccadic eye movement, participants were directed by a left or right arrow and then saccade toward the arrow-directed rectangle as soon as possible. At the same time, a face picture was presented at the same or opposite rectangle relative to arrow’s direction. In some trials, no picture was presented, which was called as no-picture condition.
The result showed that if faces were presented at left visual hemifield, fearful and angry faces presented for only 17ms could delay onset of anti-saccade. But happy faces needed 167ms presentation time to cause longer anti-saccadic latency than no-picture condition. The above-mentioned effects were absent if pictures were presented at right hemifield.
In conclusion, the threatening faces presented at left hemifield can delay onset of anti-saccadic eye movement but cannot overtly catch eyes toward. Both evolutionary view and lateralization hypothesis were preliminarily supported.

The result showed that if faces were presented at left visual hemifield, fearful and angry faces presented for only 17ms could delay onset of anti-saccade. But happy faces needed 167ms presentation time to cause longer anti-saccadic latency than no-picture condition. The above-mentioned effects were absent if pictures were presented at right hemifield.

In conclusion, the threatening faces presented at left hemifield can delay onset of anti-saccadic eye movement but cannot overtly catch eyes toward. Both evolutionary view and lateralization hypothesis were preliminarily supported.

11‧ Yi-Chin Chou
The discrimination experiment reveals the nonlinear properties of the facial expression analyzer

To understand the response properties of facial expression analyzer in the visual system, we measured the facial expression discrimination threshold at various expression valences.
We constructed a series of facial images with different happy/sad valences by morphing happy and sad faces from the POFA database (Ekman & Friesen, 1976) and divided each face image into upper and lower parts. In each 2AFC trial, either one part of the image was the same in both intervals while the other part differed in morphing level. The observer was to judge which of the two composite faces appeared happier. We used the PSI adaptive threshold seeking algorithm to measure the morphing threshold at 75% correct response level.
When the morphing level of the lower face was fixed, the discrimination threshold for the upper face was similar regardless the base morphing level. On the other hand, when the morphing level of the upper face was fixed, the discrimination threshold first decreased and then increased as the base morphing level of the lower faces increased. Our result suggests that the upper face contributes little to happy/sad discrimination. In addition, the dipper shape of the discrimination function suggests a nonlinear response property in the happy analyzer.

12‧Suzane Vassallo
Visual scanning of emotional facial expressions in traumatic brain injury: A case report

Introduction: Neurological damage may cause the interpretation of facial expression to breakdown at various stages of processing. Visual scanning of emotional facial expressions has been examined extensively in some patient groups (e.g., schizophrenia), though this is not the case in those individuals with a traumatic brain injury (TBI). We present an interesting case report which demonstrates how the visual scan path in TBI can deviate from neurologically normal controls.
Method: Participants were one adult male (LY) with impaired ability to interpret facial expression as a result of severe TBI and 3 neurologically normal male controls. Stimuli were 18 static pictures of facial expressions depicting the six basic universal emotions (sadness, happiness, anger, surprise, fear and disgust). Eye movements were recorded using a binocular infrared eye tracker.
Results: LY’s pattern of scanning differed significantly from that of the neurologically normal controls. For the controls, the majority of fixations were within the internal facial region (that area including the eyes, nose and mouth). In contrast, LY’s scanning was more dispersed (i.e., hyperscanning), with frequent foveal fixations to external peripheral regions (hair, ears, forehead, stimulus background).
Discussion: Impaired visual scanning can contribute to impaired interpretation of facial expression after a TBI.

13‧ Chia-Wei Liu
On-line recognition revealed the in-group advantage in negative facial expressions.

Facial expressions are suggested to be universal. However, studies show the in-group advantage of the recognition of facial expression (Matsumoto, 1989; 1992). People understand emotion more accurately when they are expressed by members of their own culture group. We investigated whether this in-group advantage was showed in Taiwanese by adopting an on-line emotional discrimination task in which participants were asked to discriminate positive(happy) versus negative(sad, fear, and anger)faces among Eastern and Western faces. The materials included Eastern faces from Taiwanese Facial Expression Image Database (Chen, 2007) and Western faces from Ekman and Frisen’s database (Ekman & Frisen, 1979). Both reaction times and accuracies of performance were measured. Results showed that even participants can discriminate positive and negative faces accurately; they responded significantly faster to negative Eastern faces than to negative Western faces. However, no such culture difference was found to positive faces. The results revealed the in-group advantage of the perception of facial expressions was specific to negative emotions and question the universality of facial expressions of emotion.

14‧Yung-Hao Yang
Auditory-visual integration facilitates unconscious processing of facial expression

Emotional expression of faces plays an important role for social interaction. Previous studies have shown that unconscious processing of facial expression activates the same brain areas as the visual-auditory integration of facial expression (i.e., the amygdale and posterior superior temporal sulcus). Here we examine whether auditory-visual integration could occur for audible voice but invisible facial expression. We adopt the continuous flash suppression paradigm (Jiang, Fang, Huang, & He, 2005; Tsuchiya & Koch, 2005) in which a face in one eye was suppressed by dynamic flash masks in the other eye, and the manipulated sound was presented at the same time. Experiment 1 examines whether voice is special for multisensory integration to occur for invisible face. We found that either a beep sound or a voice can facilitate the detection of an inverted face; however, only the voice can facilitate the detection of an upright face. In Experiment 2, we manipulated the semantic congruency between the voice and the face, and found a semantic congruency effect: faster detection for matched emotional prosody and facial expression than for unmatched pair. Audio-visual integration can occur even when participants are unaware of the face, and semantic congruency facilitates the formation of a unity percept.

15‧Chia-Chen Wu
Neutral features induce emotion in schematic faces

Schematic faces usually express emotion with strong emotional features, such as a downward curve represents a smile and faces with this curve induce positive emotions. In this study, we explored whether neutral schematic faces could induce emotion. Five factors: gaze directions, shape of eyes, noses, mouths, and outline of faces, were manipulated. The eyes were directly gazed or avertedly gazed, with vertical oval, circle, or horizontal oval shapes. The faces were with a round nose, a vertical line-shaped nose, or without nose. The mouth was either a horizontal line or none. The outline of the faces was either round or oval. All these features were emotionally neutral. The faces with noses must have mouths to avoid mistaken of noses as mouths; otherwise all the factors were orthogonally assigned into 72 faces. A questionnaire with these neutral faces and 20 emotional faces were submitted to data collection. Five-point Likert scale was used. Seventy-one valid questionnaires were obtained. We found that (1) averted gazes, (2) faces with mouths, and (3) faces without noses, tend to be perceived as in negative emotions. We argue that these neutral features may induce social cues in communications that relate to emotional perception.

16‧Ya-Huei Hsu
How does inferotemporal cortex respond to average faces?

Face perception is an important ability, which including face identification and recognition. Neuroimaging studies have found several related brain regions, among FFA, OFA, STS and aIT(Kriegeskorte,2007). Loffler(2005) used synthetic faces in fMRI study, and found FFA activations for the individual faces were stronger than the average faces. However, the synthetic face changes were primarily achieved by manipulating external face contours. Jiang(2009) pursued this issue and demonstrated the contour change invoked the largest response in the rFFA. This raises a concern of whether the lower FFA responds to average faces by Loffler is either caused by the face averaging process, or confounded by the change of face counter. This study morphs real individual faces to create sets of 3-, 6-, and 10-averaged faces, while controlling face contours to the same egg-shapes, and asks subjects judge the two face presentations are same or different sex. The ROI analyses show that compared to average faces, individual faces elicit the highest activations in FFA. In addition, group analyses identify and find similar activations in fusiform gyrus and aIT. We conclude that the average faces do indeed lessen FFA and aIT activations, suggesting their joint importance in face processing.

17‧ Yu-Jen Tsai
The effect of attention in face selective areas of bird experts: reevaluating the expertise hypothesis

Fusiform Face Area (or FFA) has been associated with at least two possible accounts: face specificity and perceptual expertise. These two views have exchanged over ten some years, with no firm conclusion reached. One recent claim made by supporters of the face specificity was that when experts see the objects of their expertise domain, the increased FFA response is an “overall increased attentional engagement for these stimli” (McKone et al., 2007). If this hypothesis is correct, we shall predict that when experts view the stimuli of expertise inattentively, their FFA response should be no different than that of a control object class.
In this study, we explicitly test this claim by asking bird experts and novices to engage (and disengage) their attention from faces, bird, and car images. The task consists of a central fixation, occasionally dimming (central task, or unattended condition), and six flanking and consecutive pictures pairs (periphery task, or attended condition). The results show that for experts, the FFA activation for bird-unattended condition is always larger than that of car-unattended, inconsistent with the face specificity hypothesis. This situation is not seen in bird novices.

18‧Miao Song
High Level Facial Aftereffects Induced by the Non-face Meaningful Objects

Our previous study shows that the facial aftereffect can occur not only in real face but also in the simple visual stimuli, when this simple stimuli share the same low level visual properties with the face. In terms of this observation,we introduced a novel facial identity adaptation paradigm to investigated the relationship of neural systems processing the face and the non-face meaningful objects, this issue is a long-term debate in the face study. We measured the facial identity aftereffect within the morphing test faces, using the face, the color chip, and the three kinds of non-face objects(i.e., bird, house, apple)as adapting stimuli. All the adapting stimuli share the same color, brightness, and contrast with that of the test face to minimize the influence of adaptation effect from low level. The results show that all five adapting stimuli generated the significant adaptation effect, and the adaptation effect by the real face is significantly stronger than that by the other four adapting stimuli. In terms of these observations, we discussed the theoretical implication for understanding the functional organization for faces and meaningful objects in human sensory system.

19‧Yu-Ting TING
The interaction between face- and object-selecive areas by face-like objects: an fMRI study

In daily lives, many objects with their“two-eyes-on-top, one-mouth-below" configurations, look like human faces. In one MEG study, Hadjikhani (2009) found that subjects' FFA (Kanwisher et al., 1997), responded similarly to those“face-like objects" as to faces (both peaking around 165ms), providing the first neuroscientific support of those“face-likeness”impressions. However, face-like objects are by default objects, so they should active object-selective region--LOC (Malach et al., 1995) as well. To investigate the temporal dynamics between FFA and LOC in viewing face-like objects, we use behavioral and fMRI methods to disentangle their relations. Subjects did speeded face-or-object judgments on faces, objects, and face-like objects. First, the behavioral results (N=17) show, that the reaction time of face-like objects are significantly longer than that of faces or objects, suggesting the effect of interaction between face- and object-selective areas. Second, the fMRI results (so far N=10) show that FFA activity for face-like objects is in general lower than that for faces, but higher than that for objects. In LOC, the activity for face-like objects is higher than that for faces, and similar to that for objects. We are currently working on the Granger Causality Mapping to see their inter-relations between FFA and LOC, and other face-and-object processing networks.

Object & Scene

Please tick ▼ for abstract content

20‧Yi-Wen Chen
Integrating auditory and visual information in bird experts: an fMRI sutdy

Our brain integrates the multisensory information from various senses: you can recognize the objects either by vision(seeing your mother’s face), by audition(hearing your mother’s voice), or by both. Where and how the brain integrates different multisensory information? Beauchamp(2004) referred to the Superior Temporal Sulcus(STS) as one of the important sites: they showed that not only did the visuo-auditory combinations(Sheep image + “Meeeh sound”) activated the highest, compared to either audition or vision alone, in STS, but also its activity for the congruent information was higher than the incongruent(e.g., mom’s face + papa’s voice). How does this integration ability arise? Of course our everyday experience is very important. But in order to show this point, we can’t rely on the common scenarios where all of us are expert. Instead, in the current study we recruit bird-watchers, acquiring their behavioral expertise indices by both sequential matching(visual expertise) and visuo-auditory congruency judgments(auditory expertise). The fMRI exp is the event-related version of the same visuo-auditory congruency judgment task. The fMRI results, as we calculate the response difference in STS between two conditions(|congruent – incongruent|) are highly correlated(r = .84; n=8) with birder’s expertise, supporting the importance of experience in the multisensory integration.

21‧ Yuki Yamada
The uncanny valley phenomenon is explained by deterioration of object evaluation due to categorization difficulty

When the human likeness of non-human objects increases, evaluation of the objects drastically decreases at a certain point (the uncanny valley phenomenon). The present study examined whether this deterioration of object evaluation is related to the cognitive difficulty in categorizing objects into given two categories. In Experiment 1, we employed 11 morphing images generated from images of a real person and a cartoonish person with morphing rates varying from 0 to 100% in steps of 10%. Observers were asked to categorize each image (real or cartoon), and we measured the response latency from the onset of the image as an index of categorization difficulty. Moreover, the observers were asked to evaluate likability of each image on a 7-point scale from -3 (dislikable) to 3 (likable). As a result, the longest latency and the lowest evaluation score were observed at the same morphing rate. In Experiment 2, we obtained the results similar to Experiment 1 by using dog images (real, stuffed, and cartoonish dog). These results suggest that low processing fluency in the images that were difficultly categorized induced negative evaluation of the image, explaining the uncanny valley phenomenon.

22‧ Chia-Pei Lin
Key dimensions of visual object representation in macaque AIT

One of the main bottlenecks to understanding information processing in higher cortical areas is that we don't know the key dimensions which underlie coding in these areas. At the end of the macaque ventral visual pathway, the anterior inferotemporal cortex (AIT) has been hypothesized to encode anywhere from 36 to a nearly infinite number of shape dimensions via multiple scales of functional organization. Here, we recorded from 64-site multielectrode arrays spanning 1.4x1.4 mm (distance x depth) in AIT of anaesthetized Macaca cyclopis monkeys.
We presented two stimulus sets, one consisting of color, grayscale, and silhouette versions of 112 objects, and the other consisting of grayscale images of 240 rendered 3D objects. We applied principal components analysis to the responses at each (~2 mm2) location. Remarkably, the ensemble activity patterns were consistent across the independent stimulus sets for each location and differed across locations. Each pattern was differentially driven by subsets of stimuli with opposite features, essentially defining a key dimension in stimulus space for each location. We propose that a collection of such key dimensions, mapped via piecewise reconstruction using imaging methods, would serve as a useful map for exploring object recognition computations and for cortical prosthetics.

23‧Daniel Yu-Chun Hsu
Rapid development of pose and illumination invariance in anaesthetized macaque AIT via dynamic stimulus correlation

Recent studies have demonstrated rapid (~1 hr) adult plasticity for receptive fields in macaque anterior inferior temporal (AIT) cortex. Such rapid adult plasticity is hypothesized to enable the development of invariant object recognition across changes in pose and illumination, binding the representations of multiple views of the same object. Here, we sought to test whether such rapid plasticity is dependent on mechanisms (e.g. motivation) in awake animals, and to explicitly and quantitatively characterize the role of temporal contiguity in the development of pose and illumination invariance. We presented repeated and contiguous sequences of images of objects at different poses and illuminations at 85 Hz while recording from AIT via 64-site multi-electrode arrays (MEAs). The rapid frame rate and the use of MEAs are designed to increase the likelihood of observing experience dependent plasticity via spike-timing-dependent-plasticity (STDP) mechanisms. We find examples of neurons with increased invariance and others with decreased invariance following training. The changes in invariance were specific to the modality trained (pose or illumination). These results highlight the role of local cortical plasticity excluding reward mechanisms in the rapid development of pose and illumination invariance.


25‧Chih-Ying Lin
It is mine vs. it looks familiar – ERP N250 responses

Is perceptual identification susceptible to self-relevant processing? Miyakoshi et al. (2009) failed to find electrophysiological differences between daily objects that were familiar and those owned by the participant. Current study showed a difference in cognitive neural processing between familiar and one’s own objects by adopting a repetition priming paradigm, The ERP N250 responses to consistent and inconsistent probes (i.e., the priming effect) were recorded and compared when participants made same-different judgments to the prime and the probe. Inconsistent probes resulted in stronger N250 than consistent probes (Chauncey & Holcomb, 2008) for unfamiliar and one’s own objects, but not for familiar objects. Strongest difference between consistent and inconsistent probes was observed at left prefrontal electrode sites for one’s own objects, signaling possible activation of self-memory system (Conway & Pleydell-Pearce, 2000) in the perceptual identification of one’s own objects.

26‧Kun Qian
Scintillating bar illusion

Illusory black spots are perceived within luminance patches placed at the intersection of orthogonal gray bars on a black background (scintillating grid illusion; Schrauf, Lingelbach, & Wist, 1997). The present study reported an interesting phenomenon wherein a similar kind of illusory spots were also observed within patches placed either on a horizontal or vertical bar. We named this phenomenon scintillating bar illusion, and examined its relationship to scintillating grid illusion. Both size and shape of patches are critical visual factors for scintillating grid illusion (Qian, Yamada, Kawabe & Miura, 2009). Thus, we tested the effect of these factors on scintillating bar illusion. There were three stimulus factors: patch size (0.33, 0.46, and 0.59 degrees), patch shape (circle, diamond, and square), and bars (horizontal bars only, vertical bars only, and both-bars). Consequently, illusory spots were observed in the horizontal or vertical bar condition though they were significantly weaker than those in the both-bars condition. The strengths of both scintillating grid illusion and scintillating bar illusion peaked at the similar patch size and shape. These results indicate that the two illusions share a common underlying mechanism.

27‧ Li-Feng Yeh
Functional Organization of Key Dimensions in Human Lateral Occipital Complex

An accurate understanding of the neural representation is required to decode perceptual state from activity in visual cortex. Although the key dimensions in early visual areas can be guessed with high decoding accuracy, such guesswork along semantic or category labels results in poor decoding performance for higher areas and lacks the capability to extrapolate beyond learned categories. To overcome this ‘curse of dimensionality’, we recently developed methods to extract key object shape dimensions from the spike patterns of local neuronal ensembles in macaque inferior temporal (IT) cortex. Here we show that such monkey-derived key shape dimensions activate specific BOLD patterns in the lateral occipital complex (LOC), the human homolog of IT. Four participants viewed images from two independently generated stimulus sets. Images were blocked according to the polarity of the key dimension and by stimulus set. As in monkeys, the human key dimension map appears as discrete patterns of activation tuned to opposing features. Three out of four participants showed significantly higher BOLD activation in anterior LOC for the key dimension stimuli than their low-level feature control. We surmise that key dimensions are likely shared across primate species, supporting a common code for visual object recognition.

28‧Hsin-I Liao
Retrieval of visual percept by paired association of a visual stimulus and transcranial magnetic stimulation (TMS): Objective evidence from a masking paradigm

TMS to visual cortex interacts with retinal input, altering perceptual experience. When double-pulse TMS is applied following a visual stimulus, an "instant replay" of the percept could be perceived. Furthermore, after repeating several pairs of the visual stimulus and TMS, the replay percept could be retrieved by TMS alone without accompanying with visual stimulus (Liao et al., ASSC '08; Liao & Shimojo, APCV '08). Here, we adopted a masking paradigm developed by Vasudevan et al. (VSS '09) to examine whether the retrieved percept causes similar masking effect as well as the replay. A pattern mask was paired with TMS to produce a spatially and temporally specific replay masking effect on a letter identification task. After ten or five trials, TMS followed by the target letter with identical delay is delivered without the pattern mask. In results, the replay of pattern mask was retrieved by TMS alone to degrade the identification performance. When the pattern mask appeared at a different location from the letter, no masking was observed for replay or retrieved percept. That both the replay and the retrieved percept caused location-specific masking provides objective evidence for a common perceptual mechanism underlying the replay and the retrieval.

29‧Rumi Yamada
Influence on Eye Strain from the Size of Visual Stimulus examined by accommodation response time

Elapsed time for changing focus from far to near point, accommodation response time (ART), has been studied as an objective measure for eye strain. In our previous studies ART was measured as a function of fixation duration to far point. As the duration becomes longer ART rapidly increases and gradually saturate. The influence on eye strain from the condition and task was estimated by the time constant. In this study the influence from the stimulus size was examined. Stimulus size was 0.37, 0.72, 1.45 and 2.88deg. Its luminance was 26 and 2.6cd/m2. Fixation duration to far point was 0, 10, and 20sec. Far point/near point was achieved by insertion of the convex/concave lens in front of eye. When the stimulus image through the convex lens comes into focus the observer presses the keyboard. Then the lens is automatically replaced by the concave lens after a certain delay corresponding to one of the fixation durations. The rate of increase in ART with a fixation duration was higher with the smaller stimulus. When the stimulus size is large enough there’s almost no difference in the increase rate. There might be a critical stimulus size and it seems smaller with lower stimulus luminance.

30‧Yl-Woo Lee
The effects of stimulus duration and preceding images on aesthetic experience

Two experiments were conducted to investigate the effect of stimuli duration and exposure of preceding image on aesthetic experience. The stimuli were 270 fractal images and 81 pictures of natural scene and these images were estimated on the aesthetic dimension in preparatory survey. In the experiment 1, fractal images were presented for 20, 400, 1000ms and estimated aesthetically. After the fractal image or pictures of natural scene was presented, aesthetic estimation on fractal image was measured in every trial of the experiment 2. The results showed that the shorter stimuli duration, the smaller variance of aesthetic response. After seeing the pictures of natural scenes, fractal images were evaluated as an unattractive. On the other hand, fractal images were appreciated more attractively when the preceding stimuli were unbeautiful fractal images than beautiful fractal images. These results suggest that aesthetic experience is related to the later visual information processing and aesthetic contrast is stronger than assimilation in aesthetic experience.

Locomotion & Action

Please tick ▼ for abstract content

31‧Jyh-Jong Hsieh
Tool-use and the extension of peripersonal space: Is temporal synchronization sufficient?

Previous studies reported that tool-use induced extension of peripersonal space (EPS), as evidenced by a dynamic, effector-centered congruency effect between visual and tactile perception whose pattern alternated depending on the way the tools were held. We suspect that the effector-centered EPS induced by tool-use may be a special case of temporal synchronization between one’s actions and consequential environmental events. EPS may still occur without physical contact between the effector and its acting site. To examine the sufficiency of temporal synchronization in EPS, we adopted a visual-tactile congruency task which required location discrimination of brief tactile vibrations accompanied by visual flashes on a virtual object which participant could move synchronously and remotely. In Experiment 1 the participant could move a single virtual object at the center of the display and we found the typical congruency effect between the visual and tactile stimuli. Experiment 2 included two virtual objects to create a “crossed” condition in which each hand could move an object at its opposite side, respectively, yet the congruency effect did not alter with the mapping between the hand and its effective site. Taken together, we suggest that temporal synchronization itself is not sufficient to induce plasticity in the peripersonal space.

32‧Gregory Kroliczak
Neural bases of transitive and intransitive gestures during perception and imitation

Recent fMRI results indicate comparable responses within left parieto-frontal areas during planning both familiar object use (transitive) and nonobject-related (intransitive) gestures for subsequent production with either limb (Kroliczak & Frey, 2009). Yet, neuropsychological evidence suggests a double dissociation between these gesture subtypes. Using fMRI we tested whether distinct neural representations support perception and imitation of transitive vs. intransitive gestures. Twelve right-handed, English speaking adults watched and then imitated gestures performed by an actor in 2.75-s videos. Neural responses to movies depicting transitive and intransitive gestures were used to determine if different areas might be involved in their perception. We also directly contrasted activity associated with imitation of transitive and intransitive gestures. Watching transitive gestures invoked stronger activation within several areas, including right lateral occipital cortex. The inverse contrast did not yield any activation. Whereas imitation of transitive gestures resulted in greater activity in left dorsal premotor cortex, imitation of intransitive gestures invoked stronger bilateral activation in parietal operculum and medial prefrontal cortex, and right anterior intraparietal cortex and precuneous. These findings suggest that transitive gestures are more demanding to process visually. Consistently with prominent models of praxis they also indicate that the two gesture categories can be dissociated during imitation.

33‧Mitsuhiro Yoshida
Driver's gaze control to modulate steering performance in accuracy and workload

Drivers utilize near visual information to control their car position accurately, and far visual information to make a plan of steering (Land and Horwood, Nature 1995). Drivers' steering accuracy was deteriorated by occluding near visual fields, and their workload was increased by occluding far visual fields (Kishida et al, SAE 2009). We aimed to investigate if driver's gaze control to a near or far region modulates their steering performance. Experiments were conducted using a driving simulator, where we manipulated drivers' gaze using a fixation point with varying car's speed. The car's lateral displacement from the road center and the steering angle speed were measured as a driving accuracy index and a workload index, respectively. We found that drivers' steering was more accurate with looking at a nearer region, and their workload decreased with looking at a far region. The slower car simulation increased the driving accuracy, particularly with a near gaze. The faster car increased workloads more with the near gaze than the far gaze. Thus, the gaze control to a near region is effective for the accurate steering without increasing workloads when a car is slow, while the far gaze is effective for decreasing workloads with a faster car.

34‧Jing Chen
Separating the contributions of radial flow, splay and bearing angle information to lane-keeping control

While bearing, radial flow, and splay angle information can all contribute to accurate lane-keeping control, previous studies have reported that people rely mainly on splay angles provided by lane edges for lane keeping. In this study, we aimed to separate the contributions of these three types of visual cues to lane keeping. The displays simulated observers steering down a straight path defined by either a pair of posts (providing only bearings) or a part of lane markers (providing bearings and splay angles), and the ground plane contained no flow, sparse flow, or dense flow information. Observers used a joystick to control their movement to stay in the center of the lane while facing random perturbations to both their simulated gaze direction and their lateral position in the lane. The lateral position perturbations affected the use of radial flow, bearing, as well as splay angle cues for lane keeping, but the gaze perturbations affected the use of only bearings. We found that performance improved as more flow information was added to the scene regardless of the presence of bearings or splay angles. In the presence of splay angles, observers would ignore bearings and rely mainly on splay angles for lane keeping.

35‧Joris Vangeneugden
Discrimination of locomotion direction in impoverished displays of walkers by macaque monkeys

A vast literature exists on human biological motion perception in impoverished displays, e.g. point-light walkers. Less is known about the perception of impoverished biological motion displays in macaques. We trained 3 macaques in the discrimination of facing-direction (left versus right) and forward versus backward walking using motion-capture-based locomotion displays (treadmill walking) in which the body features were represented by cylinder-like primitives. The displays did not contain translatory motion. Discriminating forward versus backward locomotion requires motion information while the facing-direction/view task can be solved using motion and/or form. All monkeys required lengthy training to learn the forward-backward task, while the view task was learned more quickly. Once acquired, the discriminations were specific to walking and stimulus format but generalized across actors. Although the view task could be solved using form cues, there was a small impact of motion. Performance in the forward-backward task was highly susceptible to degradations of spatio-temporal stimulus coherence and motion information. These results indicate that rhesus monkeys require extensive training in order to use the intrinsic motion cues related to forward versus backward locomotion and imply that extrapolation of observations concerning human perception of impoverished biological motion displays onto monkey perception needs to be made cautiously.

36‧Yuki Kawashima
Influence of vection stimuli arranged along a road on the car driver

It is common for the car drivers to judge the speed of their cars by the objects moving outside. The purpose of this study is to know whether the moving visual stimuli placed in a line along the road can yield vection to a driver so that the drivers can have the sensation of higher or lower speed than that the drivers would have in the natural scene. The computer simulated 3D highway-driving scene was presented on the hemisphere-screen of 8.5m diameter that produced a wide visual field of 120 deg. The vection stimuli consisted of gray boards placed along both side of the road. They were moving to make a forward vection with different speeds from other surrounding objects. The observers estimated their sensations of speed with the method of paired comparison. Twenty observers participated in this experiment. It was shown that the speed sensation perceived by the observer was modified by the vection stimuli, even when the surrounding objects moved at the same speed, in such a way that the speed sensation increased with the moving speed of the vection stimuli. Our results suggest the moving stimuli along a road could influence the driver's sensation of speed.

37‧Joseph Cheng
Looking where you are going does not help path perception

It has been mathematically shown that when travelling on a circular path and fixating a target on the future path, flow lines for environmental points on the path would be vertical. Here we test whether fixating a target on the future path helps path perception. Observers viewed displays (110°Hx94°V) simulating their traveling on a circular path over a textured ground (T=3 m/s, R=±3°/s or ±6°/s) for 1 s. Three display conditions were tested. In the path-fixation condition, the simulated gaze direction in the display pointed to a target along the path at 20° away from the starting position; in the non-path-fixation condition, the simulated gaze direction was on a target 10° inside or outside the path at the same distance; and in the heading-fixation condition, the simulated gaze pointed to the instantaneous heading (i.e., the tangent to the path). At the end of the trial, observers used a mouse to place the probe (10 m) on their perceived future path. We found that path perception largely depends on solving the translation and rotation problem in retinal flow, fixating a target on the future path does not help the perception of the path trajectory.


Sunday, July 25

Morning Poster Session: 11:10-12:00 (Odd numbers present)

Afternoon Poster Session: 15:40-16:30 (Even numbers present)

Motion I

Please tick ▼ for abstract content

1‧Shinji Nakamura
Effects of retinal eccentricity on jitter advantage in visually induced self-motion perception

It is well-known that uniform visual motion can induce illusory self-motion perception in the direction opposite to the inducer’s motion (vection). Recently, it has been revealed that visual expansion which contains additional jitter components (visual motion alike as external scene shot by camera which was moved forward with “camera-shake”) can induce much stronger forward self-motion perception than pure radial expansion (jitter advantage). In this investigation, the effect of retinal eccentricity on which the visual jitter was adopted was analyzed; the expanding visual pattern was presented on observer’s entire visual field, while the dots in the pattern contained jitter components only in the central or peripheral portion of it. Radius on the central area or inner radius of the peripheral area where the jitter motion was adopted was also manipulated. The result of psychophysical experiment indicated that 1) the jitter advantage was greater in the condition where the stimulus area with jitter motion was larger, and 2) the jitter advantage was slightly robust in the central visual field than in the peripheral visual field. These results would be useful in considering perceptual mechanism underlying the jitter advantage in self-motion perception.

2‧Masato Kawano
Effects of abrupt color and luminance change on reappearance in motion-induced blindness

Motion-induced blindness (MIB) is a remarkable phenomenon by which salient static stimuli disappear when surrounded by moving stimuli. Regarding causes of MIB, Bonneh et al. (2001) reported that visual attention was engaged by a superimposed moving pattern. Additionally, Theeuwes et al. (1995) described that visual attention was engaged by a change in luminance or chromaticity.
This study investigated the detection disappearance duration, as affected by abrupt luminance or chromaticity change of a target, after a stimulus’ disappearance (Experiment 1). We presented two red dots (targets) together with 49 blue cross patterns arranged in a grid (mask) rotating around its center fixation point. Observers were instructed to hold the ENTER key of a computer keyboard while gazing at a fixed point when the left target became invisible. Observers released the ENTER key when the left target became visible. Results show that the disappearance duration was reduced by the changing luminance or chromaticity.
The next experiment varied the luminance change magnitude (Experiment 2). However, no significant difference was found in the disappearance duration. Therefore, we infer that abrupt change of the target factor (luminance, chromaticity) is more important in relation to MIB than the luminance change magnitude.

3‧Hiromasa Takemura
Effects of surrounding motion on motion segregation

Motion segregation and center-surround interaction are important processes in identifying specific motion components of objects from background stimuli. However, functional relationship between motion segregation and center-surround interaction is unclear. Here we examined how surrounding motion affects motion segregation in a central field. A vertically moving random-dot pattern was presented in the surround, and in the center two random-dot patterns moving leftward and rightward were superimposed. The surrounding motion perceptually induced a vertical illusory motion in the central field (induced motion). In this stimulus, subjects perceived the central stimulus as moving in a purely vertical direction at shorter stimulus durations. However, at longer stimulus durations (around 300 ms or longer), subjects reported that the central stimulus appeared as transparent patterns moving in two oblique directions. Thus, a summation occurred between a vertical illusory motion and a horizontal physical motion in each central pattern. Durations required for segregating two oblique motion components were much longer than those required for perceiving two horizontal motions without illusory motion. Our results suggest that processes for motion segregation and center-surround interaction can interact under the constraint that computational cost for integration between induced and physical motions is higher than that for motion segregation per se.

4‧Toshio Kubodera
Multimodal integration in perceiving direction of self-motion from real somatic motion and orthogonally directed optic flow pattern

We measured observers' perceived direction of self-motion resulting from the simultaneous presentation of visual and vestibular information, each simulating a different direction of motion. Sakurai et al. (2003) reported that when observers experienced real leftward/rightward somatic motion while viewing a visual expanding/contracting optic flow pattern consistent with forward/backward self-motion, their perceived motion direction was intermediate to those specified by visual and vestibular information. Here, we extend that study to other combinations of real somatic motion and orthogonally directed optic flow patterns, ascertaining observers’ perceived directions as a function of visual amplitude at different vestibular amplitudes. Observers were seated on an oscillating motor-driven swing providing real motion (leftward/rightward, forward/backward), while they viewed a orthogonally directed visual oscillatory optic flow pattern (upward/downward, leftward/rightward, forward/backward) phase-locked to the swing motion. Observers performed a rod-pointing task and reported their perceived direction of self-motion for each combination of visual and vestibular information. Observers’ direction judgments progressively favored the direction specified by vision with larger amplitudes of optic flow. This effect was mediated by swing amplitude such that the effect of vision increased with optic flow amplitude more rapidly for small swing amplitudes than at large swing amplitudes.

5‧ Jy-Chyi Yuan
The Transparency Effect on Plaids Illusion

Plaids motion illusion is relative to motion perception, transparency effect, depth perception (figure-ground separation), and aperture problem. First experiment finds random dot can cancel the aperture problem of the single grating. Second experiment manipulate the gray level of the interposition area of the two gratings composed the plaids. The results show that when the transparency cue of the interposition area is clear, the coherence motion perception of the plaids will be replaced by the component motion perception which means the plaids illusion has been diminished. Third experiment is adding some random dot on the two gratings. The results show that plaids illusion is still being there. Fourth experiment generates random dots plaid pattern stimulus. But when the stimulus is static, subject can not detect where the boundary of the two grating is. This experiment also manipulates the dot density in the interposition area of the unseen two grating. It seems the density is not the influential variable of coherent motion illusion. The whole study tells us, random dot can cancel the aperture problem, but not the plaid illusion. The transparency cue of the interposition area is very important in gray level composed grating, but not in random dots composed plaids.

6‧ Hiroshi Ashida
Asymmetric temporal filtering underlying the ‘Rotating Snakes’ illusion

Illusory motion can be seen in repeated patterns of asymmetric luminance profiles, such as in the ‘Rotating Snakes’ by Kitaoka. While the illusion might occur partly due to latency differences for high and low contrast regions at the onset of the pattern (Backus & Oruç 2005 JoV; Conway et al. 2005 JNS), smooth percept of motion may be better explained by biased responses to retinal motion caused by small eye movements, or by the combination of these effects.
As spatial processing in our visual system is assumed nearly isotropic, asymmetry in temporal processing should be crucial for biased motion responses. We have proposed a schematic model on the basis of gradient computation with distorted temporal derivative filters (Murakami et al. 2006 Vision Res). Recently, Fermüller et al. (2010 Vision Res) proposed a comprehensive model on the basis of conventional motion energy computation (Adelson & Bergen 1985 JOSA). The basic ideas are similar, but there are differences in details. In essence, our model rely on strong second-phase inhibition while Fermüller et al. argue for the reliance on the first positive inputs. Here we review both models and discuss differences in computations for possible reconciliation and extensions.

7‧Makoto Ichikawa
Reduction of the flash-lag effect in terms of active control of visual stimulus and hand movement size

Observer’s active control of the stimulus movement would reduce the flash-lag effect (Ichikawa & Masakura, 2006 Vision Research, in press AP&P). In order to examine the effects of proprioceptive signal on the flash-lag effect, we measured the flash-lag effects for different size conditions for the hand movement. In the active condition, the position of the visual stimulus was controlled by the use of computer-mouse. There were three conditions for the hand movement size on the desk to control the vertical movement of the visual stimulus on the display (30.0, 10, or 2 cm). In the automatic condition, the stimulus moved automatically with a constant velocity that was determined by the mean velocity in the active condition. We found that the flash-lag effect in the active condition might be reduced in any of the hand movement size conditions although the reduction would depend upon the initial condition for the hand movement size for each observer. These results suggest that the proprioceptive signal to control the visual stimulus would be effective in reducing the flash-lag effect regardless of the hand movement size while that cooperation could be violated by novel relationship between the visual information and proprioceptive signal.

8‧ Satoshi Shioiri
Spatiotemporal characteristics of fast and slow motion detectors

Motion signals are analyzed by several motion detectors with different spatial frequency tunings. Recent studies revealed that there are also multiple motion detectors that have different temporal frequency tunings. For example, it has been reported that velocity-selective aftereffect was observed by adapting simultaneously to high spatial- and low spatial-frequency gratings drifting at the same temporal frequency, when tested with a static or a flicker grating [Shioiri and Matsumiya, J Vis, Vol. 9, pp. 1-15., 2009]. Their study suggests that the visual system has at least two different mechanisms for motion processing with different spatiotemporal frequency characteristics. In this study, we estimated the spatiotemporal-frequency tuning of each motion detectors by using a masking technique. The results showed that there are multiple spatial frequency-selective channels for higher temporal frequencies (> 8 Hz), whereas only one channel was found for lower temporal frequencies (one peaking at1 c/deg and 4Hz). The masking effects for the lower temporal frequency tests (< 2 Hz) were always strongest at around 4cpd test at 0.5 Hz, independent of test spatial frequency. The visual system is likely to use different types of motion detectors to analyze fast and slow motion signals.

Form & Shape

Please tick ▼ for abstract content

9‧ Han-Chang Lai
Visual short-term memory for abstract patterns: Comparing a local recognition task and a change-detection task

Previously, we used a local recognition task to assess visual short-term memory for abstract patterns across encoding durations. We found that memory for abstract pattern was more accurate when the pattern was symmetrical, had a higher connectedness, and when the probes located in the upper-left quadrant. However, the local recognition task may bias subjects to adopt a strategy that recalls local parts only. Therefore, the present study adopted a change-detection task that encourages subjects not only memorize local parts but also the whole pattern. We used E-Prime 2.0 to run the experiments on an Acer computer. Each trial began with a fixation cross, then a red sample abstract pattern (within the 5*5 grids) appeared for 300 ms, followed by a 1200 ms blank, and a test pattern that may be identical or different (a red and a white squares were switched) to the sample, appeared until response. Subject’s task was to judge whether the two patterns were the same or not. Our results (N= 19) showed that, as compared to the previous local recognition task, the symmetry advantage in the current change-detection task was augmented while the connectedness effect was decreased. Further experimentation with longer stimulus duration is in progress.

10‧ Jing-Fong Wang
Testing visual short-term memory for abstract patterns in Hepatitis C patients, depressed patients, healthy controls, and college students

The present study adopted a computerized test to explore visual short-term memory performance for Hepatitis C patients, depressed patients, healthy controls, and college students. The patient groups were tested at the clinic while the normal groups were tested in the laboratory. Each trial began with a fixation cross, then a red abstract pattern appeared for 300ms followed by a blank, and a recognition probe appeared until response. Subject’s task was to judge whether the probe overlaps with the red abstract pattern. Our results showed that Hepatitis C patients had the highest mean errors (36.0%), followed by the depressed patients (30.9%), the healthy controls (29.1%), and the college students (19.0%). We also found significant effects for pattern connectedness and symmetry. More recognition errors occurred for the low-connected than those for the high-connected patterns and the magnitudes of differences were similar among groups. Moreover, more errors occurred for the asymmetrical patterns and the differences were dissimilar among groups (college students 9.4% > healthy controls 8.3% > depression 6.1%> Hepatitis C 2.8%). Thus, it suggested that the patients groups (Hepatitis C and depressed patients) seemed to be less able to use symmetry information in a memory task. Further data collection is in progress.


12‧Toshihiro Takahashi
The effect of luminance distribution on the perception of gravitational vertical in pictures

We can perceive the direction of gravitational vertical (up-down) when observing natural pictures. It is obvious that our visual system uses the knowledge and inference regarding the objects for perceiving vertical direction. In addition, our visual system directly would use the factors processed in the early stage, such as color, luminance and edges. In this study, we investigated how such elemental factors, especially the distribution of luminance, affect the perception of gravitational vertical in pictures. Subjects observed stimulus of natural picture with one of four conditions of the presentation time, and responded the direction of gravitational vertical of the stimulus.
Results showed that subjects perceived gravitational vertical depending on the luminance distribution when presentation time was short. This tendency increased as the presentation time decreased. The results also showed that the strength of the effect was dependent on the profile of luminance distribution. These results suggest that if the information produced in the late stage of visual system is not available, the effect of luminance distribution increases. We discuss the relationship between the visual information processed in different level of the system and the perception of gravitational vertical.

13‧ Ping-Fan Chien
Coding of Relative Luminance Change in Macaque Primary Visual Cortex.

Saccade targets in the visual scene can vary widely (10^9) in luminance but our perception largely discounts these differences. While some of the lightness constancy can be explained by changes in pupil size, many of the mechanisms underlying brightness perception are thought to occur in cortex, as early as area V1. Many experiments and computational models are based on V1 encoding of absolute luminance, and the coding of relative luminance change remains unexplored. To characterize absolute versus relative luminance change encoding in macaque primary visual cortex (V1), we mapped the receptive fields of V1 neurons under rapidly changing luminance. We presented m-sequence randomized 32x32 binary patterns at 42Hz and switched them rapidly at 10.5 Hz among 4 average luminance levels. We analyzed the spiking responses via peristimulus time histogram, reverse correlation, and principal component analysis. We find neurons that encode absolute luminance (7/72), relative luminance change (20/72), both (43/72), or neither (2/72). Neurons encoding relative luminance change were primarily located in layer 4. Most of these neurons encoded relative increases in luminance. We surmise that the encoding of relative luminance change by V1 neurons is linked to the mechanism of lightness constancy.

14‧ Yusuke Matsuda
Stimulus factors to decide the perception of order and disorder

When we look at a table with some object on it, we have a feeling of order or disorder depending on the distribution of the objects. However, it is not clearly known what factors in the stimulus affect the perception of disorder. In this study, we investigated the effects of “uniformity” and “entropy” in the stimulus on the perception of disorder of visual pattern.
In the experiments, we used the stimuli consisted of dots. We defined the uniformity of the pattern as the deviation of the dot densities in local regions and the entropy of the pattern as the number of different patterns perceived to be the same. Two patterns having different values of uniformity or entropy were presented at the same time and the subject judged which pattern was more disorderly. Using the data, we represented the relation among the two factors and the magnitude of perceived disorder quantitatively.
The result showed that the perception of disorder increased as the uniformity of dot distribution increased. This result was consistent among subjects. The perception of disorder seems to depend on relatively simple factors in the pattern, such as the uniformity and entropy.

15‧David Rose
Poggendorff illusion with subjective contours

The Poggendorff illusion is a puzzle. Numerous mechanisms have been proposed to account for it, including lateral inhibition in the visual cortex, spatial blurring, assimilation to cardinal axes, apparent depth induction, distance mis-estimation, and extrapolation errors. Several such factors may be operational at once. Replacing luminance edges by subjective contours enables us to exclude some of the proposed accounts. When the Poggendorff's rectangle is demarcated by subjective contours, the illusion remains. Here, we investigate a startling reversal of the normal direction of the illusion that occurs when instead it is the diagonals that are subjective. This occurs whether the rectangle's contours are real or subjective, but only for acute angles of intersection. Eliminating the rectangle altogether reveals a persisting misalignment too, supporting extrapolation error and assimilation to cardinal axes as explanatory factors.

16‧Jong-Tsun Huang
Operating Characteristics of Blind-Spot Completion

Blind-spot completion is a unique process to construct a continuous spatial representation by integrating information over the vicinity area around the blind-spot. Three hypotheses were suggested to explain how this completion could be made possible. The first Ignorance Hypothesis held that the visual system does not process any information in the blind-spot and, henceforth, the completion is achieved by extending the surrounding processing over the squeezed and ignored blind-spot. The second Cognitive Inference Hypothesis stated that the completion is adjusted by higher cognitive and non-visual factors. The third Complementary Hypothesis proposed that one eye's blind-spot in space is covered by the corresponding area in the fellow eye, therefore, the completion can be accomplished by the non-blind-spot corresponding visual system. Three experiments were conducted to differentiate the relative explanatory power of the three hypotheses. Accidentally, these results implied that the blind-spot completion might be accomplished through luminance processing under quick excitatory operation. The area-symmetric operation around the blind-spot seems to be luminance-based.

Depth & 3D Structure

Please tick ▼ for abstract content

17‧ Naoki Nakamura
Dynamic measurement while viewing stereoscopic images of parallel method

Stereoscopic images of parallel method (conflict between accommodative and convergence stimulus, and convergence stimulus further than accommodation stimulus) reduced visual fatigue and caused accommodation relaxation. However, Ukai et al. (2007) described that an imbalance between stimulus to the accommodation and the convergence induces visual fatigue. Kim et al. (2009, Japan Ergonomics Society) reported that the latter was straight when they measured visual fatigue by subjective response.
For this study, we measured accommodative and convergence dynamic responses using a video refraction unit while viewing normal images (non-conflict between accommodative and convergence stimuli) and stereoscopic images of parallel methods.
Results showed accommodative transient overshoot in some subjects. Additionally, we compared the relation between the CA/C ratio and the accommodation response: high-CA/C-ratio subjects have the appearance of overshoot. This result resembles findings obtained using stereoscopic images and the cross method (conflict between accommodative and convergence stimulus, and accommodative stimulus further than convergence stimulus) (Fukushima et al., 2009)
Based on results of the present study, we infer that stress on accommodation and convergence systems caused by viewing stereoscopic images exists whether using the cross method or parallel method.

18‧Kazuho Fukuda
Effect of vertical-size disparity on binocular corresponding points

Vertical-size disparity produces slant perception (Ogle, 1938). In recent studies, it was revealed that internal eye position signal created by vertical-size disparity recalibrates the relationship between horizontal disparity and perceived slant and that the Ogle’s effect was due to miss-application of the recalibration mechanism. There are two possible mechanisms for the recalibration. One is that vertical-size disparity recalibrates the relationship between the output signal of horizontal disparity detectors and the slant perception. The other is that vertical-size disparity recalibrates binocular corresponding points for horizontal disparity detection. Here, we examined whether vertical-size disparity recalibrates binocular corresponding points. The stimuli were composed of horizontal lines to introduce vertical-size disparity and a probe to measure binocular corresponding points. Within a short adaptation to a vertical-size disparity pattern, subject observed an apparent motion created by the probe presented dichoptically. The first stimulus of the probe appeared in one eye and the second appeared in the other eye with variable horizontal positional differences. Subjects judged the direction of the apparent motion. Results showed that there is little effect of vertical-size disparity on the binocular corresponding points, which suggests that the assumption that vertical-size disparity recalibrates corresponding points would not be appropriate for the slant perception.

19‧Masayuki Sato
Depth reversal as a function of disparity-modulation spatial frequency, number of cycles, and modulation amplitude

It is known that perceived depth from binocular disparity is sometimes in the opposite direction to the geometrical prediction (depth reversal). In the present study we measured the apparent depth direction as a function of some spatial factors such as (i) spatial frequency, (ii) number of cycles (i.e. overall stimulus size), and (iii) amplitude of disparity modulation. Binocular disparity introduced into a random-dot stereogram was modulated sinusoidally along a horizontal axis to produce a vertical corrugation in depth. The peak or trough was at the center of the pattern. The observer’s task was to indicate the direction of apparent depth. The spatial frequency of disparity modulation was either 0.1, 0.2, or 0.4 cycles/deg. The number of cycles ranged from 1 to 8. The amplitude of disparity modulation was up to 0.5 deg. The duration of the test stimulus was 0.5 s. All five observers reported reversed depth in the condition with the highest spatial frequency and larger stimulus size. Two observers also reported reversed depth in the condition with the lower spatial frequency and smaller stimulus size. It appears that multiple factors cause depth reversal.

20‧Saori Aida
Magnitude of perceived depth in two stereo-overlapping surfaces is larger than that in three stereo-overlapping surfaces

The geometric relational expression of horizontal disparity, viewing distance, and depth magnitude between objects in stereopsis suggests that magnitude of perceived depth would be the same as long as disparity size and viewing distance are constant. However, we found that this is not necessarily the case, using two and three stereo-overlapping surfaces, in two experiments. In Experiment 1, 10 observers performed a depth reproduction task and showed that the magnitude of reproduced depth between the two stereo-surfaces was larger than that between outer two surfaces of the three stereo- surfaces for each of the disparities used (6.28′, 9.43′, and 18.86′). In Experiment 2, 7 observers performed a depth discrimination task and showed that the perceived depth between two stereo-surfaces was larger than that between outer two surfaces of the three stereo-surfaces for the smallest and largest disparities used (6.28′ and 18.86′) but not for the middle disparity used (9.43′). We discussed these results in terms of “incomplete” disparity averaging.

21‧ Wen-Jing Lin
Influence of regularity of geometric structures and types of object landmarks on wayfinding behavior

People often rely on landmarks to find their ways, but relatively few studies have examined how different types of landmarks influence wayfinding. A previous study from our lab showed that participants had better wayfinding performance in environments with local object landmarks than with global ones when the geometric structure inside the environment was irregular. We speculated that the superior performance with local landmarks could be explained by the fact that global landmarks only provided rough directional guidance and were thus less informative references during wayfinding. In the current study, we adopted a factorial experimental design to further examine how regularity of geometric structures and types of object landmarks conjointly influence wayfinding. The results showed that participants learned to find their ways quickly in both regular and irregular environments with local landmarks, and they kept performing well even when the target was removed from sight. In contrast, for environments with global landmarks, their performance became much worse in the regular than in the irregular environment when the target was invisible. These results suggested that participants referred to both landmarks and geometric structures in the irregular maze, and confirmed that global landmarks offered little help when one attempted to achieve precise wayfinding.

22‧Nobuyuki Tanaka
Comparison of simulator sickness between active and passive observations

The simulator sickness is thought to be caused by the discrepancy between self-motion perceptions induced by the visual and the vestibular systems.In motion sickness, we know empirically a driver hardly gets carsick while a passenger suffers.Both inputs from visual and vestibular systems a passenger receives are supposedly identical to those a driver does.An active observation, such as drivers do, may also reduce the simulator sickness.In this study simulator sickness in active and passive observation conditions are compared by means of subjective evaluations.After the observation of optic-flow for 10 minutes subjects filled out questionnaires consisting of items of the simulator sickness questionnaire and evaluated strength of vection itself and that of the sense of acceleration/deceleration (change in vection).The expanding/contracting optic-flow simulates the motion of dots filling a space when you move forward/backward.In the active condition, a subject presses the keyboard to accelerate or decelerate his/her own motion in response to the color change of the central fixation.In the passive condition, a subject just views the exact same sequence of optic-flow as in the active condition without any active involvement.The results of questionnaire did not show an obvious difference.

23‧ Yoshimura Tatsuya
Enhancement and Inhibition of Vection by Peripheral Optic-flow Pattern

When stationary observers view an optic-flow pattern, they feel as if they move themselves in the opposite direction to the pattern. This perception is called visually induced self-motion perception, or vection. In our previous studies the strength and the duration of vection was examined as a function of field size of optic-flow. As expected, the larger field size of expanding or contracting optic-flow induces the stronger and the longer perception of forward or backward vection. The interaction between two optic-flow patterns presented in central and peripheral fields is examined in the current study. The circular field of 90deg was divided into the central circle and the annular ring. Optic-flow of both fields simulate flying dots of 0.9cm in diameter filling up the space of 20m in depth with a density of 0.125 /m3. The dots in the central field were flying -40km/h (backward) or 40km/h(forward). The velocity and direction of surround dots were -80, -40, -20, 0, 20, 40, or 80km/h. The observer was to keep responding either forward-, backward-, or null-vection while watching the optic-flow. The duration of vection was shorter with the surrounding optic-flow of opposing direction to the center and vice versa.

24‧Kazuya Matsubara
The Effect of Luminance Contrast and Stimulus Distance on the Subjective Depth

Stimuli with higher luminance contrasts tend to be perceived closer in depth. In order to quantify this effect, we measured perceived depth of the test stimulus as a function of luminance contrast, using a depth matching technique. The depth of a reference stimulus was matched to that of the test stimulus by changing the binocular disparity. We also varied the test spatial frequency (from 0.5 to 8 cpd) and the disparity (crossed, uncrossed, or zero). The results showed that for all disparity conditions the test stimulus was perceived much closer with a higher luminance contrast. This suggests that the effect of luminance contrast on depth perception is independent of the binocular disparity. Luminance contrast contributes to a monocular depth process that does not interact with the mechanism of stereopsis. The difference between different spatial frequency conditions suggests that more than one spatial frequency channels contribute to the monocular depth process. Although the effect of luminance contrast was similar for all spatial frequencies, the slopes of the contrast-depth function are steeper with lower spatial frequencies. This is not easily explained by a single mechanism.

Attention III

Please tick ▼ for abstract content

25‧Hiu-Mei Chow
Effects of emotion on attentional blink

Attentional blink (AB), a failure to identify the second target (T2) when it is presented closely after the first target in a rapid sequential visual presentation, was employed as a paradigm to probe on how emotion is in relation to attention. Previous studies found emotional words as T2, as well as an induced sad mood in participants compared to happy mood, would reduce AB in both conditions. Current study investigates whether there is an interaction between the emotion of the participant and the emotion in the stimuli.
Participants were induced with positive or negative mood with music before they completed an AB task with emotional T2 (positive, neutral and negative words). It was found that emotional T2, especially negative words, had an effect of resisting attentional blink, replicating the previous research. The performance in AB between positively and negatively induced mood groups did not differ significantly, nor the congruency effect of the two variables. Yet, the study revealed potential weakness of previous studies of attentional blink task using words. Suggestions were made to improve the experimental design to demonstrate congruency effect in future research.

26‧Shuo-Heng Li
Competition of emotional words for attentional resource

Zeelenberg, Wagenmakers, and Rotteveel (2006) have shown that emotional words (negative and positive) are better identified than neutral words. However, there was no difference in performance between positive and negative words. We hypothesize that this may be due to their using only a single word as target, unlike previous studies showing different performance often presenting multiple words in the display (e.g., visual search or dot-probe task). To test whether words with positive and negative valence have different effects only when they are presented with competitive items, we presented two words (a target word and a competitive word) briefly and the participants had to respond to the target in a two-alternative forced-choice (2AFC) paradigm. Results showed a competitive advantage for positive words, compared to neutral words. Biased competition affects processing of emotional stimuli, causing the difference in performance between conditions when emotional stimuli are presented alone versus under competition.

27‧ Ya-Ling Shih
Can attentional inhibition to emotional stimuli affect emotion vulnerability?

Attentional orienting toward negative information has been shown to increase, but not reduce, emotion vulnerability (EV). With impaired inhibition ability among highly anxious people, we suggested that improving attention control (through attentional inhibition) may be critical to reduce EV. The current study contained two phases: inhibition training (IT) and EV assessment. In the IT were two groups of participants trained to inhibit negative stimulus or neutral stimulus. There were a series of Go/No-Go task in this phase. Participants were presented with a pair of Chinese words (consisting of a negative word and a neutral one), and then one of them changed color to serve as a Go/No-Go cue. Participants responded to the cue location (top or bottom) in the Go cue condition; otherwise, they held their responses. In the EV assessment, participants needed to complete stress task, and their emotion reactivity was recorded before and after stress task. We assess EV both before and after the IT to examine whether attentional inhibition would affect EV. Results are expected that attentional inhibition toward negative stimulus leads to a decrease in emotion vulnerability. Studies focusing on the inhibition to negative stimulus have critical implications regarding clinical therapies based on cognitive change.

28‧Winnie W. L. Chan
Using target-distractor discriminability to examine specific task strategies in Repetition Blindness

Repetition Blindness (RB) refers to a reduced performance in reporting repeated items in a rapid serial visual presentation (RSVP) stream. Whether the size of RB is influenced by using targets and distractors of different categories (e.g., digits versus letters) has been a controversial issue in recent years. In this study, we investigated the effect of categorical discriminability on RB, by asking participants to either report the targets or to detect repetition. In Experiment 1, targets were letters and distractors were digits. Significant RB was found in the detection task, but priming was found in the report task. In Experiment 2, distractors were collections of letter segments, which rendered targets less discriminable from the distractors. RB was still found in the detection task, but unlike Experiment 1 no priming was found in the report task. Comparing the two experiments, categorical discriminability had an effect on performance in the report task, but not in the detection task. This result suggests that different strategies may be involved across tasks, so that target-distractor discriminability influences RB differently in the two tasks.

29‧Apollo M. H. Chu
Repetition blindness with objects having parts deleted and added

Repetition blindness (RB) refers to the finding that participants often fail to report a repeated target in rapid serial visual presentation (RSVP) streams. An earlier study from our lab (Hayward, Zhou, Man, & Harris, 2010) reported that RB for familiar objects is based upon processing of isolated features. The current study investigates whether RB for objects is affected by addition or deletion of object features. In two experiments RB for familiar objects was studied with the first target being a complete object and the second target having some parts removed or extra parts added (that is, if the first target was a dog, the second target might be half the dog or might be a dog with six legs). While the magnitude of RB was found to decrease as a part of the object was removed, it remained invariant as extra parts were added. These results support the contention that processing of independent object components provides the basis for RB.

30‧Melanie Murphy
Nature vs Nurture – The Relationship Between Acute Stress and Vision

Purpose: Although genetic factors are recognised, environmental influences play a notable role in the development of the most prevalent refractive error, myopia. Recent investigations highlight the impact of nearwork and stressful lifestyle as important determinants in the incidence of this condition. The present investigation examined whether acute psychological stress (assessed verbal university presentation) contributes to the myopic shift previously observed to occur over the course of a 3 year degree. Method: First, the relationship between refractive state and demographic characteristics of first and senior year Australian University students was investigated. Then sympathetic nervous system activation (indicated by heart rate, salivary cortisol, systolic and diastolic blood pressure), pachymetry, intraocular pressure (IOP) and refractive state in first year students at baseline and following acute stress was measured. Results: Myopia was significantly greater in the senior cohort and was associated with greater amounts of near work. Levels of myopia significantly increased following acute stress, in conjunction with elevated cortisol levels and an increase in IOP. Conclusion; Acute stressful events at University contribute to a myopic shift in refraction over the course of tertiary study, though whether these shifts are transient or permanent remains to be tested.

31‧ Yuki Miyahara
Synchronism of perceptual reversals involving two horizontally presented ambiguous figures

Previous studies demonstrate that eye movement and visual attention with ambiguous figures are related to perceptual reversals. Our study examines processing of perceptual reversals by presenting two ambiguous figures simultaneously. For our experiments, we chose two figures from ambiguous figures of two types (depth reversals and figure-ground reversals), then situated them horizontally. In Experiment 1, the dispositions and combinations of figures were varied. In Experiment 2, distances between two figures were varied in six steps. Participants evaluated the synchronism of perceptual reversals on a scale of 0–9 in each experiment. Experiment 1 results show that synchronism evaluations of stimuli including reversal figures of two types were less than those of stimuli including figures of the same type. This result suggests that global visual processing cannot be conducted in parallel because processing of reversals has two mutually differing types. Experiment 2 results show that evaluations decreased with increased distance on each combination of figures. This result suggests that participants pay less visual attention to shape perception as stimuli occupy a larger area.

32‧Jhih-Yun Hsiao
Auditory semantic context modulates the conscious perception of bistable figures

We perceive integrated percepts in our daily life by combining the information available in multiple sensory modalities. Here we tested whether the soundtrack that a person hears can influence the perception of bistable figures and whether any such crossmodal influence interacts with attentional control. In Experiment 1, participants
viewed the figure “my wife or my mother-in-law” which typically induced percept that alternates between the old woman and young lady, while listening to soundtrack with either the voice of old woman, young lady, or else without sound. The results showed that participants report a particular percept as being dominant for significantly more of the time when listening to a semantically-congruent auditory soundtrack. However, this effect was eliminated when the participants tried to actively maintain a particular percept. That is, the sound of the young lady could not bias percept toward young lady when participants actively maintained the percept of the old woman. These results imply that sounds may induce an involuntary shift of attention to the percept that is congruent with the sound, or, through voluntarily attentional control, to help maintain the percept once established, and this effect was mitigated when voluntary attention was directed simultaneously with the sound.

33‧Wei-Ming Huang
Test of the automaticity in depression: An example of the Stroop effec

Patients with depression have been found to have specific impairments in the Stroop tasks. Our previous study suggested that the patients with depression were intact in controlled processes in Stroop tasks (Huang, Jingling, & Su, 2010). The goal of this study was to further investigate whether they have deficits in automatic processes. We used the diluted Stroop paradigm, in which ink color and words were separately presented at different locations. The separation should shorten response times for subjects could focus on the ink color more easily and ignore the automatic activations of the words. Thus, the Stroop effect should be smaller. In the experiment, the subjects were instructed to name the ink color, either on a bar (diluted condition) or on a color word (classical condition). Preliminary results with twelve patients and nine healthy controls showed that the reduction of the Stroop effect was observed only in controls but not in depression. In other words, patients still experienced large Stroop effect even when color and words were separated. Our findings thus suggest that the worse Stroop performance in patients with depression might be attributed to their abnormal automatic activations.

34‧Lok-Teng Sio
Attention modulated binocular suppression in non-amblyopic population

We demonstrated that observers with no known history of amblyopia can suppress the input to one of the eyes with attention. The stimuli in each trial contained a Gabor target presented in one of the four possible locations (up, down left and right) in one of the eye and a Gabor distractor at another location either in the same or the different eye of the target. Other possible locations for stimulus presentation in both the left and the right eye images were filled with white noise. A cue at the center of the display indicated the location of the target. The cue was presented to only one eye to control the ocular attention of the observers. The task of the observer was to detect the target. Thus, to perform the task, the observer had to attend to the cued eye. When the target and the cue were presented to the same eye, the target threshold reduced nearly 2-fold (5.2dB) in the dominant eye and more than 3-fold (10.6dB) in the non-dominant eye compared with the condition when the cue and the target were presented to the different eye. Thus binocular suppression can be voluntarily controlled in normal population as well.

35‧Ming Liang
Dual code principle for integration of bottom-up and top-down attentional control

Visual attentional selection includes both bottom-up saliency and top-down guidance, involving multiple visual cortical areas. It is still unresolved how the two types of attentional control are integrated. We propose to extend the dual code principle from the theory of a bottom-up saliency map in V1 to post-V1 processes and top-down control, accordingly input information and attentional control signals are represented in the same neural population at each visual stage. In a prototype model implementing the proposal, the saliency signal is the modulation of neural population responses to input, and is thus transmitted from V1 to V4 along the visual hierarchy. Consequently, the responses to low level image features in V1 and to object shape features in V4 not only code input by the identities of the responding neurons, but also the saliency strength by the response levels. Top-down guidance further modulates the V4 responses, and is thus naturally integrated with bottom-up control. In detecting target objects in cluttered scenes, the model performs better than alternative models having artificial separation and integration of the input representation and attentional control. Information theory measures indicate that the dual code representation significantly improves the discriminability between objects and background clutters by V4 responses.

Computational Issues

Please tick ▼ for abstract content

36‧ Osamu Watanabe
Analytical method for investigating the nonlinearities of observer's judgments with psychophysical reverse correlation

Many physiological researches have utilized the reverse correlation technique to probe receptive field structures of neurons in early visual cortex. Recently, this technique is also employed in psychophysical researches. The psychophysical reverse correlation aims at exploring a whole brain mechanism for an observer's judgment, while physiological researches, in general, investigate a filter property of a single neuron. Brain mechanism for observer's judgments should be made up of many neurons interacted each other and would be a complex nonlinear system. Therefore, the psychophysical reverse correlation requires an analytical method considering the system's nonlinearity. Here, we formulate a general framework to probe the nonlinearities of brain mechanisms with the psychophysical reverse correlation. Similar to physiological researches, the difficulty of the psychophysical reverse correlation analysis is that observer's responses to a stimulus are stochastic and only have discrete values (such as in yes/no and n-AFC tasks). We show that the joint cumulants for stimuli and observer's responses have the information about nonlinear kernels of brain mechanisms and present a general method to extract the kernel information from the cumulants. We can obtain the joint cumulants for stimuli and responses experimentally. We also discuss the application of this analytical method to psychophysical experiments.

37‧Zong-En Yu
A Neural Model for Counting and Subitizing

Counting involves repeatedly identifying items during the search, and incrementing internal representation of number. Therefore the response time are observed to be proportional to the number of items. On the other hand, subitizing for small numbers yields much quicker response than counting. This implicates two distinct mechanisms might take place in turns dealing with different scale of items.
In this research, a neural model describing the possible neural mechanism of subtizing and counting is proposed. The model deals with the interplay between the intraparietal sulcus, area TE, and the prefrontal cortex. It consider subregion LIP in the intraparietal sulcus works as a visual buffer which tagging items been identified and incrimented, subregion hIPS as a functional unit coding semantic representation of numbers, area TE as a pattern recognition unit tuned to the pattern of the input image, and the prefrontal cortex as an association area mediating other representations of numbers, such as visual and auditory representations.
The proposed model is expected to replicate the response of normal subject during a numerosity task. Furthermore, this model describes the counting and subitizing mechanisms of patients with simultagnosia, by blocking the IPS layer in the model.


©2010 Asia-Pacific Conference on Vision