張貼日期:Sep 09, 2014 9:31:25 AM
Keynote speech
Lecture Hall
Friday, 26th September 2014
09:15 - 10:05
A First Tutorial on Static General Recognition Theory (STGRT)
James T. Townsend, Indiana University
10:25 - 11:15
Identification of Mental Architectures in Face Perception Using the Systems Factorial Technology
Mario Fific, Grand Valley State University
Failure to selectively attend to a facial feature, in the part-to-whole paradigm, has been taken as evidence of holistic perception in a large body of face perception literature. In this study we demonstrate that although failure of selective attention is a necessary property of holistic perception, its presence alone is not sufficient to conclude that holistic processing has occurred. One must also consider the cognitive properties that are a natural part of information-processing systems, namely, a processing order (serial, parallel), a stopping rule (self-terminating, exhaustive), and process dependency. We demonstrate that an analytic model (nonholistic) based on a parallel mental architecture and a self-terminating stopping rule can predict failure of selective attention to a facial feature. The new insights in our approach are based on systems factorial technology (SFT), which provides a rigorous means of identifying the holistic/analytic distinction. The main goal of the study was to compare potential changes in architecture when two second-order relational facial features are manipulated across different face contexts. Supported by simulation data, we suggest that the critical concept for modeling holistic perception is the interactive dependency between features. We argue that without conducting tests for architecture, stopping rule, and dependency, apparent holism could be confounded with analytic perception. A brief tutorial about how to use SFT in face perception experiments is provided.
11:15 - 12:05
How do information processing systems deal with conflicting information?
Daniel R. Little, The University of Melbourne
In this presentation, I will discuss how different information processing architectures deal with incongruent or conflicting information. For instance, a robust empirical finding is that RTs are slower when dealing with incongruent sources of information (e.g., determining whether a whale is a fish or a mammal? A whale has biological properties which make it a mammal but lives in an environment typically populated by fish) than when dealing with congruent sources of information (e.g., determining whether a trout is a mammal or a fish). I will argue that the effect of conflicting information depends on the processing architectures and derive a new measure of information processing called the conflict contrast function by drawing an analogy to the capacity of an information processing system. By varying the salience of the conflicting information source, I will show that serial, parallel and coactive information processing architectures predict qualitatively different conflict contrast functions. I will show results from several experiments and show that this novel measure provides convergent evidence to other measures based on the Systems Factorial Technology (Townsend & Nozawa, 1995; e.g., the Mean Interaction Contrast and the Survivor Interaction Contrast) about the underlying processing architecture.
13:30 - 14:20
Context effects in spatial vision: Six dimensions of configuration dependence
Chien-Chung Chen, National Taiwan University
The visual performance to a visual target can be changed by the presence of another visual stimulus (Flanker) nearby. Many theories have been proposed to explain the flanker effects. We used a dual-task paradigm, in which an observer is to detect a target superimposed on a pedestal and in the presence of flankers, to investigate the mechanisms underlying such flanker effects. The typical result is that, compared with the no-flanker conditions, the target threshold vs. pedestal contrast (TvC) functions shifted horizontally on log-log coordinates when the flanker appears. This result can be explained by a model in which the effect of the flankers is to modulate both the excitatory and the divisive inhibitory sensitivity of the target mechanism to both the target and the pedestal. The flanker effect depends on spatial configuration of stimuli, such as the relative distance, location, orientation, phase, and depth between the target and the flankers. Such spatial configuration dependence suggests that the flanker effect is close related to border ownership coding and contour and surface completion.
14:40 - 15:30
Fast categorization of objects in natural scenes : the first 200 ms of visual processing
Michèle Fabre Thorpe, University of Toulouse
The efficiency of the primate visual system at recognizing hundreds of thousands of objects in complex and rich environments is astonishing. When required to categorize animals, faces, vehicles… in natural photographs flashed for only 20ms, humans and monkeys are fast and very accurate even in extreme conditions of presentation (lack of color, low contrasts, strong masking…).
Their fastest behavioral responses show that the visual processing can be completed in 100-150 ms. The conference will describe the temporal course of object categorization at various hierarchical levels (animals/birds; vehicles/cars), and illustrate two main characteristics of early visual processing : its massive parallelism and its surprising automaticity. Objects can be categorized up to the basic level (car, dog, bird…) with no focused attention.
When natural photographs are used as stimuli, objects are embedded in their natural settings. It might be argued that fast categorization relies on fast extraction of image global statistics rather than on true object processing. The categorization of scenes with controlled spectral signatures shows that primates (both humans and monkeys) rely on object processing whereas models based on spectral signature perform at chance. On the other hand, early and reliable interference in the processing of objects and their contexts of presentation can also be demonstrated.
All data support the efficiency of the biological visual system, its ability to shape itself by experience and to extract –implicitly- the regularities of the surrounding visual world.
15:30 - 16:20
How can we recognize sensory stimuli that we have not experienced for several decades?
Simon J. Thorpe, University of Toulouse
For over 25 years, I have been arguing that the speed of processing in the human visual system puts major constraints on the underlying computational mechanisms. Behavioral and electrophysiological data show that at least some biologically important stimuli (including animals and faces) can be identified and localized in just 100 ms. This is particularly impressive given the limitations of the underlying neural hardware, and suggests that a great deal of processing can be achieved with a single feed-forward sweep though the many successive stages of the visual system. For a long time, it was thought that the ability of the human visual system to process complex natural scenes would remain forever beyond the capacity of artificial vision systems, but it is now clear that the state of the art in computer vision is starting to catch up. Interestingly, the best artificial systems use processing architectures built on simple feed-forward mechanisms that look remarkably similar to those used in the primate visual system. However, the procedures used for training these artificial systems are very different to the mechanisms used in biological vision. In this talk I discuss the possibility that spike-based processing and learning mechanisms may allow future models to combine the remarkable efficiency of the latest computer vision systems with the flexible and rapid learning seen in human scene processing.
Sunday, 28th September 2014
9:00 - 9:50
Quantifying Configural Superiority with the Capacity Coefficient
Joseph W. Houpt, Wright State University
Configural superiority effects are an important component of our understanding of visual perception of many types of stimuli. We propose the capacity coefficient as common framework for measuring configural superiority across a wide range of stimulus types. This measure has a number of advantages. The coefficient is based on a comparison of responses to the configuration with a baseline of unlimited-capacity, independent, parallel processing of each of the parts. Response times for processing the parts in isolation are used to estimate that baseline performance. Better than baseline performance, or better than unlimited-capacity, independent parallel processing, of a configuration of parts, indicates configural superiority. Furthermore, because the capacity coefficient accounts for the difficulty of processing each part, the capacity coefficient for one type of configuration can be compared to the capacity coefficient of another configuration, even if the parts are not exactly the same. We applied the capacity coefficient to three domains in which configural superiority effects have been previously demonstrated: the orientation of a pair of dots, words, and faces. We found that participants had better than baseline performance for detecting differences in the location of dots relative to reference points if there was also a difference in the orientation. The capacity coefficient was much higher than when there was not a difference in orientation; in fact, when there was no difference in orientation, the capacity coefficient indicated worse than baseline performance. Likewise, we found that participants performed better than baseline with words. Participants’ capacity coefficients were higher for words than random consonant sequences, which tended to have equal to or worse than baseline capacity coefficients. Finally, we found that participants had better than baseline performance with aligned upper and lower face halves, but lower capacity coefficients with misaligned face halves, usually below baseline.
10:20 - 11:10
The Modal Research Design for SFT Application in Face Perception: Conjunctive/Disjunctive-rule Stimulus Structures (AND/OR) and a Novel Signature of Holistic Face Perception: “Snake Wiggle”
Mario Fific, Grand Valley State University
In the research practice in general, analytic and holistic perception suffered from not being operationally defined by the cognitive processes but by experimental procedures. We operationally define holistic processing by the coactive processing model, which is defined to have the following properties: parallel, exhaustive and process-dependent perceptual strategy. The novel diagnostic assessment is provided to learn about the underlying cognitive processes that are engaged in holistic perception of faces: systems factorial technology (SFT) is a suite of tools designed to analyze the organization of mental processes underlying cognitive tasks. The SFT tools provide information about whether processes are organized in serial or parallel, whether the processing could self-terminate or be exhaustive, and whether the processes are independent or dependent. Within SFT the coactive processing is identified by the presence of a unique signature obtained from analyzing reaction time distribution functions.
The SFT approach could be applied on (facial) features varying on continuous stimulus dimensions. An example of continuous stimulus dimensions could be configural second-order facial features, such as eye separation, or nose position, among many others. In this study the two proposed stimulus structures were used for a research design: the disjunctive and conjunctive-rule stimulus structures, so called OR and AND designs. The factorially designed disjunctive and conjunctive stimulus dimensions (OR and AND) permits SFT tests for processing order, stopping rule, and process dependency, and thus provides a rigorous means of identifying the holistic/analytic distinction. The OR design is analytic-face friendly as it allows subjects to use the terminating stopping rule and to recognize faces only on individual facial features. The AND design is holistic-face friendly as it forces subjects to use the exhaustive stopping rule and recognize whole faces. A brief tutorial about how to construct such designs using SFT is provided. The individual subject results, from a face study, showed that subjects would prefer analytic models of face perception in the OR design, in contrast to more holistic face perception in the AND design.
We also report a new diagnostic signature of holistic visual perception of faces which we call a “snake wiggle”. The signature is based on the dynamic change of the shape of RT distributions across multiple trials during which subjects learned to classify a set of faces. The present study shows how the coactive signature evolves during the learning phase of the experiment: from the analytic signatures to the coactive one. The animation of the evolution process reveals a dynamically changing signature appearing as a snake wiggle. The “snake wiggle” signature could be easily generalized for similar phenomena (gestalt, configural perception, object perception etc), and across practically all stimulus dimensions.
11:10 - 12:00
An Introduction to Response Time General Recognition Theory (RTGRT)
James T. Townsend, Indiana University
Workshop
Lecture Hall
Friday, 27th September 2014
9:00 - 10:20
Systems Factorial Technology with R
Joseph W. Houpt, Wright State University
In this workshop, l will focus on the powerful mathematical modeling framework Systems Factorial Technology (SFT), developed to assess broad classes of information processing models engaged in any cognition or perception task requiring decisions about two or more sources of information. I will review the basic theory of SFT, including the nonparametric measures for response time data, and illustrate their implementation in a new R package (http://cran.r-project.org/web/packages/sft). Participants will learn to design a Double Factorial Paradigm experiment, which enables the easy application of all SFT measures within a single empirical study. Participants will try SFT analysis of data from a simple visual detection task using the sft package, and, time permitting, will conclude the tutorial with sample designs for experiments in their own research domains.
10:40 - 12:00
Tools for simulating from parametric models of information processing architecture in [R]
Daniel R. Little, The University of Melbourne, Australia
In this workshop, I will demonstrate how to use [R] to generate the predictions of the logical rule models of categorization (Fific, Little & Nosofsky, 2010; Little, 2012). Here we will use a General Recognition Theory and Decision Boundary Theory (Ashby & Townsend, 1986; Ashby & Gott, 1991) “front-end” to generate drift-rates, which we will use to generate response time predictions from the Linear Ballistic Accumulator (Brown & Heathcote, 2008). No knowledge response time modelling is required but some understanding of how to manipulate variables and functions in [R] would be useful.
Readings: Fific, Little & Nosofsky (2010); Brown & Heathcote (2008).
Oral Presentation
Lecture Hall
13:30-13:50 Saturday, 27th September 2014
1. Conditions under which affine representations are also Fechnerian under the power law of similarity on the Weber sensitivities
Yung-Fong Hsu, Department of Psychology, National Taiwan University
In a recent paper, Hsu, Iverson, and Doble (2010) examined some properties of a (weakly balanced) affine representation for choices, showing that using the Fechner method of integrating jnds, one can reconstruct the scales in the representation from the behavior of (Weber) sensitivities $\xi_{s}(x)$ in a neighborhood of s=0. In this paper we impose a power law of similarity on the sensitivities and study its impact on the affine representation. Especially, we specify the conditions for the first- and second-order derivatives of $\xi_s(x)$ with respect to s (and evaluated at s=0) under which the affine representation is degenerate to a Fechnerian one. We also link the results to the solutions in Iverson (2006) that were solved for the more general law of similarity within the Fechnerian framework.
2. Temporal characteristics of the collinear masking effect in visual search
Li Jingling & Ching-Wen Chiu, Graduate Institute of Neural and Cognitive Sciences, China Medical University
A salient target is usually easier to find in visual search. However, perceptual grouping, collinearity in particular, can reverse this observation. The phenomenon, called the collinear masking effect, was observed for a search display that was filled with horizontal bars while a column of bars were vertical (the collinear column). This collinear column thus is salient and well-grouped. The task was to discriminate a small oriented bar either in the collinear column or in other columns in the background. The collinear masking effect refers to the phenomenon that discrimination was slower for a target on the collinear column comparing to that in the background. Since feature saliency affect visual search in relatively early (e.g., 40 - 70 ms) while feature conjunctions late (e.g., 150 -300ms) in information process stages, we examined the time course of the collinear masking effect in this study. With limited presentation durations (40, 70, 150, or 300 ms), different masks lead to different results. Nevertheless, the collinear masking effect was observed at the time point when the target discrimination was above chance level. Further experiment showed that if the search display was presented preceded the target, reliable collinear masking effect was obtained as early as 40 ms. Our results imply that the collinear masking effect occurred as fast as saliency was calculated, suggesting that collinear grouping interferes with visual search very early in the information processing stream.
3. Specialization and generalization in developing the other-race effect
Sarina Hui-Lin Chien1 & Tsung-Ren Huang2 & Jing-Fong Wang1, 1Institute of Neural & Cognitive Sciences, China Medical University & 2Department of Psychology, National Taiwan University
It has been a longstanding debate whether perceptual development is primarily a pruning or branching process. While the developmental literature on perceptual narrowing emphasizes declined sensitivities to rarely exposed stimuli, research of perceptual learning highlights improved sensitivities to frequently exposed stimuli. By examining infants' developmental changes in processing own- and other-race faces, studies of the other-race effect provide a unique window into understanding the dominant form of experience-shaped neuroplasticity and have favored the pruning view (i.e. perceptual narrowing). However, here we reveal the branching aspect of perceptual development (i.e. perceptual learning). Using the familiarization paradigm, we investigated 4-, 6- and 9-month-old Taiwanese infants perceptual discriminability with oval-masked faces of three ethnic groups (Taiwanese, Caucasian, Philippine) and each with three levels of discrimination difficulty. Our results showed that the infants exhibited an own-race advantage at 4 months and improved rather than declined discriminability for other-race faces between 4 and 9 months. This race-generalized improvement of face recognition implies a more specialized neural representation of the facial features shared by own- and other-race faces. Our study and other studies that used unmasked faces complementarily suggest that, during perceptual development, the neural representations of familiar and unfamiliar perceptual features may compete and undergo branching and pruning processes, respectively.
*This project was supported by Taiwanese Ministry of Science and Technology Grant: NSC 99-2410-039-003-MY3 to Dr. S. H. L. Chien.
4. Male observers perceived female facial parts more attractive than the whole face
Chihiro Saegusa & Katsumi Watanabe, Research Center of Advanced Science and Technology, The University of Tokyo
To investigate potential differences in the use of configural and parts-based information for facial attractiveness judgment, in this study, we examined how the attractiveness rating of female facial parts (eyes, mouth and nose) and whole face would differ between male and female observers. Observers (16 males and 20 females) saw the eyes, mouth, nose and whole face of 58 Asian female facial photographs in 4 separate sessions and evaluated how attractive they were with a 7-point scale with no time constraint. The results showed that, with the present set of female photographs, the male observers evaluated the whole faces as less attractive than the female observers, while there was no difference in attractiveness evaluation for each facial part. In order to examine how the attractiveness rating for each facial part would predict the attractiveness rating for the whole faces, we performed multiple regression analyses. There was no significant difference in coefficients of each facial part between the female and male observers, indicating the female and male observers combined the information from each facial part for attractiveness rating for the whole faces in a similar way. However, there was a significant difference in the constant between male and female observers, with constantly lower attractiveness ratings made by the male observers. These findings suggested that (a) the influence from each facial part on the attractiveness judgment of the whole faces was similar between female and male observers, and (b) the attractiveness ratings for the whole faces made by male observers might be influenced by facial configuration, skin texture, etc. Although the results may be confined to the present set of photographs and observers, the present findings imply the possible gender difference in contribution of configural and parts-based information to facial attractiveness judgments.
5. Order Effect and Time Varying Category
Lee-Xieng Yang & Hao-Ting Wang, Department of Psychology, National Chengchi University
The main purpose of this study is to examine how people learn time-varying categories as well as whether order effect exists in such learning. To this end, we design three types of category structures, in which the stimuli vary along trials in an ascending, descending, and quadratic trend. Also, tendency to repeat preceding category label as current response is regarded as evidence for order effect. The results show a clear order effect in these experiments. The modeling results reveal that GCM, which is modified to be sensitive to trial order and SDGCM, which relies on the similarity and dissimilarity to the exemplars for categorization, provide a good fit for all experiments. However, the rule-based model used by Navarro, Perfors, and Vong (2013), which changes the boundary trial by trial has difficulty accommodating the learning pattern in quadratic trend.
Poster Session
Lecture Hallway
15:30-17:00 Saturday, 27th September 2014
(1) Observing cognitive processes in solving geometry problems requiring rotations with eye tracker and handwriting devices
John J. H. Lin & Sunny S. J. Lin, Institute of Education, National Chiao Tung University
This study used eye-tracking approach to investigate problem solvers’ online processes while solving geometry problems with handwriting devices. We are interested in the reasons that lead to increasing difficulties in solving a geometry problem with a pair of triangles that requires rotations in a geometry problem solving scenario using eye tracker. In addition, we investigated whether increase the distance between the pair of triangles that require rotation would reduce difficulties. Solvers’ outcomes, cognitive load, and eye movement were integrated as a probe to investigate the source of difficulties when they fail to solve the problems. Sixty-two high school students were required to complete a pretest, solve problems, and write a questionnaire regarding source of difficulties, as well as cognitive load in a classroom setting. Incrementally providing information makes it possible to observe the diagram-comprehending processes with high resolution. The results indicated rotations with specific angles might be the source of difficulties while a similar problem is solved. Separate triangles could not increase pass rates but could reduce cognitive load. The result of eye movement analyses revealed that solvers’ fixated at the diagrams longer while the triangles were separately. The analysis of the eye transition patterns on specific interest of areas suggested that no significant difference between successful and unsuccessful solvers.
(2) Representational Change Theory: An eye movement study of matchstick arithmetic insight problem
Yu Chen Lin & Ya Lun Liang & Yassiri Y Cruz, Department of Psychology, National Cheng Kung University
Representational Change Theory and Dual Factors Theory interpret insight problem solving process differently (Tseng, Chen, Chen, Sung, & Chang, 2014). When using matched arithmetic problem as experiment material, the former emphasizes representation only, and claims that every participant falls into inappropriate representations in the prior period of problem solving. The latter holds that the averting fixedness is also important. Thus, due to averting fixedness in the prior period of problem solving, the successful participants can do better in changing representation later.
However, since the previous experiment of Dual Factors Theory has unequal sample size between groups (12 and 24) (Tseng et al., 2014), it’s doubtful whether the significant difference is actually evidence of averting fixation of the successful group. We assume, since it’s the first time encountering insight problems, both groups will have similar patterns. The successful group will not avert fixation in the prior period.
The result shows that both groups have similar patterns in the prior period and their percentage of fixation duration on fixation region (the left side |||) are alike. Though not averting fixation in the prior period, the successful group starts to find possible answer with larger variance on fixation region in the middle period and increases fixation duration on the key region in the final period. The results obtained supports the Representational Change Theory more than Dual Factors Theory, for the two groups show no difference in fixation region in the prior period but the successful group finds the solution at the end.
(3) Inverted faces do not lead to the gender-
category dependency on the sequential judgment
Aki Kondo & Kohske Takahashi & Katsumi Watanabe, Research Center for Advanced Science and Technology, The University of Tokyo
The preceding decision and the preceding item bias one-by-one decision making for sequentially presented items. This phenomenon is known as the sequential effect (Holland & Lockhead, 1968). Previously, we showed that sequential face attractiveness judgments were also biased toward those in the preceding trials (Kondo, Takahashi, & Watanabe, 2012). Furthermore, the magnitude of the bias was smaller when the genders of the faces in the current and the preceding trials were different (between-gender condition) than when they were same (within-gender condition) (Kondo, Takahashi, & Watanabe, 2013). In the present study, we examined whether the smaller bias in the between-gender condition would be due to the lower perceptual similarities (i.e., distinguishability) between male and female faces and would interact with holistic (configural) processing of faces. For this purpose, we presented the upright and inverted images of the faces. In Experiment 1, forty-eight pictures of male and female faces were presented in a random sequence. Participants were asked to rate attractiveness of each face on a 7-point scale. The upright and inverted faces were presented in separate sessions. We found that the sequential bias of the upright faces was weaker in the between-gender condition than in the within-gender conditions, whereas they were comparable when the faces were inverted. In Experiment 2, the participant’s task was to indicate the gender of each face by pressing keys as quickly as possible. We found that the gender judgment was quicker for the upright faces than for the inverted faces. However, the correlation analysis revealed that the distinguishability of gender (i.e., the reaction time of each face in Experiment 2) could not be related to the magnitude of sequential bias in Experiment 1 (upright, within, r = .02, t(46) = 0.14; upright, between, r = -.18, t(46) = -1.22; inverted, within, r = -.08, t(46) = -0.52; inverted, between, r = .06, t(46) = 0.43, all ps > 0.2). Thus, perceptual (dis)similarity and possible difference in holistic/configural processing could not account for the weaker sequential bias between female and male faces. Taken together, these results implied that (1) the decision for the faces is biased when the preceding face belongs to the same gender-category, (2) the gender-category is automatically considered when the participants perform attractive judgment for the upright faces, and (3) the gender-category has no effect when the faces are presented in the inverted orientation. In the sequential effects, the preceding events affect a point of reference for the current judgments (Holland & Lockhead, 1968; Stewart, Brown, & Chater, 2005). Our finding points to the possibility that the reference for male faces and that for female faces are (at least partly) separate when observers view the faces in the upright orientation, while the references for inverted faces are inseparable among different gender categories. The present study would provide an important implication regarding the interaction of configural process (i.e., holistic face perception) and subjective decision making (i.e., face attractiveness judgments).
(4) Individual Difference on Eye Movements for Re-examining the Cognitive Processes in a Pictorial Matchstick Arithmetic Insight Problem
Jon-Fan Hu1 & Ming-Liang Wei2 & Yu Chen Lin1 & Ya Lun Liang1 & YassiriY Cruz1, 1Department of Psychology, National Cheng Kung University & 2Department of Electrical Engineering, National Cheng Kung University
Representational Change Theory and Dual Factors Theory interpret the success/failure for solving visually drawing insight problem processes differently (Tseng, Chen, Chen, Sung, & Chang, 2014). When using a pictorially presented matched arithmetic problem as experimental material, the former emphasizes representation only, and claims that participants’ failure is due to inappropriate representations in the prior period time while solving similar problems. The latter holds that the averting fixedness is a more important factor deciding the success or failure. However, since the latest experiment supporting Dual Factors Theory has unequal sample size between groups (12 and 24) (Tseng et al., 2014), it is doubtful whether the significant difference reported is actually convincing for that theory. In contrast, we assume, during the earlier time encountering insight problems, both groups should have similar patterns. Afterwards, successful group would have more active processes in searching for possible solutions than unsuccessful group, which would lead to the final performance at the end. In the present study, participants were invited and completed the eye-tracking study with the same design as Tseng et al. (2014). Almost the same number of successful and unsuccessful participants for solving the problem, the results show that both groups indeed have similar patterns in the prior period and their percentage of fixation duration on fixation region (the left side |||) are alike. Though not averting fixation in the prior period, the successful group starts to find possible answer with larger variance on fixation region in the middle period and increases fixation duration on the key region in the final period. The findings tend to support the Representational Change Theory more than the Dual Factors Theory, for the two groups show no difference in fixation region in the prior period but one group finds the solution at the end.
(5) Comparisons between Component, Configural, and Composite Face Processing: Testing the Utility of a Taiwanese Face Database
Vicky Ying-Hsi Chen & Gary C.-W. Shyi, Department of Psychology and Center for Research in Cognitive Science, National Chung Cheng University
The utility of creating an ethnically-specific and culturally-endowed face database lies not so much in the fact that a large number of face images has been collected and normed in accordance with acceptable standardized procedure. Rather, the utility lies in how to put those face images into use in tasks that presumably would invoke different kinds of processing, configural and holistic processing in particular, in widely used face-processing tasks. Here we present results from a study where tasks tapping, respectively, component, configural and holistic processing of faces selected from a Taiwanese face database (Shyi, Huang & Yeh, 2013). Three tasks were used: For the component task, pairwise face images with altered components (eyes or mouth) were presented and participants were asked to judge the identity between the two faces. For the configural task, distances between the eyes and between the eyes and mouth were simultaneously altered outwardly or inwardly by 4% or 8%. Finally, for the composite task, composite faces were created by joining the top and bottom parts of difference faces to create the appearance of entirely novel faces. In each trial for all three tasks, two faces were presented simultaneously but displaced vertically to reduce point-by-point image comparison, and participants were asked to decide, whether or not the two faces were the same or different for both the component and configural tasks, or whether or not the top parts of displayed faces were exactly the same for the composite task. It should be noted that a complete design was used for the composite task (Gauthier & Bukach, 2007), where results from both complete and partial design can be evaluated.
Using d’, the results of the present study, while largely replicating many findings existed in the literature, also revealed a number of notable differences: First, for the configural task, reliable inversion effect was found with 8% alteration, but was absent with 4% alteration, replicating Yang’s (2010) parametric study on configural processing. For the composite task, the congruence effect—as an index of holistic processing based on complete design—was larger for composite faces that were aligned than those that were misaligned, replicating many of the findings reported by Gauthier and her colleagues. Finally, somewhat unexpectedly, not only reliable inversion effect was found with the component task, but also its magnitude was greater than that for the 8% configural task, which may have been due to the shape change involved in alteration of components (McKone & Yovel, 2009). In summary, results of the present study show that the Taiwanese face database may be a useful source of stimuli for a variety of tasks assessing different aspects of face processing.
(6) 3D Surface configuration effect in Glass Pattern perception
Pei-Yin Chen & Chien-Chung Chen , Department of Psychology, National Taiwan University
We investigated how 3D surface configuration affects 2D Glass pattern perception. Glass patterns consist of randomly distributed dot pairs, or dipoles, whose orientations follow a geometric transform. The stimuli were concentric, radial or spiral Glass patterns (2.5deg radius) consisted of dots (2.3’ x 2.3’) with 4% density. Two kinds of 3D surface modulation were achieved by manipulating binocular disparity of dots: 1) slanted, where the depth changed linearly from one side to the other, and 2) concave/convex, where the depth changed with a projectile centered at the fixation.
We measured the coherence threshold for detecting the Glass patterns on these surfaces at 75% accuracy with a 2AFC paradigm. In the slanted conditions, the coherence threshold was always the same as that measured on the frontoparallel plane regardless the slant of the surface. In the concave/convex conditions, however, the threshold increased with surface curvature linearly on log-log coordinates. Our result suggests that the Glass pattern detection is independent from surface slant and thus may have an underlying 3D representation. In addition, such 3D representation is linear rather than quadratic in space.
(7) Taiwanese Face Memory Test (TFMT): A Preliminary Investigation
Ya-Hsin Cheng & Kuan-Hao Cheng & Gary C.-W Shyi, Department of Psychology and Center for Research in Cognitive Sciences, National Chung Cheng University
The main goal of this study was to establish a test that utilizes face images collected and normed in the Taiwanese society in order to provide a culturally calibrated tool for assessing face memory ability. To that end, we largely followed the procedure by which the Cambridge Face Memory Test (CFMT) was created and popularized by Duchaine and Nakayama (2006). The face images were drawn and modified from the recent database created by Shyi, Huang and Yeh (2013). Unlike the CFMT, our stimuli included both male and female face images. Six individuals, three males and three females, were chosen randomly as target faces. All faces were cropped so that no hair was visible and facial blemishes were removed. The faces were posed with neutral expressions. The test was divided into three stages: the same images, the novel images, and the novel images with noise. In the same-images stage, three face images were presented on each test item, one of which was identical to the target face. In the novel-images stage, face images in each test item were varied in terms of lighting, pose, or both. In the third stage, four different levels of Gaussian noise were added to the face images. Fifty-two participants, 27 female and 25 male, were tested. In accordance with Duchaine and Nakayama (2006), we anticipated that the participants would perform well on the first stage, and then declined on the second and third stages. As predicted, the accuracy of the first stage was .94, and dropped to .71 and .70, respectively, for the two remaining stages, showing an overall pattern very similar to that reported by Duchaine and Nakayama. However, an interaction between gender of face image and stage reveals that female face images were more negatively influenced by variations in stages 2 and 3 than male images. Moreover, unlike Duchaine and Nakayama, female participants performed equally well on male and female face images, whereas male participants performed better on male than on female images. In summary, while showing a general pattern similar to that of CFM, the differential results of TFMT highlights the importance to create culturally calibrated test for assessing face memory ability.
(8) Is Emotion of A Schematic Face Just A Mouth Line? Emotional Advantage Beyond Simple Features
Yang-Ming Huang1 & Yei-Yu Yeh2, 1Department of Psychology, Fu-Jen Catholic University & 2Department of Psychology, National Taiwan University
We investigated whether emotional advantage of schematic faces is purely based on the processing of simple features. Multiple tokens of each valence category were presented briefly, manipulating face orientation and target valence. A four-dot mask, indicating the target for valence classification, was terminated simultaneously with the search array in the common-offset condition or delayed after array offset in the delayed-offset condition. Face probes and word probes were used in Experiments 1 and 2, respectively. The results of both experiments showed the object substitution effect with better performance in the common-offset condition than in the delayed-offset condition. When face probes provided perceptual features in the decision, a main effect of emotional advantage was observed. Without the perceptual features in the word probes, emotional advantage was observed only with upright faces and not with inverted faces. Object masking did not modulate the emotional advantage. The results suggest that configural aspects of an emotional schematic face are stronger than those of a neutral face to compete against the delayed mask in process prioritization. Moreover, reactivating the configural representation is necessary to observe the emotional benefit of a schematic face. Configural features are important for conveying the emotion of a schematic face.
(9) Perceptual grouping facilitates attention allocation in a dynamic environment for older adults
Pei-Chen Chiu1 & Yu-Wei Liao1 & Cheng-Ta Yang2 & Chun-yu Kuo1 & Yei-Yu Yeh1, 1Department of Psychology, National Taiwan University & 2Department of Psychology, National Cheng Kung University
We investigated whether grouping can enhance older adults’ ability to allocate attentional resources in a dynamic object tracking task. In Experiment 1, 12 young (M = 22 years old) and 12 elderly (M = 68 years old) participants performed the multiple object tracking task in three conditions. In the baseline condition, eight balls in gray color moved with four balls as the tracked targets and four balls as the ignored distractors. In the color-grouping condition, the tracked targets were in one color while the ignored distractors were in a different color. In the color-mixture condition, four different colors were used in the target and distractor groups so that each target shared a color with a distractor. The results showed an age effect in the baseline and color-mixture conditions, with young participants showing higher hit rates in identifying which objects were the targets after movement ended. The age effect disappeared in the color-grouping condition, with both groups showing high hit rates. In Experiment 2, 12 young (M = 22 years old) and 12 older adults (M = 65 years old) participants tracked 5 target balls while ignoring another 5 distractors. Four conditions were constructed: baseline, color-grouping, speed-grouping, and color-speed-grouping. The results showed a significant age effect in three conditions except for the color-speed-grouping condition. Taken together, the results from these two experiments showed that configural processing based on perceptual grouping can facilitate attention allocation in a dynamic environment for older adults.
(10) You must remember the whole for the reappearance of a part to capture your attention
Chun-Yu Kuo & Yei-Yu Yeh, Department of Psychology, National Taiwan University
When an external stimulus meets an internal representation maintained in working memory, it captures attention. This working memory-based attentional capture has been demonstrated in various contexts using different stimuli. Yet, few studies have investigated the context for a part of a visual array maintained in working memory to capture attention. In three experiments, we addressed this issue using a dual-task paradigm. The participants remembered the spatial locations of four colored disks for later recognition, judged the direction of a moving stimulus along with a static distractor, and recognized whether a probe array matched with the memorized locations. Color was irrelevant to both the memory and attention tasks. In the valid condition, the color of the moving target was identical to the previously displayed color; and in the invalid condition, the color of the static distractor was the same as the previously displayed color. In Experiments 1 and 2, four empty disks were used in recognition to probe the to-be-remembered locations. In Experiment 3, a solid line drew the outline contour of the four locations for recognition. Colored disks were used as stimuli in the motion task for Experiments 1 and 3 whereas colored crosses were used in Experiment 2. The results showed working memory-driven attentional capture only in Experiment 1, with two colored disks presented in the motion judgment task and four colored disks probed for location memory. The contrast in the results between Experiments 1 and 2 showed that external stimulation and internal representation must share the same part for it to capture attention. Moreover, probing with a whole display appears to highlight the disks in the whole so that the reappearance of a colored disk can capture attention. Probing with an outline contour may decrease the importance of the disks and thus the reappearance of a colored disk does not capture attention. Remembering the whole display that configured from individual parts is necessary for the reappearance of a part to capture attention in a selection task.
(11) An encoding advantage for faces of own-race over other-race in Taiwanese participants: A morphed face study
Mandy Chun-Man1 & Nick Shun-Fu Hu2 & Dennis Chu-Li Tai2 & Sarina Hui-Lin Chien2, 1Department of Biological Science and Technology, China Medical University &
2Graduate Institute of Neural & Cognitive Sciences, China Medical University
Literature on the other-race effect (ORE) in face recognition had lent a strong support for a significantly higher accuracy, adults or children alike, in recalling faces of their own race. Yet little was known as to the exact stage of visual processing at which ORE came into play. Using a morphed face paradigm, Walker and Tanaka (2003) reported a small but significant processing advantage as early as the stage of visual encoding for faces of own-race over other-race in Canadian Caucasians. The present study sets out to take Walker and Tanaka’s approach to further explore the encoding advantage hypotheses in Taiwan, a predominantly Han society. A swathe of morphed pictures of Caucasian and Taiwanese male and female face images were used as the stimuli. The participants were to perform a sequential same/different face discrimination task. In each trial, the participant viewed an Asian or Caucasian parent face for 1 s and followed by either the same parent face or a different morph face. The stimulus probability for the “same” and “different” trials was set equal. In the “different” trials, the Asian (or Caucasian) parent face will be followed by a morph face consisting of 90%, 80%, 70%, 60%, 50%, or 40% contribution from that parent face and the remaining percent contribution provided by the Caucasian (or Asian) parent face. The results (N=20, 10 females) showed that, on average, Taiwanese participants had a higher % correct rejection for Asian parent condition than that of Caucasian parent condition. In addition, the psychometric function on the percent of correct rejections for Asian- and Caucasian-parent conditions were separately fitted with an accumulative normal function (with a centroid parameter μ and a slope parameter δ). The participants showed a significantly smaller discrimination threshold (15.22% for Asian vs. 35.79% for Caucasian faces) and a steeper slope (1.15 for Asian vs. 0.909 for Caucasian faces) in the own-race conditions, regardless of gender. Taken together, the present study provides evidence supporting a processing advantage at encoding stage for faces of own-race over other races. Further data collections for international Caucasian student participants are in progress.
*This project was supported in part by Taiwanese Ministry of Science and Technology Grant “MOST103- 2410-H-039-002-MY3” to Dr. Sarina. H. L. Chien and partly by the Student Research Grant “MOST103-2815-C-039-013-H” to Mandy C. M. Chen.
(12) Fluency and engagement due to optimal information complexity is sometimes unfavorable
Wang Man-Ying1 & Da-Lung Tang2 & Jie-Li Tsai3, 1Department of Psychology, Soochow University & 2Department of Mass Communication, Tamkang University & 3Department of Psychology, National Chengchi University
What is the optimal level of information complexity for the viewer? We examined consumers' attitude towards banner ads and shopping Apps of different information complexity. A measure of information complexity was developed that divides the number of information sources by the number of salient regions in the ad or App. Participants liked the ad or app the best when the information complexity measure was approximately 1.Their mean fixation duration was longer, the number of fixations was smaller and the saccade amplitude was larger, showing an engaged mode of processing. Participants' deliberate viewing strategy or goal served as the modulator. When they were high in construal level and did not rely on fluency for their ad judgments or when price information was present in the App, their preference for information complexity and eye movement behavior were different. These findings also suggested that eye movement behavior is associated with experienced fluency that, in turn, affected one's preference.
(13) The effect of the chromaticity of image elements on symmetry detection
Chia-Ching Wu1 & Chien-Chung Chen2, 1 Department of Psychology, Fo Guang University & 2Department of Psychology, National Taiwan University
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.
(14) Working memory capacity predicts workload capacity
Ju-Chi Yu & Ting-Yun Chang & Cheng-Ta Yang, National Cheng Kung University
We investigated the relationship between working memory capacity (WMC) and workload capacity (WLC). An operation span task and three redundant-target detection tasks were used to measure an individual’s WMC and WLC. Both nonparametric and parametric analyses showed that high-WMC group had larger WLC than low-WMC group when the redundant information came from visual and auditory modalities, indicating superior processing capacity. Such difference was eliminated when multiple processes required processing for a single working memory subsystem. These results highlighted the role of executive control in integrating information from the two subsystems of working memory for perceptual decision making.