Learning and memory

The effect of the subjective congruency of linguistic labels on boundary extension

Kentaro Inomata

Boundary extension (BE) is a phenomenon wherein participants remember seeing more of a scene than was actually shown to them (Intraub & Richardson, 1989, Journal of Experimental Psychology: Learning, Memory, and Cognition, 15(2), 179-187). In previous studies, boundary extension was investigated by using pictures that contain a main object. Boundary extension was observed when linguistic labels were presented with pictures that did not have a main object (Inomata, 2011, iPerception, 2, 958). This suggests that semantic understanding of a scene is necessary to elicit boundary extension. However, it is possible that linguistic labels increase visual information that yields boundary extension. Therefore, this research aims to determine whether a congruency between abstract pictures and linguistic labels elicit boundary extension. Participants were presented with 12 pictures with either a congruent label, an incongruent label, or without any label. After these pictures had been presented, the participants were required to rate the differences between the test pictures and the remembered pictures. Subsequently, they rated the semantic congruency between the pictures and labels. The test pictures were identical to the presented pictures. Results revealed that boundary extension occurred only when labels were congruent with pictures subjectively.

Always trust the congenitally blind man

Achille Pasqualotto and Michael J Proulx

Human memory has been found to be easily influenced by the nature of the learnt material, such as when lists of semantically related words are learnt. In this case we are likely to report words that were not in the list, but which are related to them, namely 'false memories'. Here we investigated the role of visual experience on false memory generation. Previous studies suggested that congenitally blind people possess superior verbal-memory and verb-generation skills. Additionally, visual cortex was active during these memory tasks. We tested congenitally, late blind, and sighted participants by adopting the false memory paradigm. Results showed that congenitally blind participants were less affected by false memories than the other two groups. Moreover, congenitally blind participants reported more words from the original. That is, they might 'stick'closely to the words they heard rather than being 'fooled'by semantic associations across them. The extensive structural and functional reorganisation in congenitally blind's brain, especially across visual areas, might provide the neural bases for these superior verbal memory skills.

Short-term changes in expectations flexibly modulate the appearance of ambiguous stimuli

Marcus Rothkirch, Katharina Schmack, Richard Murphy and Philipp Sterzer

Expectations can shape the way we perceive the world. Here, we tested whether short-term changes in expectations modulate the perception of ambiguous visual stimuli. Participants performed a probabilitic learning task. High or low tones were associated with right- or leftward rotation of sphere stimuli that were disambiguated by disparity cues. Randomly interspersed were ambiguous versions of the sphere without disparity cues but indistinguishable in appearance from the unambiguous stimuli. The association between tones and rotation direction changed unpredictably in a probabilistic fashion every 16 to 32 trials. Participants indicated rotation direction and their confidence about the reported percept on every trial. Stimulus-response mapping was randomised across trials to ensure that button presses were uncorrelated with participants' perception. In unambiguous trials, reaction times were faster matching the predominant tone-rotation association in a given block compared to non-matching stimuli. Most importantly, ambiguous stimuli were more frequently perceived as rotating in the direction matching the currently predominant association. This effect was also observed when only considering trials with the highest confidence level and correlated with the effect of reaction times in unambiguous trials. These results indicate that even rapidly changing prior expectations strongly influence the contents of our visual experience.

Using recollection and familiarity to investigate view generalization in object recognition

William Hayward

Objects can sometimes be recognized across changes in view with relatively trivial costs to performance, but at other times even a small change in viewpoint can result in large recognition impairments. One way to explain this apparent contradiction is whether the memory experience of the object is based upon the entire encoding episode (also termed 'recollection') or whether it is based on the object itself, devoid of context (called 'familiarity'). In two experiments, different techniques were used to separate recollection and familiarity judgments for rotated objects. Participants studied familiar objects and then had to perform recognition judgments for the objects across changes in rotation. In Experiment 1, participants had to remember the temporal context of the objects, and the process-dissociation procedure was used to get independent estimates of recollection and familiarity. In Experiment 2, the 'Remember-Know' procedure was used to assess participants'memory experiences for each object, and again separate estimates of recollection and familiarity were obtained. Both studies showed that recollection was influenced by the similarity between views at study and test, but familiarity was viewpoint invariant. These results show that information is routinely encoded from objects to support both view-specific and view-invariant patterns of recognition.

Nicotine facilitates memory consolidation in perceptual learning

Anton Beer, Devavrat Vartak and Mark Greenlee

The neurotransmitter acetylcholine is known to enhance performance in perceptual tasks. However, little is known about the role of the cholinergic system for implicit memory consolidation processes in perceptual learning. Here we compared two groups of non-smoking men who learned the same visual texture discrimination task (TDT). Following the TDT training, one group received chewing tobacco containing nicotine for one hour. The other group received a similar tasting control substance without nicotine. Participants were randomly assigned to the groups and blind to the substance. Electroencephalographic (EEG) recordings during substance consumption showed a reduced alpha activity and P300 latency in the nicotine group compared to the control group. When re-tested on the TDT the next day, both groups responded more accurately and more rapidly than during training. These improvements were specific to the retinal location and orientation of the texture elements of the TDT suggesting that learning involved early visual cortex. A group comparison showed that learning effects were significantly more pronounced in the nicotine group than in the control group. The EEG findings suggest that oral consumption of nicotine enhances the efficacy of nicotinic acetylcholine receptors. Our findings further suggest that enhanced efficacy of acetylcholine receptors facilitates the consolidation processes involved in perceptual learning that take place following task completion.

Individual Differences in Semantic and Spatial Scene Gist Processing

Anne Hillstrom and Davina Patel

Scene gist processing is influenced by individual differences in speed of perceptual processing [Võ and Schneider, 2010, Visual Cognition, 18(2): 171-200]. As both spatial and semantic content is supposed to be influential during scene gist processing, the current study explored whether verbal and spatial skill differences would independently affect scene gist processing. 75 university staff and students participated. Short versions of the Alice Heim test (AH5) assessed spatial and verbal skills. Judgements of whether or not sentences matched pictures and judgments of the relative spatial location of probe objects were carried out on separate sets of photographs. For half the trials in each task, a 250 ms preview of the picture (without sentences or probe objects) preceded the judged picture. Spatial and verbal skills were highly correlated, so partial correlations were used to measure skill to task relationships. Spatial skill correlated significantly with preview benefit in the spatial task, but not with spatial task performance itself, nor with preview benefit or base performance on the sentence task. Verbal skill did not correlate significantly with anything. The two preview benefits did not correlate significantly with each other. Thus, spatial skill but not verbal skill affects early processing of scenes.