From parallel to serial processing: a computational study of visual search

A novel computational model of a preattentive system performing visual search is presented. The model processes displays of lines, reproduced from Wolfe, Friedman-Hill, Stewart, and O'Connell's (1992) and Treisman and Sato's (1990) visual-search experiments. The response times measured in these experiments suggest that some of the displays are searched serially, whereas others are scanned in parallel. Our neural network model operates in two phases. First, the visual displays are compressed via standard methods (principal component analysis), to overcome assumed biological capacity limitations. Second, the compressed representations are further processed to identify a target in the display. The model succeeds in fast detection of targets in experimentally labeled parallel displays, but fails with serial ones. Analysis of the compressed internal representations reveals that compressed parallel displays contain global information that enables instantaneous target detection. However, in representations of serial displays, this global information is obscure, and hence, a target detection system should resort to a serial, attentional scan of local features across the display. Our analysis provides a numerical criterion that is strongly correlated with the experimental response time slopes and enables us to reformulate Duncan and Humphreys's (1989) search surface, using precise quantitative measures. Our findings provide further insight into the important debate concerning the dichotomous versus continuous views of parallel/serial visual search.     

The Lower Visual Search Efficiency For Conjunctions Is Due To Noise and Not Serial Attentional Processing

Models of human visual processing start with an initial stage with parallel independent processing of different physical attributes or features (e.g. color, orientation, motion). A second stage in these models is a temporally serial mechanism (visual attention) that combines or binds information across feature dimensions. Evidence for this serial mechanism is based on experimental results for visual search. I conducted a study of visual search accuracy that carefully controlled for low-level effects: physical similarity of target and distractor, element eccentricity, and eye movements. The larger set-size effects in visual search accuracy for briefly flashed conjunction displays, compared with feature displays, are quantitatively predicted by a simple model in which each feature dimension is processed independently with inherent neural noise and information is combined linearly across feature dimensions. The data are not predicted by a temporally serial mechanism or by a hybrid model with temporally serial and noisy processing. The results do not support the idea that a temporally serial mechanism, visual attention, binds information across feature dimensions and show that the conjunction-feature dichotomy is due to the noisy independent processing of features in the human visual system.     

A vision research apparatus for broad luminance range displays.

Lightness, the perceived gray shade of a surface, and the perception of self-luminous surfaces--that is, surfaces that appear to glow--have most often been studied with paper displays and computer-generated stimuli presented on CRT monitors. Although both methods are often effective, experiments that require a wide range of luminance values in the same display are often difficult to conduct with paper and computer displays alone. Also, color mode appearance is often an issue when surface color perception is the topic of research; CRT monitors are essentially light sources themselves and often appear in the luminous mode of color appearance. Here, we describe an apparatus in which the target is an undetected aperture whose luminance is adjustable. Whereas a typical CRT monitor offers a luminance range of about 100:1, much broader luminance ranges are possible with the described apparatus. Unlike a CRT monitor, the stimulus background will always appear in the surface mode of color perception, and the target(s) can appear as either surface colors or luminous colors. Apparatus modifications are possible, including the addition of a stereoscope or an embedded CRT for creating an adjustable region that is computer controlled.     

A vision research apparatus for broad luminance range displays

Lightness, the perceived gray shade of a surface, and the perception of self-luminous surfaces—that is, surfaces that appear to glow—have most often been studied with paper displays and computer-generated stimuli presented on CRT monitors. Although both methods are often effective, experiments that require a wide range of luminance values in the same display are often difficult to conduct with paper and computer displays alone. Also, color mode appearance is often an issue when surface color perception is the topic of research; CRT monitors are essentially light sources themselves and often appear in the luminous mode of color appearance. Here, we describe an apparatus in which the target is an undetected aperture whose luminance is adjustable. Whereas a typical CRT monitor offers a luminance range of about 100:1, much broader luminance ranges are possible with the described apparatus. Unlike a CRT monitor, the stimulus background will always appear in the surface mode of color perception, and the target(s) can appear as either surface colors or luminous colors. Apparatus modifications are possible, including the addition of a stereoscope or an embedded CRT for creating an adjustable region that is computer controlled.     

Odd-item search by pigeons: Method,instrumentation, and uses

In odd-item visual search, subjects confront a display on which a number of stimulus items appear. All but one of these items are identical; the subject must respond to the one item (the target) that in some way differs from all the others (the distractors). The time required to find the target reflects the similarity between the target form and the distractor form. A matrix of search times for all possible pairs of a set of 20 or more items can be obtained in a single session. Such similarity matrices may reflect stimulus features, dimensions, and categories, among other things. A method is described through which pigeons learn odd-item search rapidly and perform with high accuracy despite the appearance of each form as a target on some trials and as a distractor on others. The paper also describes the essential apparatus and exemplifies displays and data.     

A powerful,inexpensive experiment controller or IBM PC interface and experiment control language

A very powerful, but inexpensive, advanced technology (16 MHz 80C188EB) experiment controller is described. It can be programmed in either a BASIC-like (ECBASIC) or an ALGOL-like (ECL) procedure specification language. It provides 1-msec resolution, optional transparent total data logging, and is designed to function as a remote peripheral processor in a network with virtually any computer acting as the network supervisor (e.g., IBM compatible or Macintosh). Each serial port on the host computer can support up to 10 simultaneous experiments. The various optional I/O modules provide for opto-isolated normally open or normally closed switch operation input, high current output, as well as D/A and A/D functions. Our I/O modules can also be plugged directly into an IBM PC parallel port by using a small adaptor board. In this way, the advantages of using ECBASIC or ECL to control experiments are available to researchers who wish to simply interface their host computer directly to the apparatus in order to minimize expense.     

The changing pattern of perceptual analytic strategies and response selection with practice in a two-choice reaction time task

In the experiment reported here the primary hypothesis proposed, which was supported by the data, was that subjects performing a two-choice self-paced serial RT task would learn to change their strategy of perceptual analysis of signals presented to them as they became progressively more practised. Early in practice each signal is identified as a particular state of the display and an appropriate response is then made. Well practised subjects, however, select their responses by reference to the change or constancy between successive displays. This strategy implies that the assumptions made by the generally accepted simple S-R connectionist model of two-choice serial RT may be quite misleading, since the choice of any particular response must be determined not only by the display state but also by what the previous response had been.     

Sequential dynamics in visual short-term memory

Visual short-term memory (VSTM) is thought to help bridge across changes in visual input, and yet many studies of VSTM employ static displays. Here we investigate how VSTM copes with sequential input. In particular, we characterize the temporal dynamics of several different components of VSTM performance, including: storage probability, precision, variability in precision, guessing, and swapping. We used a variant of the continuous-report VSTM task developed for static displays, quantifying the contribution of each component with statistical likelihood estimation, as a function of serial position and set size. In Experiments 1 and 2, storage probability did not vary by serial position for small set sizes, but showed a small primacy effect and a robust recency effect for larger set sizes; precision did not vary by serial position or set size. In Experiment 3, the recency effect was shown to reflect an increased likelihood of swapping out items from earlier serial positions and swapping in later items, rather than an increased rate of guessing for earlier items. Indeed, a model that incorporated responding to non-targets provided a better fit to these data than alternative models that did not allow for swapping or that tried to account for variable precision. These findings suggest that VSTM is updated in a first-in-first-out manner, and they bring VSTM research into closer alignment with classical working memory research that focuses on sequential behavior and interference effects.     

Serial control of phonology in speech production: a hierarchical model

A dynamic oscillator-based model of the sequencing of phonemes in speech production (OSCAR) is described. An analysis of phoneme movement errors (anticipations, perseverations, and exchanges) from a large naturalistic speech error corpus provides a new set of data suitable for quantitative modeling and is used to derive a set of constraints that any speech-production model must address. The new computational model is shown to account for error type proportions, movement error distance gradients, the syllable-position effect, and phonological similarity effects. The model provides an alternative to frame-based accounts, serial buffer accounts, and associative chaining theories of serial order processing in speech.     

风险寻求中的图形表征效应

在负性风险信息沟通中,采用文本表征与图形表征,结果发现图形表征方式更能引起风险回避行为。基于此,本研究以获益方案作为风险沟通信息,分别采用文本表征与图形表征来探索信息表征方式对风险寻求的影响。结果显示:在文本表征方式中,个体决策主要受备择方案之间风险大小的影响;而图形表征方式中,个体不仅受备择方案之间风险大小的影响,而且更受备择方案之间获益大小的影响。结果说明:相对文本表征,图形表征方式更能引起决策者为了获益而冒更大的风险,从而表现出风险寻求偏好,从而证实"图形效应"在风险决策中存在普遍性,为我们风险决策信息提供依据。     

Pattern redundancy and detection in very short-term memory

A tachistoscopically presented target letter is detected better in a background of redundant items than in nonredundant material. The present experiment extended this finding by showing that detection rate increased systematically across three levels of background letter redundancy. A conclusion based on the earlier findings was that a serial process of letters in memory could not account for the effects of background redundancy on detection accuracy. Contrary to this conclusion, the present analysis showed that detection for semiredundant displays was successfully predicted by two variants of a serial process model. One model assumed that the scan of a letter is faster when the letter is the same as any prior letter scanned in the matrix. The other model assumes that nonredundant letters are processed serially and that redundant letters are processed all at once, i.e., in parallel.     

A computer-based test battery for the assessment of static and dynamic spatial reasoning abilities

A battery of 10 computerized tests of spatial ability is described. It includes 5 tests that require reasoning about static spatial displays and 5 tests that require reasoning about dynamically displayed spatial information. An integrated software package for task presentation and data analysis is described, along with a summary of results from a validation study comparing performance on the computer-based tests with performance on standardized paper-and-pencil tests of spatial abilities. Finally, research applications of the current battery are discussed.     

Recording and analyzing eye-position data using a microcomputer workstation

This paper presents a PC-based eye-position data collection and analysis system. Software routines are described that supplement hardware calibration procedures, improving data-collection accuracy and reducing the number of unusable trials. Collected eye-position data can be remapped over a displayed stimulus image and spatially and temporally represented by parameters such as individual fixations, clusters of fixations (Nodine, Carmody, & Kundel, 1978), cumulative clusters, and gaze durations. An important feature of the system is that the software routines preserve scan-path information that provides a sequential dimension to the analysis of eye-position data. A “hotspot” analysis is also described, which cumulates, across 1 or more observers, the frequency of eye-position landings or “hits” on designated areas of interest for a given stimulus. Experimental applications in the fields of radiology, psychology, and art are provided, illustrating how eye-position data can be interpreted both in signal detection and in information-processing frameworks using the present methods of analysis.     

Interactive clustering

An interactive strategy for applying cluster-analytic techniques in behavioral research is presented. The two-part approach stresses the use of on-line computers for both data collection and analysis. In data collection, an extension of multidimensional unfolding to clustering reduces the number of judgments required of subjects by as much as 50%, During data analysis, the interactive procedures described permit the testing of multiple clustering models from an extensive family. With each selection, the goodness of fit of the model to the data can be tested. In addition to improving efficiency, the interactive strategy promoted here combines the advantages of the original nonmetric clustering procedures (e.g., Johnson, 1967) with those of the latest linear additive models (e.g., Sattath & Tversky, 1977; Shepard & Arabie, 1979).     

On-line computers in psychology: A laboratory course for advanced psychology majors

This course is structured as a series of increasingly difficult projects which involve students in all aspects of using computers for experimentation. Projects range from programming simple experiments utilizing monitor displays and keyboard responses to the use of hardware interfaces to connect complex peripherals. Students not only design and develop their projects, but also engage in data collection and analysis. The course provides experience with hardware and software as well as the intricacies of debugging the resulting project.     

Implications of search accuracy for serial self-terminating models of search

In two experiments, the data from a total of 35 participants was compared to predictions from two competing serial search models, one with perfect memory and one with no memory. Reaction times were collected from an unlimited exposure duration, 4-alternative, target identification search task with array sizes of 8, 12, or 16 items. From the linear function relating RT to number of distractors a hypothetical “inspection time/item” was computed under the assumption of SSTS. Target identification accuracy was obtained from the same participants using the same displays, but with pre- and post-masks used to limit total inspection time (to 120, 180, 300, and 540 ms). Neither the memoryful nor memoryless model provided a good enough fit to the data to support the assumption that serial inspection is the primary process determining search time. Despite direct evidence for serial allocation of attention in visual search, these results suggest that scholars should seriously consider the possibility of hybrid search models in which serial and parallel strategies operate jointly.     

Perception of two-dimensional Markov constraints within visual displays

A method is described for constructing visual displays in which statistical constraints are encoded within two spatial dimensions without introducing one-dimensional linear constraints. Within each local group of four elements, the state of one element was determined with a given probability by the previously generated states of the other three. Ss rated such displays on a scale from “lumpy” (or crude texture) to “lacy” (or even texture). The consistency of classification obtained for displays with strong aggregated (“lumpy”) properties was substantially higher than that obtained for displays with strong distributed (“lacy”) properties. An incidental feature of the Ss’ behavior was their deliberate degrading of the visual quality of the displays. Comparison is made with one-dimensional displays concatenated in two dimensions.     

Using context to build semantics

Latent semantic analysis (LSA) is a model of knowledge representation for words. It works by applying dimension reduction to local co-occurrence data from a large collection of documents after performing singular value decomposition on it. When the reduction is applied, the system forms condensed representations for the words that incorporate higher order associations. The higher order associations are primarily responsible for any semantic similarity between words in LSA. In this article, a memory model is described that creates semantic representations for words that are similar in form to those created by LSA. However, instead of applying dimension reduction, the model builds the representations by using a retrieval mechanism from a well-known account of episodic memory.     

A microprocessor control and recording system for instrumental behavior research

This paper describes a miroprocessor system developed for experimental control and data collection in real time. By connecting the experimental equipment to the microprocessor via a serial communication interface, transmission of various animal test-chamber functions over greater distances with only a few cables is facilitated. The microprocessor is programmed on a host computer (PDP/8). A combined terminal station with a cassette deck is used for data and program storage.     

Conditioned suppression as a psychophysical technique for the cat

A modified conditioned suppression paradigm coupled with a new training apparatus has been applied to gathering sensory data in the cat. The method excels in that training, data collection, and reliability are quickly achieved and it can be used with subtle stimulus changes.