Visual search asymmetry: The influence of stimulus familiarity and low-level features

Wang, Cavanagh, and Green (1994) demonstrated a pop-out effect in searching for an unfamiliar target among familiar distractors (U—F search) and argued for the importance of a familiarity difference between the target and the distractors in determining search efficiency. In four experiments, we explored the generality of that finding. Experiment 1 compared search efficiency across a variety of target-distractor pairs. In Experiments 2, 3, and 4, we used Chinese characters and their transforms as targets and distractors and compared search performance between Chinese and non-Chinese participants. We demonstrated that search asymmetry and search efficiency in the U—F condition are influenced by the presence of low-level feature differences between the familiar and the unfamiliar stimuli. Our findings suggest that the familiarity of the distractors, rather than the familiarity difference between the target and the distractors, determines search efficiency. We also documented a counterintuitive familiarity-inferiority effect, suggesting that knowledge of search stimuli may, sometimes, be detrimental to search performance.     

Visual search for global/local stimulus features in humans and baboons

Fagot and Deruelle (1997) demonstrated that, when tested with identical visual stimuli, baboons exhibit an advantage in processing local features, whereas humans show the “global precedence” effect initially reported by Navon (1977). In the present experiments, we investigated the cause of this species difference. Humans and baboons performed a visual search task in which the target differed from the distractors at either the global or the local level. Humans responded more quickly to global than to local targets, whereas baboons did the opposite (Experiment 1). Human response times (RTs) were independent of display size, for both local and global processing. Baboon RTs increased linearly with display size, more so for global than for local processing. The search slope for baboons disappeared for continuous targets (Experiment 2). That effect was not due to variations in stimulus luminance (Experiment 3). Finally, variations in stimulus density affected global search slopes in baboons but not in humans (Experiment 4). Overall, results suggest that perceptual grouping operations involved during the processing of hierarchical stimuli are attention demanding for baboons, but not for humans.     

Visual search asymmetry in pigeons

Pigeons received an odd-item search task that involved an array of 12 patterns containing 11 similar distractors and a single target. Pecks to the target resulted in the delivery of food. Accuracy was greater on trials when a distinctive feature was located in the target but not in the distractors, rather than when the feature was in the distractors but not in the target. This search asymmetry was influenced by the similarity of the target to the distractors. The results are similar to those obtained with humans.     

The effects of stimulus rotation and familiarity in visual search

Rauschenberger and Yantis (2006) observed that an intersecting circle-line combination enjoyed significantly greater search efficiency when it was oriented to resemble a "Q" than when it was oriented so that the intersecting line was vertical (cf. Treisman and Souther, 1985). Although a control experiment made it unlikely that the obliqueness of the line was responsible for the observed benefit, there was no direct evidence that this benefit was attributable to the "Q-ness" of the stimulus. In the present study, a subset of Rauschenberger and Yantis's experiments was repeated with Chinese subjects, who had never been exposed to the Latin alphabet. For these subjects, there was no benefit for the"Q"-like stimulus, in contrast to the results of Rauschenberger and Yantis's study. These results show that a simple 45 degrees rotation of a stimulus can affect search efficiency significantly--but only when this rotation bestows meaning, or familiarity, to that stimulus.     

Visual search asymmetry with uncertain targets

The underlying mechanism of search asymmetry is still unknown. Many computational models postulate top-down selection of target-defining features as a crucial factor. This feature selection account implies, and other theories implicitly assume, that predefined target identity is necessary for search asymmetry. The authors tested the validity of the feature selection account using a singleton search task without a predefined target. Participants conducted a target-defined and a singleton search task with a circle (O) and a circle with a vertical bar (Q). Search asymmetry was observed in both tasks with almost identical magnitude. The results were not due to trial-by-trial feature selection, because search asymmetry persisted even when the target was completely unpredictable. Asymmetry in the singleton search was also observed with more complex stimuli, Kanji characters. These results suggest that feature selection is not necessary for search asymmetry, and they impose important constraints on current visual search theories.     

Visual search asymmetry for viewing direction

In visual search experiments, we examined the existence of a search asymmetry for the direction with which three-dimensional objects are viewed. It was found that an upward-tilted target object among downward-tilted distracting objects was detected faster than when the orientation of target and distractors was reversed. This indicates that the early visual process regards objects tilted downward with respect to the observer as the situation that is more likely to be encountered. That is, the system is set up to expect to see the tops of these objects. We also found a visual field anisotropy, in that the asymmetry was more pronounced in the lower visual field. These findings are consistent with the idea that the tops of objects are usually situated in the lower visual field and less often in the upper field. Examination of the conditions under which the asymmetry and the anisotropy occur demonstrated the importance of the three-dimensional nature of the stimulus objects. Early visual processing thus makes use of heuristics that take into account specific relationships between the relative locations in space of the observer and 3-D objects.     

Visual search for features and conjunctions in development.

Visual search performance was examined in three groups of children 7 to 12 years of age and in young adults. Colour and orientation feature searches and a conjunction search were conducted. Reaction time (RT) showed expected improvements in processing speed with age. Comparisons of RT's on target-present and target-absent trials were consistent with parallel search on the two feature conditions and with serial search in the conjunction condition. The RT results indicated searches for feature and conjunctions were treated similarly for children and adults. However, the youngest children missed more targets at the largest array sizes, most strikingly in conjunction search. Based on an analysis of speed/accuracy trade-offs, we suggest that low target-distractor discriminability leads to an undersampling of array elements, and is responsible for the high number of misses in the youngest children.     

The influence of stimulus incongruity on the familiarity effect in visual selection

A consistent finding in the literature concerning visual selection is that a stimulus exposed to S subsequently loses some of its potential to elicit looking responses. Studies examining the familiarity effect in visual selection have not systematically examined the influence on that effect of various stimulus characteristics possessed by the familiarized stimulus. The present experiments examined the relationship between the magnitude of the familiarity effect and the level of incongruity represented by the familiarized stimulus and found it to be increasing and monotonie. In addition, Experiment 3 examined the generality of reports that familiarized stimuli are “liked” less than their unfamiliarized counterparts. The ease of explaining the results of the experiments using an “affect change” model of the familiarity effect and an “inforjnation-conflict resolution” model of that effect are discussed.     

Activation and inhibition of stimulus features in conjunction search

Two visual search experiments were conducted to examine how activation of target-relevant features and inhibition of target-irrelevant features are involved in conjunction search. The hypothesis was that if an excitatory mechanism is involved, it should be revealed as facilitation when a target and distractors are repeated in two successive displays. If an inhibitory mechanism is involved, suppression should be obtained when distractor features from one display determine the target in the following display. The results of Experiment 1 showed that facilitation was observed consistently across two set sizes (5 and 9), whereas suppression was obtained only with larger set size (9). This pattern of results was replicated and extended in Experiment 2 with different set size. It seems that both excitatory and inhibitory mechanisms are involved in conjunction search. The excitatory mechanism seems to play a role in conjunction search regardless of search difficulty, whereas the inhibitory mechanism might play a role only when set size is larger.     

Visual identification rate in search of a missing stimulus

Subjects searched for a digit which was not presented in an array of nine different digits. The identification rate was estimated from the percent correct. Well-trained subjects identified 6.5 digits with a 50-msec. exposure. This fast identification rate was expected from our procedure using a single answer, an inconsistent target, and equal probability that every digit would be a target. Naive subjects markedly improved their performance in the first two weeks of the trials which were given for a total of 24 days.     

Looking time: The effects of stimulus complexity and familiarity

During a 10 min. stimulus familiarization period, three groups of 60 Ss each received either 0, 10, or 20 sec. of familiarization on each of 30 experimental stimuli: 10 each of low, medium, and high stimulus complexity. All Ss then viewed the experimental stimuli in a second task, during which they could look at each stimulus for as long as they wished (free looking). For half the Ss in each group, free looking was administered immediately after the familiarization period. The remaining Ss received free looking 48 hr. later. The results replicated earlier research which has shown that free looking time is inversely related to stimulus familiarity, and directly related to stimulus complexity. Unlike earlier findings, the data suggested that with a 48 hr. delay between familiarization and free looking, a stimulus can, at least partially, recover from the decrement in looking time produced by 10 sec. of familiarization.     

Visual training improves perceptual grouping based on basic stimulus features

Training on visual tasks improves performance on basic and higher order visual capacities. Such improvement has been linked to changes in connectivity among mediating neurons. We investigated whether training effects occur for perceptual grouping. It was hypothesized that repeated engagement of integration mechanisms would enhance grouping processes. Thirty-six participants underwent 15 sessions of training on a visual discrimination task that required perceptual grouping. Participants viewed 20?×?20 arrays of dots or Gabor patches and indicated whether the array appeared grouped as vertical or horizontal lines. Across trials stimuli became progressively disorganized, contingent upon successful discrimination. Four visual dimensions were examined, in which grouping was based on similarity in luminance, color, orientation, and motion. Psychophysical thresholds of grouping were assessed before and after training. Results indicate that performance in all four dimensions improved with training. Training on a control condition, which paralleled the discrimination task but without a grouping component, produced no improvement. In addition, training on only the luminance and orientation dimensions improved performance for those conditions as well as for grouping by color, on which training had not occurred. However, improvement from partial training did not generalize to motion. Results demonstrate that a training protocol emphasizing stimulus integration enhanced perceptual grouping. Results suggest that neural mechanisms mediating grouping by common luminance and/or orientation contribute to those mediating grouping by color but do not share resources for grouping by common motion. Results are consistent with theories of perceptual learning emphasizing plasticity in early visual processing regions.     

Decision-bound theory and the influence of familiarity

In this article, we derive a nonparametric prediction from decision-bound theory (DBT). The crucial aspect that is tested is whether or not familiarity of a stimulus affects response time in categorization. We show that, for our design, DBT, extended with some reasonable and testable assumptions, predicts no familiarity effect. Our prediction is nonparametric in that, rather than fit a specific instantiation of general DBT, we posit only some general assumptions of this theory and derive the prediction from these assumptions. It is found that familiarity did have a strong impact on response time for at least half of our participants. We suggest that DBT is in itself incomplete and should be extended to account for the full range of available data.     

Effects of stimulus familiarity upon judged visual duration

Five experiments investigated the influence of the differences in stimulus familiarity among a dot-matrix letter, word, and nonword upon the relative judged duration of 30-msec flashes. Figure-ground contrast was manipulated by varying the number of dots comprising each display letter. With 30-msec presentations, judgments ranked subjective durations of the displays nonword > word > letter. Differences in judged duration among stimulus forms were greater when both forward and backward masking reduced recognition probability to chance level than when no mask was employed, and discriminations among masked presentations were easier for subjects. Figure-ground contrast influenced apparent duration judgments only as increases in figure-ground contrast contributed to clarity of familiarity differences among displays. The data were interpreted to indicate that differences in stimulus familiarity operate as early in visual processing as do differences in figure-ground contrast, with greater familiarity facilitating an automatic contact between a stimulus and its memory representation and therein reducing experienced duration.     

The role of stimulus familiarity in context-dependent recognition

Memory & Cognition - Investigations of context-dependent recognition, a phenomenon in which recognition is better for items studied and tested with the same rather than different context, have...     

Erratum: Stimulus familiarity and the delayed emergence of stimulus equivalence or consistent nonequivalence

The Psychological Record -     

The effects of stimulus structure and familiarity on same-different comparison

It was hypothesized that performance in the same-different comparison task is based on two modes of processing: (1) structural processes that organize the detailed parts of a stimulus into a well-formed whole, and (2)analytic processes that decompose the stimulus into features. This hypothesis was supported, but with the unexpected finding of individual differences in the mode of processing underlying “same” responses. Those Ss in the “same” condition whose reaction times were faster for symmetrical than for asymmetrical patterns supported the hypothesis for structural processes. The remaining Ss in the “same” condition, as well as all the Ss in the “different” condition, were unaffected by symmetry. These Ss supported the hypothesis for analytic processes. Although familiarity effects were obtained for both structural and analytic Ss, the rotation of the familiar patterns into an unfamiliar orientation virtually eliminated familiarity effects for the structural Ss, but left them intact for analytic Ss.     

Identification asymmetry of parafoveal stimulus pairs

Three experiments, involving 8, 10, and 10 naive participants, respectively, were conducted to examine the phenomenon in which the peripheral member of a parafoveally appearing stimulus pair is more accurately identified than the central. In the first experiment, the asymmetry was observed with stimulus pairs which had but one distinguishing feature; this could have implications for the notion that feature perturbations between stimuli provide the basis for the effect. The second experiment eliminated an explanation based upon acuity at the position of the midpoint of the stimulus configuration, since the asymmetry remains when this position is held constant. The involvement of response or decision factors seems unlikely, since the third experiment found the effect with blocked presentations involving responses to only one member of the pair. It is concluded that accounts which appeal to difficulties in separating stimuli from the overall pair configuration fail to explain more accurate identification of the peripheral stimulus when the pair is positioned at a constant distance from fixation.