In 1998/1999, three participants trained for up to 74-h-long sessions to find a target present on half the trials in visual displays of 1, 2, or 4 initially novel objects. There were four targets and four foils that never changed. Displays occurred simultaneously, or the objects occurred successively, or the four features of each object occurred successively. When successive, the SOAs were short (17, 33, or 50 ms), so the displays appeared simultaneous, making it likely that the search strategy was the same in all conditions. A 2004 publication examined only the simultaneous condition and found evidence suggesting serial search as well as some small amount of automatic attention to targets and occasional early or late termination of search. A 2021 publication examined only the displays with single objects, obtaining evidence for dynamic perception of features. These studies drew conclusions from modeling subtle aspects of the response time distributions; extending such modeling to all conditions would have been complex making it difficult to understand the main processes at work. Here, we present a simple way to extend the 2021 model to the conditions with multiple item displays. It is a hybrid model with parallel automatic processing of features from all display items, processing that finishes during the first comparison, combined with serial comparisons that terminate when a target is found, or when none is found. When objects occur sequentially, there is a tendency to compare first the first object presented that probability rising with SOA. This model gives a good qualitative account of the accuracy and median response times from all the conditions. This success suggests that a more complex model incorporating the dynamic processes of the 2021 model would provide an excellent quantitative account for the accuracy and response time distributions for all the conditions of this visual search study.

Author(s): Harding, Samuel; Shiffrin, Richard | Abstract: We explore the dynamic coordination of perception, decision, and action underpinning perceptual choices by recording cursor movements during a binary response task. Stimuli were presented sequentially to control the time-course of perception, and we utilized a Hidden Markov Model (HMM) to relate measured movements of the mouse cursor to latent cognitive processes. Stimuli were simple perceptual objects comprised of two features, one of which was fully diagnostic of the correct response, while the other provided a probabilistic cue. The order of their arrival varied across trials, allowing us to manipulate the order of feature processing. The model builds upon response time methods and makes predictions about when individual features were perceived and the accumulation of evidence towards a response, every 10 milliseconds of each trial.

Participants gave recognition judgments for short lists of pictures of everyday objects. Pictures in a given list were an equal mixture of three types that varied according to the way they were used as targets and foils earlier in the same session. Under consistent-mapping (CM), targets and foils never switch roles; under varied-mapping (VM), targets and foils switch roles randomly across trials; whereas all-new (AN) items are novel on each trial of the experiment. Past research has shown that markedly enhanced performance occurs in CM conditions, leading to conclusions that item-response learning takes place in CM, perhaps automatically. However, almost all past research has compared CM, VM, and AN performance in between-blocks designs in which participants may adopt different cognitive strategies and criterion settings across the conditions. The present mixed-list design holds constant the strategy and criterion settings that are used for CM, VM, and AN items, and produced patterns of performance dramatically different than those observed in pure-list control conditions. We develop an extended version of an exemplar-based random-walk model of probe recognition to account for the major qualitative effects in the data. The data and the modeling provide evidence for strong item-response learning for CM foils but weak item-response learning for CM targets. We consider possible explanations for these effects in our General Discussion. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Eight initially novel objects with four features were learned by three participants over about 70 sessions in a variety of present-absent search tasks. This article analyzes and models trials with a single object presented for test. The features of the object were presented simultaneously, or successively at rates fast enough that the objects appeared to be simultaneous (inter-stimulus intervals were 16, 33, or 50 ms). Classification of a test object as target or foil required a conjunction of two features. When successively presented, features diagnostic for target presence could arrive first or last, and vice versa for features diagnostic for foil presence. Two results were particularly important: (1) the order in which target-diagnostic or foil-diagnostic features appeared produced large changes in accuracy and response times; (2) simultaneous feature presentation produced lower accuracy than sequential presentation with target-diagnostic features arriving first, despite the delay in such features arriving. The results required a dynamic model for perception and decision. The model has features perceived at independent times. It accumulates evidence at each moment based on the features perceived up to that time, and the diagnosticity of those features for classifying the test object as target or foil. The model also assumes that configurations of features provide evidence as processing continues: when all four features of an object are perceived the evidence points without error to the correct response. The results and modeling support the view that perceptual and decision processes operate concurrently and interactively during identification, recognition, and classification of well-learned objects, rather than in successive stages.

Black, relative to White, individuals have experienced discrimination for centuries in the United States. Recent work suggests that subtle differences in how novel Black faces are initially perceived relate to prejudicial behavior. One such difference is that non-Black people attend more to the eyes of White versus Black novel faces. The present study sought to better characterize this difference by assessing how distinct individual differences widely shown to relate to prejudicial behavior—internal motivation to respond without prejudice (IMS), external motivation to respond without prejudice (EMS), and implicit race bias—relate to disparities in attending to the eyes of novel Black and White faces. Participants viewed novel Black and White faces one at a time on the right or left side of the display. Replicating a race-based disparity in visual attention to the eyes, non-Black perceivers fixated more on the eyes of White in comparison to Black faces. Individual differences among perceivers corresponded with the extent of this race-based disparity. IMS had a negative relationship with a race-based disparity in attention to the eyes, such that higher levels of IMS among perceivers corresponded with lower disparities in attention. Implicit race bias had a positive relationship with this disparity, such that higher levels of implicit race bias among perceivers corresponded with higher disparities in attention. Together, these findings illustrate that two individual differences known to affect prejudicial behavior are associated with preferential gaze patterns in visual attention toward faces on the basis of race.

Infants' development of joint attention shows significant advances between 9 and 12 months of age, but we still need to learn much more about how infants coordinate their attention with others during this process. The objective of this study was to use eye tracking to systematically investigate how 8- and 12-month-old infants as well as adults dynamically select their focus of attention while observing a social partner demonstrate infant-directed actions. Participants were presented with 16 videos of actors performing simple infant-directed actions from a first-person perspective. Looking times to faces as well as hands-and-objects were calculated for participants at each age, and developmental differences were observed, although all three groups looked more at hands-and-objects than at faces. In order to assess whether visual attention was coordinated with the actors' behaviors, we compared participants looking at faces and objects in response to gaze direction as well as infant-directed actions vs. object-directed actions. By presenting video stimuli that involved continuously changing actions, we were able to document that the likelihood of joint attention changes in both real and developmental time. Overall, adults and 12-month-old infants' visual attention was modulated by gaze cues as well as actions, whereas this was only partially true for 8-month-old infants. Our results reveal that joint attention is not a monolithic process nor does it develop all at once.

In a world that relies increasingly on large amounts of data and on powerful Machine Learning (ML) models, the veracity of decisions made by these systems is essential. Adversarial samples are inputs that have been perturbed to mislead the interpretation of the ML and are a dangerous vulnerability. Our research takes a first step into what can be an important innovation in cognitive science: we analyzed human’s judgments and decisions when confronted with targeted (inputs constructed to make a ML model purposely misclassify an input as something else) and non-targeted (a noisy perturbed input that tries to trick the ML model) adversarial samples. Our findings suggest that although ML models that produce non-targeted adversarial samples can be more efficient than targeted samples they result in more incorrect human classifications than those of targeted samples. In other words, non-targeted samples interfered more with human perception and categorization decisions than targeted samples.

The present research infers aspects of spatial attention from movement to targets (and preferably not to foils) of a mousecontrolled cursor on a computer monitor. The long-term goal is a data-rich and rapid assessment technique that can be used to diagnose individual and clinical deficits of attention. The aim of this present research is validating the approach using a college population of subjects. In the experiment, participants attempt to move a cursor toward three spatial positions at which targets appear rapidly but at irregular times, and attempt to inhibit movements toward foils appearing at those positions. We assume that cursor movements toward a position indicates attention has been directed toward that position. A modified Hidden Markov Model (HMM) uses five sources of evidence, all based on parameters to be estimated, to predict the time varying movement of attention: four aspects of cursor movement and a probability that attention will move from one time interval to the next. Five minutes of data are used to estimate parameters for each subject that produce a predicted attention trajectory which best matches what the subject is instructed to do. These parameters are used to predict the attention trajectory for the remainder of the hour of testing. The predictions of attention movements are then matched to the appearance of targets and foils to infer such components of attention as ability to respond to targets vs foils, times to do so, and changes in these components over time. The results illustrate a promising approach to assessment of attention that could likely be employed for both adults and children in clinical settings requiring short testing periods.

Previous research suggests that 9-month-old infants will develop a response bias in the A-not-B search paradigm after only observing an experimenter search for a hidden object on A-trials. In the current experiment, we tested whether infants would persist in making errors when only the hands-and-arms of the experimenter were visible. Three different conditions were included: (1) the experimenter was silent while hiding and finding the object, (2) the experimenter communicated with the infant via infant-directed speech, or (3) the body of the experimenter was visible during the training phase before his head and body were occluded during the test phase. Unlike previous studies, the results revealed that a significant proportion of infants searched correctly when the body of the experimenter was not visible, and only the combination of infant-directed speech and familiarization with a fully-specified body resulted in a majority of infants committing search errors. These results are interpreted as suggesting that the likelihood of infants committing search errors is dependent on their motor simulation of the experimenter’s reaching. The strength of this simulation is graded by the similarity between the observed action and the motor representation.

Social stimuli are a highly salient source of information, and seem to possess unique qualities that set them apart from other well-known categories. One characteristic is their ability to elicit spatial orienting, whereby directional stimuli like eyegaze and pointing gestures act as exogenous cues that trigger automatic shifts of attention that are difficult to inhibit. This effect has been extended to non-social stimuli, like arrows, leading to some uncertainty regarding whether spatial orienting is specialized for social cues. Using a standard spatial cueing paradigm, we found evidence that both a pointing hand and arrow are effective cues, but that the hand is encoded more quickly, leading to overall faster responses. We then extended the paradigm to include multiple cues in order to evaluate congruent vs. incongruent cues. Our results indicate that faster encoding of the social cue leads to downstream effects on the allocation of attention resulting in faster orienting.

Infants' understanding of a pointing gesture represents a major milestone in their communicative development. The current consensus is that infants are not capable of following a pointing gesture until 9-12 months of age. In this article, we present evidence from 4- and 6-month-old infants challenging this conclusion. Infants were tested with a spatial cueing paradigm in Experiment 1 (500-ms stimulus-target onset asynchrony [SOA]) and Experiment 2 (100-ms SOA). The results revealed that the younger infants shifted their attention in the cued direction when presented with a pointing gesture and with a foil (i.e., same size and shape as pointing gesture) at both SOAs. Older infants shifted their attention only in response to the pointing gesture at 100-ms SOA. Experiment 3 tested infants' preferences for the social stimulus (i.e., pointing gesture) relative to the foil and a non-social stimulus (i.e., an arrow). The results revealed that infants are biased to selectively attend to the pointing gesture. Taken together, these results suggest that 4- and 6-month-old infants are capable of preferentially selecting and following a pointing gesture. It is theorized that this early capacity assists infants in their developing understanding of triadic forms of communication.

The individuation of objects via spatiotemporal continuity and surface features, such as texture and orientation, was studied with a streaming-bouncing paradigm. In a typical display, two moving discs start on opposite sides of the screen and move toward each other, coincide, and then reverse directions. If the discs share identical surface features, the perception of the event is ambiguous and the discs could be perceived as streaming or bouncing. If the discs are not identical, distinctive features could disambiguate this percept. The following experiments were designed to explore how surface features interact with spatial-temporal information and bias the percept toward streaming or bouncing. Four experiments were conducted in which participants judged whether the two discs (1.5[sup]0[/sup] visual angle) streamed or bounced (constant velocity = 3.5 horizontal deg/sec) as they moved across a grey background. Between-subject variables included trajectory (horizontal vs. diagonal) and surface features (luminance gratings or random dots). The orientations of the gratings for the two discs differed by 0, 15, 30, or 45 deg; the probability density distributions for the random dots differed by 0, 10, or 20%. In each experiment, there were three conditions mediating the intersection of the two discs (occlusion, transparency, or visible occluder), which were programmed to bounce on every trial. Participants were presented with a total of 288 trials with 24 (luminance grating) or 32 (random dots) per intersection condition by surface feature difference (ΔF). Consistent with previous studies, the results revealed a significant bias to perceive streaming (on average 80% with ΔF=0) in the transparent and occluder conditions, which decreased as a function of ΔF. By contrast, the likelihood of perceiving streaming in the occlusion condition rarely exceeded 30%. These results also interacted with trajectory and surface feature. Taken together, these results suggest that multiple processes are involved in individuating objects. Meeting abstract presented at VSS 2013