James P. Thomas

A signal detection model predicts the effects of set size on visual search accuracy for feature, conjunction, triple conjunction, and disjunction displays

Recently, quantitative models based on signal detection theory have been successfully applied to the prediction of human accuracy in visual search for a target that differs from distractors along a single attribute (feature search). The present paper extends these models for visual search accuracy to multidimensional search displays in which the target differs from the distractors along more than one feature dimension (conjunction, disjunction, and triple conjunction displays). The model assumes that each element in the display elicits a noisy representation for each of the relevant feature dimensions. The observer combines the representations across feature dimensions to obtain a single decision variable, and the stimulus with the maximum value determines the response. The model accurately predicts human experimental data on visual search accuracy in conjunctions and disjunctions of contrast and orientation. The model accounts for performance degradation without resorting to a limited-capacity spatially localized and temporally serial mechanism by which to bind information across feature dimensions.

Bandwidths of orientation channels in human vision

The ability of observers to discriminate between stimuli differing in orientation was measured using low contrast, foveally viewed stimuli. Detection and discrimination performance were measured simultaneously. A model is presented which permits bandwidths of orientation-tuned mechanisms to be estimated from the data. In a group of five observers, half-amplitude bandwidths varied from 10 degrees to 20 degrees.

Detection and identification: how are they related?

Detection and identification are described as using the same base of sensory information but applying different decision processes. Consistent with this view, there is no evidence of different sensory cutoffs for the two tasks, and accuracies in the two tasks vary as a function of stimulus strength in closely related fashion. Identification accuracy also depends on the extent to which stimuli are processed independently, and the quantitative relationship between identification and detection can be used to estimate the degree of independence.

PARALLEL PROCESSING IN VISUAL SHORT-TERM MEMORY

Visual short-term memory for the contrast and spatial frequency of sinusoidal gratings was measured in a delayed discrimination task in which the 2 stimuli to be compared were separated in time by 1-10 s interstimulus intervals (ISIs). Delayed discrimination thresholds for spatial frequency and contrast were compared, both when the 2 types of thresholds were measured in separate blocks of trials and when the 2 types of measures were randomly intermixed in an uncertainty paradigm, which required participants to process information about both dimensions on each trial. In both cases, accuracy of memory for spatial frequency was independent of ISI, but memory for contrast decreased as ISI increased. Performance was lower in the uncertainty case, but only by an amount predicted by statistical decision theory for independent sources. The results are consistent with a model assuming a set of parallel special-purpose visual discrimination and short-term memory mechanisms.

Neural recoding in human pattern vision: model and mechanisms.

We describe a model of neural recoding in spatial vision that specifies how the outputs of selected units akin to VI cells are normalized and combined to signal information about particular stimulus attributes. The recoding portion of the model is linked to psychophysical behavior via a two-stage signal-detection decision module that specifies how the outputs of the combining mechanisms are used in making fine spatial discriminations. We describe how masking and cue summation experiments isolate each of the processing stages, how earlier results from such studies guided development of the model, and we demonstrate how these procedures permit empirical estimates of model parameters as well as tests of alternative formulations. An important part of our work describes the characteristics of two complementary types of higher-level mechanisms isolated from previously published discrimination data. One sums normalized primary-level responses over disparate frequencies to signal precise information about the orientation of a stimulus; the other sums over all orientations to signal the spatial grain of texture-like patterns. We demonstrate how the model accounts for a large body of previously published discrimination data, and present the results of a new quantitative test of model predictions.

When orthogonal orientations are not processed independently

We studied the extent to which two gratings, superimposed at orthogonal orientations, are processed independently in making spatial frequency discriminations. Our findings suggest that independence is not preserved in this task, and that it breaks down in two distinct ways: (1) judgments about one component are reduced in accuracy by the presence of the second; and (2) information about the two component gratings is not used independently in making judgments about the compound stimulus formed by them. These results place constraints on the applicability of models that assume information from spatially tuned mechanisms is directly available in performing a discrimination task.

Simultaneous visual detection and identification: theory and data

A signal-detection model is presented for the two-alternative forced-choice procedure that gathers simultaneous detection and identification judgments. By using vector representation, a measure of independent processing within the visual system is derived. The independence measure is applied to data on the detection and identification of stimuli that differ along the dimensions of orientation, spatial frequency, and size. Data are presented that favor the signal-detection model over a high-threshold model.

Configural effects constrain fourier models of pattern discrimination

Many models of spatial pattern discrimination assume that judgments are based on information directly available from mechanisms tuned to limited ranges of spatial frequency and orientation. We tested the validity of this assumption for spatial frequency, orientation, and contrast information in a series of complex pattern discrimination experiments. Observers discriminated between simple gratings, between gratings masked by components that differed widely in spatial frequency and/or orientation, and between patterns that presented two cues to discrimination, one in each frequency and/or orientation band. Component cues were combined either in rigid-object correspondence (e.g. both components were rotated clockwise in one pattern and counterclockwise in the other) or in opposition (e.g. in one pattern one component was rotated clockwise, the other counterclockwise; the direction of rotation was reversed for each component in the second pattern). The results demonstrate that information from tuned pathways is not always used directly in making spatial judgments, but in some cases is combined across wide regions of the Fourier domain prior to the discrimination decision. We find two distinct patterns of combination that appear to independently signal information about texture and edges. These findings provide a potential link between low-level, spatially tuned analyzers in the visual system and higher-level pattern processing mechanisms.

Underlying psychometric function for detecting gratings and identifying spatial frequency.

Observers detected and/or identified the spatial frequencies of grating stimuli. Spatial frequency varied from 3.8 to 5.5 cycles per degree, and contrast varied from 0.001 to 0.33. In nearly all cases, the psychometric functions that relate performance on the different tasks to contrast are multiples of one underlying function, provided that the functions are expressed in standard normal deviates. The underlying function is positively accelerated at contrasts less than 0.01 and levels off at contrasts greater than 0.05. A vector model interprets the results and relates them to the responses of individual spatially tuned mechanisms.

Detection and discrimination of simple and complex patterns at low spatial frequencies

Abstract These experiments examined the extent to which low spatial frequencies are processed independently. The assessment was carried out with respect to both detection and discrimination performance. For simple sinusoidal gratings, pairs of stimuli could be discriminated when their contrasts reached threshold, if the ratio of their spatial frequencies was 3:1 or larger, suggesting idependent processing in separate channels. For smaller frequency ratios, slightly more contrast was required for discrimination than for detection, suggesting that stimuli were not processed by entirely separate channels. The detection and discrimination thresholds of complex grating stimuli fell within the ranges which would be expected if probability summation effects and summation of different closely spaced harmonic frequencies within single channels are considered, supporting the hypothesis of independent processing of low-spatial frequency information. The single exception to this involved discrimination of a square wave and a square wave with its fundamental component removed. In this case, discrimination required considerably more contrast than detection, even when factors of probability summation and within-channel summation of harmonics are considered.