Multidimensional visual statistical learning

Recent studies of visual statistical learning (VSL) have demonstrated that statistical regularities in sequences of visual stimuli can be automatically extracted, even without intent or awareness. Despite much work on this topic, however, several fundamental questions remain about the nature of VSL. In particular, previous experiments have not explored the underlying units over which VSL operates. In a sequence of colored shapes, for example, does VSL operate over each feature dimension independently, or over multidimensional objects in which color and shape are bound together? The studies reported here demonstrate that VSL can be both object-based and feature-based, in systematic ways based on how different feature dimensions covary. For example, when each shape covaried perfectly with a particular color, VSL was object-based: Observers expressed robust VSL for colored-shape sub-sequences at test but failed when the test items consisted of monochromatic shapes or color patches. When shape and color pairs were partially decoupled during learning, however, VSL operated over features: Observers expressed robust VSL when the feature dimensions were tested separately. These results suggest that VSL is object-based, but that sensitivity to feature correlations in multidimensional sequences (possibly another form of VSL) may in turn help define what counts as an object.     

Implicit visual learning and the expression of learning

Although the existence of implicit motor learning is now widely accepted, the findings concerning perceptual implicit learning are ambiguous. Some researchers have observed perceptual learning whereas other authors have not. The review of the literature provides different reasons to explain this ambiguous picture, such as differences in the underlying learning processes, selective attention, or differences in the difficulty to express this knowledge. In three experiments, we investigated implicit visual learning within the original serial reaction time task. We used different response devices (keyboard vs. mouse) in order to manipulate selective attention towards response dimensions. Results showed that visual and motor sequence learning differed in terms of RT-benefits, but not in terms of the amount of knowledge assessed after training. Furthermore, visual sequence learning was modulated by selective attention. However, the findings of all three experiments suggest that selective attention did not alter implicit but rather explicit learning processes.     

Modeling cross-situational word-referent learning: prior questions

Both adults and young children possess powerful statistical computation capabilities--they can infer the referent of a word from highly ambiguous contexts involving many words and many referents by aggregating cross-situational statistical information across contexts. This ability has been explained by models of hypothesis testing and by models of associative learning. This article describes a series of simulation studies and analyses designed to understand the different learning mechanisms posited by the 2 classes of models and their relation to each other. Variants of a hypothesis-testing model and a simple or dumb associative mechanism were examined under different specifications of information selection, computation, and decision. Critically, these 3 components of the models interact in complex ways. The models illustrate a fundamental tradeoff between amount of data input and powerful computations: With the selection of more information, dumb associative models can mimic the powerful learning that is accomplished by hypothesis-testing models with fewer data. However, because of the interactions among the component parts of the models, the associative model can mimic various hypothesis-testing models, producing the same learning patterns but through different internal components. The simulations argue for the importance of a compositional approach to human statistical learning: the experimental decomposition of the processes that contribute to statistical learning in human learners and models with the internal components that can be evaluated independently and together.     

Effects of postresponse visual stimuli on visual discrimination learning in the rhesus monkey

To learn more about the mechanism (or mechanisms) involved with postresponse stimulus processing during discrimination learning, a series of studies was conducted with monkeys to determine why the combined relevant and irrelevant stimuli impair learning more than irrelevant stimuli appearing alone. It was found that: (a) the greater size and complexity of the combination of stimuli were not responsible for the greater deficit, while the presence of the relevant stimuli (S^D and S^Δ) within the stimulus combination apparently was; (b) the more similar the postresponse irrelevant stimuli were to the relevant stimuli the greater the deficit that resulted; and (c) monkeys that had earlier learned to discriminate the relevant and irrelevant features of a combination showed no learning impairment when this same stimulus combination was later presented after the response during a new learning problem. These results were interpreted as evidence that: (1) processes associated with learning a discrimination problem do not end with the execution of a choice response; (2) postresponse stimuli produce greater impairment in discrimination learning when they are distorted versions of the relevant stimuli; and (3) the impairment resulting from postresponse irrelevant stimuli occurs primarily when this misinformation is processed and misperceived as being relevant to learning the discrimination problem.     

Touchscreen-enhanced visual learning in rats

The efficiency of traditional levers and of modern touchscreen technology for training rats on a computerized visual discrimination was studied in a series of observations. When compared with a lever-based discrimination procedure, the use of touchscreens supported the faster development of signal tracking behavior and acquisition of a two-stimulus simultaneous visual discrimination. It did not affect the final level of accuracy. Factors related to spatial proximity of the responses with the stimuli, sign-tracking, and increased ease of touchscreen motor responses were suggested as possible reasons for the touchscreen training advantage. This increased efficiency allows large numbers of animals to be tested quickly, a necessary requirement for studies involving genetic and physiological interventions.     

Association learning and visual discrimination

The effect of learned stimulus associations on visual discrimination was measured in four experiments. The stimuli were visual shapes which had been scaled for similarity. Two shapes were selected as discriminanda, and each S was pretested and posttested for discrimination of these briefly presented simultaneous pairs of “same” and “different” shapes. During the training, each discriminandum was paired with another simultaneously presented associated shape on a paired-associate response-learning task. The two associated shapes were very similar, intermediate in similarity, or very dissimilar. There was more improvement in posttest discrimination following training with dissimilar associates. We conclude that learned stimulus associations affect visual discrimination.     

Implicit learning of ignored visual context

Humans process a visual display more efficiently when they encounter it for a second time, showing learning of the display. This study tests whether implicit learning of complex visual contexts depends on attention. Subjects searched for a white target among black and white distractors. When the locations of the target and the attended set (white distractors) were repeated, search speed was enhanced, but when the locations of the target and the ignored set (black distractors) were repeated, search speed was unaffected. This suggests that the expression of learning depends on attention. However, during the transfer test, when the previously ignored set now was attended, it immediately facilitated performance. In contrast, when the previously attended set now was ignored, it no longer enhanced search speed. We conclude that the expression of visual implicit learning depends on attention but that latent learning of repeated information does not.     

The automaticity of visual statistical learning

The visual environment contains massive amounts of information involving the relations between objects in space and time, and recent studies of visual statistical learning (VSL) have suggested that this information can be automatically extracted by the visual system. The experiments reported in this article explore the automaticity of VSL in several ways, using both explicit familiarity and implicit response-time measures. The results demonstrate that (a) the input to VSL is gated by selective attention, (b) VSL is nevertheless an implicit process because it operates during a cover task and without awareness of the underlying statistical patterns, and (c) VSL constructs abstracted representations that are then invariant to changes in extraneous surface features. These results fuel the conclusion that VSL both is and is not automatic: It requires attention to select the relevant population of stimuli, but the resulting learning then occurs without intent or awareness.     

Multiple effects of visual format on TV news learning

Two experiments were conducted on edited TV newscast sequences to clarify effects of film accompaniment on learning from heard news text. In Experiment 1,150 British subjects viewed a sequence with either film format throughout or alternating film and ‘talking head’ format between items. Those items that were presented by ‘talking heads’ in the mixed sequence were learned better with film format, in which the heard text was accompanied by appropriate moving pictures. However, no effect of uniform context was found on the remaining items. In Experiment 2, 91 German subjects viewed one of four versions of a bulletin, one with ‘talking head’, one with film throughout, the other two having complementary mixed-format patterns. Besides confirming a beneficial effect of film presentation over ‘talking head’ which accords with the findings of all studies of learning from material in the form of national network newscasts, the results showed an impairing effect of uniform visual format. This can explain ‘contradictory’ findings, notably with atypical test material.     

Temporal constraints on visual learning: a computational model

Given a constant stream of perceptual stimuli, how can the underlying invariances associated with a given input be learned? One approach consists of using generic truths about the spatiotemporal structure of the physical world as constraints on the types of quantities learned. The learning methodology employed here embodies one such truth: that perceptually salient properties (such as stereo disparity) tend to vary smoothly over time. Unfortunately, the units of an artificial neural network tend to encode superficial image properties, such as individual grey-level pixel values, which vary rapidly over time. However, if the states of units are constrained to vary slowly, then the network is forced to learn a smoothly varying function of the training data. We implemented this temporal-smoothness constraint in a backpropagation network which learned stereo disparity from random-dot stereograms. Temporal smoothness was formalized with the use of regularization theory by modifying the standard cost function minimised during training of a network. Temporal smoothness was found to be similar to other techniques for improving generalisation, such as early stopping and weight decay. However, in contrast to these, the theoretical underpinnings of temporal smoothing are intimately related to fundamental characteristics of the physical world. Results are discussed in terms of regularization theory and the physically realistic assumptions upon which temporal smoothing is based.     

Constraint on the semantic flexibility in visual statistical learning

We examined whether the expression of visual statistical learning (VSL) is flexible at the superordinate-categorical level. In the familiarization phase, participants viewed a sequence of line drawings. In the test phase, participants observed two test sequences (statistically related triplets versus unrelated foils) that consisted of the same objects as those presented during the familiarization phase (the same condition) or different objects that shared the same categorical information with drawings during the familiarization (the different condition). They then decided whether the first or the second sequence was more familiar. The results of Experiment 1 showed greater familiarity above chance levels for statistically related triplets only in the same condition. In Experiment 2, even where the word stimuli representing each superordinate-level category were included in the test phase (the categorical condition), the results showed VSL only in the same condition. Our findings suggest that the semantic flexibility of VSL is limited to the basic-level category.     

Age differences in visual statistical learning

Recent work has shown that older adults' lessened inhibitory control leads them to inadvertently bind co-occurring targets and distractors. Although this hyper-binding effect may lead to the formation of more superfluous associations, and thus greater interference at retrieval for older adults, it may also lead to a greater knowledge of information contained within the periphery of awareness. On the basis of evidence that younger adults only show learning for statistical regularities contained within attended information, we asked whether older adults may also show learning for regularities contained within to-be-ignored information. Older and younger adults viewed a series of red and green pictures and performed a 1-back task on one of the colors. Unbeknownst to participants, both color streams were organized into triplets that occurred sequentially. Implicit memory for the triplets from both the attended and ignored streams was tested using a speeded detection task. Replicating previous work, younger adults demonstrated more learning for the attended triplets than the unattended triplets. Older adults, however, demonstrated similar learning for both the attended and ignored triplets, suggesting that contrary to popular belief, they may actually know more than younger adults about the world around them, including how seemingly irrelevant events co-occur. (PsycINFO Database Record (c) 2012 APA, all rights reserved).     

Perceptual learning in visual category acquisition

The attribute structure of a set of dot patterns was studied by having subjects segment (parse) the dots of each pattern into parts or subunits by drawing circles around groups of dots from each pattern. These parsing data were obtained for subjects who had no prior experience with the patterns and for subjects who had previously learned to identify the patterns as members of one of four categories. Analyses of the parsing data indicated that category learning increased the salience of large subunits that were similar in orientation for patterns that were members of the same category. This evidence for perceptual learning was obtained even when the category training procedure required learning to identify the patterns individually, suggesting that attribute abstraction and item learning are not incompatible. It was also obtained without an increase in overall intersubject agreement. The latter result led us to question the usefulness of intersubject agreement as an index of category knowledge.     

Touchscreen-enhanced visual learning in rats.

The efficiency of traditional levers and of modern touchscreen technology for training rats on a computerized visual discrimination was studied in a series of observations. When compared with a lever-based discrimination procedure, the use of touchscreens supported the faster development of signal tracking behavior and acquisition of a two-stimulus simultaneous visual discrimination. It did not affect the final level of accuracy. Factors related to spatial proximity of the responses with the stimuli, sign-tracking, and increased ease of touchscreen motor responses were suggested as possible reasons for the touchscreen training advantage. This increased efficiency allows large numbers of animals to be tested quickly, a necessary requirement for studies involving genetic and physiological interventions.