Session Detail

How binocular matching happens on stereoscopic vision with and without 2D-boundary-cues are our basic interesting. In this study we compare two kinds of presentation ways (goggle-glasses & two-mirror-screens) and three kinds of images which have different richness of 2D-boundary-cues (Random-Dots-Stereogram, dot-pattern, real scenes). Two images of left-right eyes superimposed in goggle-glasses presentation, so produces new binocular-disparity boundary cues or ghost. Two-mirror-screen presentation doesn’t have superimposed binocular-disparity generated 2D-boundary-cues, so RDS fusion supposed to be more difficult in this presentation way. Result shows reaction time of 3D perception of RDS will be shorter by goggle-glasses presentation, but not by two-mirror-screen presentation. Dot-pattern and real scenes don’t have this difference. We think it’s because of RDS need to do local matching of zero-crossing first then perceived 3D. With contour pictures (dot-pattern and real scenes) might direct match 2D-boundary-cues globally. So we conclude that stereoscopic fusions are dynamic interaction of global and local matching.

 

We hypothesized that vertical distances are perceived as larger than horizontal because motion on vertical direction acquires more effort because of the gravity. Based on that we asked what would happen if observer would estimate distances by walking toward stimuli. Two experiments were done in a gym, with 31 (14+17) participants matching distances of two stimuli on horizontal and vertical direction. Participants would look at the stimuli (on 1m, 1.5m or 2m), then put a blindfold on eyes, count backwards by three, and walk or climb on leathers for the same distance. In first experiment we measured anisotropy effects and participants walked/climbed on direction opposite to stimuli, while in the second we controlled for the effects of fear from climbing and they walked/climbed on the same direction. If paricipants walked/climbed on opposite direction there were significant effects of direction and distance, and their interaction. Participants were matching shorter vertical distances to longer horizontal ones, which means that vertical distances were perceived as longer. If they walked on same direction only effect of stimuli distance was significant, but no effects of direction. These findings are consistent with the hypothesis on gravity integration into perception action schemes.

 

Although previous studies reported depth adaptation of disparity-defined corrugation involved both phase-dependent and independent processing (Graham and Rogers, 1983; He and Shigemasu, APCV2016), whether there is any shape-level depth adaptation is still unclear. Using dynamic random-dot stereograms in phase randomly-changing adapting condition, we compared the depth aftereffects with the horizontally-orientated sinusoidal corrugations, plaids and noise-shape as adaptation stimuli separately, and horizontally-orientated corrugations as test stimuli. Three adaptors had the same amount of disparity but different distributions. The plaids were composed of horizontally- and vertically-orientated corrugations, which had the same peak-to-trough amplitudes. The dots of noise-shape adaptor distributed in random positions without exact shape structure. In each condition, after adapting to two horizontally-positioned adaptors with larger-smaller or middle-middle amplitudes separately, participants were asked to judge which side of the test stimuli had larger amplitude and PSE was calculated. Results showed significant differences among three conditions. The horizontally-orientated adaptor had the same shape as test stimuli and caused the largest aftereffects. The plaid adaptor caused the middle amount of aftereffects. The noise-shape adaptor had totally different shape and caused the smallest amount of aftereffects. These results might suggest the shape-level depth adaptation, which was independent of disparity-specified depth adaptation.

 

If one eye is patched for a period of 2.5 hours, transient changes in ocular dominance result with the previously patched eye’s contribution being strengthened. Similar changes result from opaque and translucent occlusion suggesting that it is the form information not the luminance information that drives these neuroplastic changes. However, this does not rule out the possibility that interocular luminance imbalances per se cannot produce changes in ocular dominance, indeed based on what we know of the physiology, where the contrast gain of visual neurons is luminance dependent, one would expect it can. We show that if the mean luminance of one eye is reduced 1000 fold for a period of 2.5 hours, there are subsequent changes in ocular dominance and that this critically depends on the absolute luminance of the deprived eye rather than the relative interocular imbalance. We argue that this suggests that the site of action is subcortical, before binocular combination.

 

Horizontal disparity, i.e., horizontal difference between images seen by two eyes, is a cue for depth perception. There are individual differences in the processing of binocular disparity for depth perception even for people who have stereoscopic vision. To investigate the causes for individual differences in stereopsis, we measured stereo acuity with appropriate optical corrections, eye refractive state with no optical corrections and personal refractive state over the years to assess the effects of the latter two factors on stereo acuity. We also measured upper limit to detect disparity for depth perception. Results revealed that stereo acuity for myopic participants was significantly higher than that for emmetropic or hyperopic participants. Participants with a large horizontal astigmatic power tended to have low stereo acuity. Further, upper limit to detect disparity was correlated to stereo acuity. Participants to have high upper limit tended to have low lower limit to detect disparity (high stereo acuity). We discuss these results in terms of individual differences in the quality of retinal image and personal experience of space environment.

 

The perceived depth from disparity in random dot stereogram can be affected by luminance contrast (Chen, Chen & Tyler, 2016). Here, we further investigated whether interocular contrast difference can influence the perceived depth from disparity. The test stimuli were rectangular random dot stereograms (1.27 x 3.44 degree) whose binocular disparities were modulated horizontally to generate the percept of a single cosine surface (0.29 cycle/deg). The maximum test disparity ranged from 0 to 20 arc min while the luminance contrasts of the left- and right-eye image were assigned independently from 5% to 80%. The observers’ task was to adjust the length of a horizontal bar to match the perceived depth difference in the test stimuli. When the interocular contrast ratio (C(High) / C(Low)) was small, the perceived depth increased with the luminance contrast in either eye. However, when the interocular contrast ratio was larger than 4, the perceived depth decreased with the interocular contrast ratio. The effect was the same regardless where the higher contrast was placed in the dominant eye or not. Thus, the perceived depth from disparity depends on not only luminance contrast but also interocular contrast ratio in the stimuli.

 

“Skin transparency” has been used as a popular evaluation term for human facial skin in Japan. However, there is no common definition on “transparency” of skin because it relates to individual personal impressions. In this study, we confirmed the consistency between numerical values of the skin transparency obtained by using a magnitude estimation and a paired comparison methods. In addition, we investigated the relationships between the skin transparency and the image statistical values of the skin. We conducted a pairwise comparison experiment and a subjective evaluation experiment on skin transparency. In the two experiments, participants and stimuli were identical. Eight young students participated and 14 skin patches of women with different ages were used as visual stimuli. As a result, strong correlations between results of the two experiments were shown, suggesting that there is a common and consistent scale of skin transparency and that the subjective evaluation of the skin transparency is an effective evaluation method of the skin appearance. Furthermore, we analyzed the correlations between the skin-transparency and image statistical values. As a result, it was found that the mean luminance value of skin is strongly correlated to the skin transparency.

 

It has been shown that spatial structure of the scene affects lightness of an object. However, most of the previous studies focus on the context of the luminance distribution surrounding the object. Our visual system could estimate illumination of a scene from a surrounding spatial structure based on a correlation between the scene and illumination. For example, an object does not receive light from above when there is something to shield above the object. This study investigated whether lightness is influenced by the spatial structure with no explicit information of luminance. The stimulus consisted of random dots with 3D structure defined by binocular disparity was used to eliminate the influence of luminance and texture. The magnitude of recognized illumination was manipulated by changing the spatial structure of the scene. Observers responded to the lightness of the test patch presented in the stimulus space by adjusting the luminance of the comparison patch placed outside the scene. Results showed that the responded luminance increased when the test patch was interpreted to receive weak illumination. This result suggests that the visual system can infer the illumination of a scene from spatial structure with no information of luminance distribution.

 

The dress photograph, which is posted on the Tumblr in Feb 2015, had produced an interesting phenomenon of visual perception. Some people claim that’s the combination of blue and black color, some people see that’s the combination of white and gold color, and some people’s perception change between the two combination. The purpose of this study is to create a picture, which have the same effect as the dress, can divide people into three types of color combination. In this study, we use two online questionnaires. The first one contains two dress pictures: one is partial, the other is whole; the second one contains 22 blue-black stripped pictures. In the first questionnaire, the result shows that 55 of 268 participants perceive whole dress picture as white and gold color, the rest perceive blue and black. In the white-and-gold groups, 24 of 55 participants perceive white and gold in the partial picture, 22 of 55 participants perceive blue and black. In the second questionnaire, the result shows that 6 of 30 participants only see blue and black in stripped pictures. Moreover, the top five pictures, that are ranked by different groups separated by the perception of partial picture, are the same.

 

We investigated the relationship between word impression and color impression by selecting hue for the word in color paired comparison method, in which two colors were simultaneously presented from 132 combinations and an observer selected the one color closer for the impression of an evaluating word. In the experiment, six trials for each combination were performed and 10 subjects participated in. We selected the evaluating words by referring the results of the factor analysis to the data in the previous literature. 12 surface colors from vivid tone in the PCCS were represented under D65 on a monitor with luminance of 77.2 cd/m2. The results of the selection were converted to z-score and the principal component analysis on the averaged z-score of each color for the words was performed. Consequently, the first and second principal components had 95% cumulative contribution rate and more than 75% of model predicted values by the linear combination of the two components were within the 95% confidence interval. In addition, plots of each color distributed elliptically in the scatter diagrams with axes of the two principal components. Therefore, those 12 colors are handled according to hue impression in expressing word concepts.

 

We conducted an fMRI experiment to clarify which regions in the human brain are involved in lightness perception. The stimulus was a flat surface placed on a checkerboard. The surface was light or dark gray and was or was not shadowed by a cylinder placed behind the viewpoint. The surface also had variations with two levels in form and orientation. The stimuli were presented successively in pair, where the paired stimuli were different in terms of whether the surface was shadowed or not in the half of the pairs. The subjects were asked to report whether the paired surfaces were the same or different in terms of lightness, form, or orientation while the stimuli used in those three tasks were identical. If certain regions are more activated during the lightness discrimination task than another task, the regions would be involved in lightness perception. We found that a region in the vicinity of the calcarine sulci was more activated in the lightness task than in the form task. This result indicates that the low level visual region such as the primary visual cortex is involved in lightness perception.

 

A gray patch surrounded by a white area appears darker than that surrounded by a black area. This “simultaneous lightness contrast” effect is enhanced when luminance gradients from black (outer part) to white (inner part) are added in the outer region of the surrounding white area (Agostini & Galmonte, 2002). This fact shows the effect of luminance gradients on lightness perception. Then, how about chromatic saturation gradients? It is known that simultaneous saturation contrast occurs as well as lightness contrast. In this study, we examined if saturation gradients enhance saturation contrast effect. We employed stimuli containing isoluminant saturation gradients, and asked participants to perform a chromatic saturation matching task using the method of adjustment. The results demonstrated that saturation gradients enhance simultaneous saturation contrast in the same way that luminance gradients do lightness contrast. Although the relationship between the effects of saturation gradients and luminance gradients needs to be clarified, we assume that the results of this study can be attributed to the albedo hypothesis that Agostini and Galmonte (2002) adopted.

 

A retina receives light stimulation and transforms the detected signal into spikes, which are then transmitted to the brain. However, how light stimulation information is conveyed by spike trains is not fully understood. Here, we utilize a multi-electrode array to record firings of the bullfrog’s retinal ganglion cells under a whole field stochastic light stimulation. Our goal is to understand the mutual information shared between stimulus light intensity and the firing rate of RGCs. The result from the stimulation with different correlation time shows that if the correlation time is longer, the retina takes shorter processing time. Furthermore, the retina can even detect the hidden variable in the stochastic stimulation to make prediction. We utilize the time-shifted mutual information to quantify the information encoded in the firing rate and used it as an interpretation of prediction.

 

There is the third class of photoreceptors in addition to cones and rods in primates. The photoreceptive retinal ganglion cells (pRGCs) containing photopigment melanopsin are known to support various non-image forming functions including circadian rhythm and pupillary light reflexes. In physiology several studies have shown that pRGCs have a long latency to light stimuli, much slower than that for cones. The slow response seems ecologically advantageous in terms of irradiance encoding of environmental light. Although the pRGCs in the retina are slow signals originated from pRGC in entire pupillary pathway is unclear. The aim of this study is to clarify a difference in time between cone-mediated fast signals and pRGC-mediated slow signals in the pupillary pathway. We used test stimuli that consist of cone and pRGC stimuli. The cone stimulus modulated cones alone while pRGC stimulus modulated pRGC alone. These two stimuli were summed with various temporal phases in order to estimate an intrinsic temporal difference between cone- and pRGC-mediated signals. The results showed that an amplitude of the pupillary response was largest to the test stimuli in which the phase shift of the pRGC stimulus was advanced by about 1.7 seconds from the cone stimuli.

 

Recent RNAseq analysis has demonstrated bidirectional changes in neurotransmission and metabolism, structural and immune pathways during induction of optically induced myopia. Thus, the aim of this study was to investigate whether similar gene pathways are also related to the excessive axial growth during the induction and recovery from form-deprivation myopia (FDM) in chick. Gene set enrichment analysis (GSEA) software was used to determine enriched pathways in both archived (accession # GSE6543) genomic transcriptome data and new FD recovery data correlated with biometric data. Significant changes in mitochondrial energy metabolism, neurotransmission, G protein coupled receptor signalling and complement cascades were identified during the10 days of induction of profound myopia and found to correlate well with change in axial dimensions. Bile acid and bile salt metabolism pathways (cholesterol/lipid metabolism and sodium channel activation) were significantly upregulated during recovery. The pathways altered during induction of FDM are established indicators of oxidative stress, and are consistent with the choroidal thinning, axial elongation and hyperosmotic ion distribution patterns across the retina and choroid previously reported in FDM.

 

The power-free photovoltaic retinal prosthesis is currently under studied aiming to restore vision of patients with retinitis pigmentosa (RP) and age-related macular degeneration (AMD). The major challenge of the photovoltaic device is the limitation of power efficiency. In collaboration with Prof. Chung-Yu Wu at the National Chiao Tung University, our photovoltaic prosthesis implements an unique divisional power supply scheme (DPSS) system, which provides electricity to only a subset of electrodes at any moment in time with the power generated by all solar cells. This design significantly increases the power efficiency for each electrode, but could potentially reduce the spatiotemporal resolution of retinal prosthesis in human patients. The present study was to systematically characterize the performance of spatiotemporal integration in various DPSS conditions for human subjects using a psychophysical approach. A 16x16 pixels LED array controlled by Arduino was used to simulate the output signal of the DPSS design, and human performance corresponding to different visual stimulations at various update frequencies was used to assess the spatiotemporal resolution of retinal prosthesis. The results showed that the visual angle, pattern complexity and familiarity, division configuration, and contrast polarity significantly influence the optimal update frequency of the DPSS system, while the division number and stimulation order apparently do not affect the performance. These findings provide an insight into the optimization of the photovoltaic retinal prosthesis with the DPSS design, which could be developed into a power-free device able to restore vision in the future.

 

In our previous experiment, we discovered that the slightly increased culture medium concentration as a result of the elevated evaporation rate can significantly promote neurite outgrowth of ganglion cells in mouse retinal explants. To verify this observation, the overall concentration of the culture medium was increased 1.25 and 1.5 folds. The results showed that the 1.25X, but not 1.5X, increased concentration can significantly enhance neurite outgrowth. However, the elevated concentration of culture medium not only increases the relative amount of forskolin but also changes the osmolarity of the medium. This elevated osmotic concentration in the culture medium may stimulate the neurite outgrowth by turning on the stress compensation mechanism. To verify this hypothesis, overall temperature increase from 35 to 38 C for one hour each day (mild stress) and constant exposure at 38 C for five days (strong stress) were used to assess its effect on neurite outgrowth. The results showed that the neurite outgrowth under the mild stress condition is much better than under the control and strong stress conditions. These findings provide an insight into the cellular mechanism of retinal axon regeneration under the stress condition, which could potentially develop into a new method in neural regeneration.

 

Dichoptic stimulation using frequency-tagged flicker has been used as a tool to observe the change of response at the cortical level. Nevertheless, studies of the flickering effect in the inter-ocular competition of the human visual system are limited. This is because even though correlations between the stimulus in one eye affecting the other eye’s network are observed in SSVEP responses, it has been challenging to confirm any causality. The observed amplitude coupled SSVEP response might in fact be a coincidental event due to the lack of time-control. Therefore, we designed a new paradigm for detecting if indeed there is inter-ocular inhibitions by simultaneously combining amplitude modulation (AM) with a sine wave stimuli from the other eye. The amplitude fluctuations from the AM signal provides the time-control manipulation for verification of any interaction between monocular neurons. We found a modulation from AM stimuli inhibiting the 25 Hz sine wave in a phase-dependent manner. Moreover, AM stimulation originating from the dominant or non-dominant eye did not interfere in the modulation of 25 Hz suggesting that AM signal stimuli is eye dominance independent. This is the first time that a tool has been developed for revealing causality in inter-ocular competition studies.

 

After viewing a hue, matches made from memory produce difference-thresholds as a function of wavelength which closely resembled hue difference-thresholds obtained by simultaneous matching. This resemblance persisted despite a negative exponential decay of matching accuracy over time. These results were later replicated, and similar results found to occur with memory difference-thresholds for orientation of single lines and gratings. Since sensory difference-thresholds reflect sensory responses, this similarity suggests that what is retrieved from memory resembles sensory responses. If sensory responses are stored in memory and subject to random decay, can their characteristics endure sufficiently to produce memory difference-thresholds that resemble sensory functions? This question was tested with a neural network model of memory-matching using stored sensory patterns subject to increasing noise being correlated with stimulus patterns that varied from and included the original stimulus. The model demonstrates that difference-threshold functions obtained with noise increasing over time could continue to resemble functions obtained without noise. This suggests that when multiple matches are made from memory to measure difference-thresholds, their averaging reveals sensory activity stored within accumulating noise.

 

In human visual research, measuring the visual response latency by using steady state visually evoked potentials (SSVEPs) to identify which peak of response corresponds to which peak from the stimuli signal is still challenging and inaccurate. To overcome the drawback, this study developed a new method using amplitude-modulated visual stimulation generated by a LED flickering to induce the SSVEP response. To assure the amplitude-modulated waveform is correctly generated and to measure the onset time of the LED, a photodiode was attached with the LED. The advantage of this method is that the envelope of the SSVEP signal can be extracted to then compare its peaks with the corresponding peaks of the envelope of the photodiode signal for gauging the response latency. The results demonstrated that the SSVEP power strongly distributed in the occipital lobe showing a response latency (signal from Oz channel) following single-eye and dual-eye stimulation of approximately 100 ms (range from: 90~120ms). Furthermore, SSVEP responses in anterior channels (e.g.: Cz and Fz channels) had longer latencies (130~160ms) than those in the Oz channel. In addition, no significant difference was observed between the latency from the dominant and non-dominant eye.

 

The oddball paradigm combined with recording of event-related potential (ERP) can elicit the P300 waves. Researchers suggested that P300 reflects attention and memory functions. Although the generator sites are still unclear, the subcortical structures, such as the locus coeruleus and the thalamus, play an important role. The aim of the present study was to investigate the effects of different brain lesions on the P300 component. We recorded the P300 component by using visual and auditory oddball paradigms. Patients with putaman or thalamic stroke were recruited as participants. Healthy young and age-matched participants were also included as control groups. Patients underwent full clinical examination and MRI scan. All participants accepted the Mini-Mental State Examination (MMSE) for evaluation of general cognitive functions. Decreased visual P300 peak amplitude also be found in patient groups as well as in age-matched group suggesting an aging effect. Compared to young and age-matched healthy controls, patients with putamen or thalamus stroke showed delayed auditory P300 peak latency. It suggests the impairment of the auditory P300 may stem from lesions at various location and different lesion type. Our study illustrates the important role of corticolimbic structures in the generation of the P300 potentials.

 

The human visual system processes different attributes of an object separately and then integrates them to elicit a specific response. Based on the paired-attribute model in which bound feature pairs are units of multi-attribute stimulus-response association, our previous study suggested that learning the association of a color-location pair or a shape-location pair with a response is more difficult than that of a color-shape pair. Present study investigated on which pairs the association of a stimulus comprising color, shape, and location with a response depended upon. Participants learned mapping of eight items comprising color (red/green), shape (circle/triangle), and location (left/right) to four responses. Thereafter, they were tested with transfer blocks with a pattern comprising a colored shape and a colored switch indicating left or right (Condition of two pairs), a pattern comprising a colored shape and a gray switch (Condition of a color-shape pair and a location singleton), and a pattern comprising a gray shape and a colored switch (Condition of a color-location pair and a shape singleton). The result revealed that correct response rate dropped only for color-location pair and shape singleton conditions, suggesting that color-shape pair is more important in stimulus-response association than the color-location pair.

 

Increasing evidence has shown that the ability of statistical learning (SL) contributes to the acquisition and processing of individual words in people’s native and foreign languages. However, whether such ability contributes to sentence reading is less explored. The present study examined whether individual difference in SL, as indexed by participants’ sensitivity to nonadjacent visual shapes in sequence, correlates with their ability in reading sentences with relative clauses (RCs). Previous research on Chinese sentence processing has established that sentences with subjective RCs (SRCs) are more difficult to read and to comprehend than those with objective RCs (ORCs). The present results replicated the SRC disadvantage in a self-paced reading task, and further showed that individual difference in the magnitude of such disadvantage, as measured by online reading time especially at the head noun position, significantly correlated with participants’ performance in the SL task. On the other hand, individual difference in the magnitude of SRC disadvantage did not correlate with participants’ IQ or working memory. These findings suggest that SL of dependency between nonadjacent visual shapes might serve as the underlying mechanism to support acquisition and processing of linguistic structures in written sentences.

 

According to the film editor, Walter Murch, the viewer is compelled to re-evaluate the novel scene context following a scene cut when the scene change is large. However, when the scene change is small, viewers may notice the scene change, but the novel scene may not be different enough to compel viewers to re-evaluate the scene context. Yet, the failure to re-evaluate the scene context could also be due to the viewer not being aware of a scene change. In this study, we conducted two experiments to 1) examine how the magnitude of scene change influences the re-evaluation of the novel scene, indicated by the initial centering response (ICR), and 2) examine how the ICR relates to viewers’ conscious awareness of a scene change. In Experiment 1, pairs of images were presented for several seconds, one after the other, separated by a brief mask (15ms), as image difference was varied. We found that the ICR increased with the magnitude of scene change in the image pairs. In Experiment 2, we further asked participants to respond to the scene change. Our results suggest that viewers were aware of a small scene change even if they did not re-evaluate the novel scene.