Human learning/Implicit learning
Friday, September 30th, 2011 [17:20 - 19:20]
PS_1.061 - Positive feedback at the end of an unsolvable test (but not at the beginning) makes students believe they did well
Orgaz, C. , Matute, H. & Vadillo, M. A.
University of Deusto. Bilbao, Spain
Recent research has shown that providing easy items at the beginning of a test makes students believe they do better in it. This could be due to the earlier items being easier, but it could also be due to students’ subjective, auto-administered positive feedback during the first part of the task. In two experiments we kept the test items constantly unsolvable and manipulated whether positive or negative feedback was presented during the first or the second part of the test. In Experiment 1, the students showed recency. That is, they thought that they did better when they received positive feedback during the last part of the test and they thought that they did worst when they received negative feedback during the last part of the test. Experiment 2 replicated this result and showed that this recency effect vanishes after one hour, which is consistent with the abundant literature on recency-to-primacy shifts in various experimental paradigms. Thus, well-known learning and memory effects such as recency, primacy, and spontaneous recovery affect not only the contents of memory but also perceived self-efficacy.
PS_1.062 - On the interplay between information, behavior and judgments in the illusion of causality
Yarritu Corrales, I. & Matute, H.
Universidad de Deusto, Bilbao, Spain
In two experiments we test how the presentation of biased information before the actual experiment can induce causal illusions. The potential cause was a fictitious medicine and the effect was recovery from a health crisis. Group Induced was pre-informed about the number of patients in which the effect (recovery from the crisis) occurred when they took the medicine. Group Warned was pre-informed about the number of patients that had recovered without taking the medicine. Both groups were then presented with 100 trials (one per patient). In Experiment 1 the patient could have taken the medicine or not and could feel better or not. The drug-recovery contingency was zero. However, participants in Group Induced showed an illusion that the drug was effective. In Experiment 2 the medicine had to be administered (or not) by the participant (rather than the patient). In this case, our pre-information manipulation resulted in participants in Group Induced administering the drug significantly more often than participants in Group Warned. This in turn produced a larger number of coincidences between the drug and the recovery and a larger illusion. In other words, the illusion of causality was mediated by the behavioral increase produced by the pre-information manipulation.
PS_1.063 - The representation and activation of evaluative connotations - behavioral and ERP correlates of the evaluative priming effect
Schmitz, M. & Wentura, D.
A long-standing debate in the field of cognition and emotion deals with the question of how evaluative connotations are represented in semantic memory. For this purpose, evaluative priming (i.e., processing valent targets following either valence-congruent primes or incongruent primes) with non-evaluative tasks (e.g., semantic categorization task) was considered a useful paradigm. However, published results are still inconsistent: congruence effects, null results, and even incongruence effects were found. We put forward the assumption that congruence effects may be traced back to the target-encoding facilitation by a valence-congruent prime, whereas incongruence effects may be attributed to the prime-activation maintenance by a valence-congruent target and a subsequent increase of response conflicts. To test this assumption, we used the semantic categorization task (person vs. animal) and a negative stimulus-onset asynchrony between prime and target to minimize target-encoding facilitation and to maximize prime-maintenance. As expected, response priming effects (i.e., faster responses if prime and target are response compatible compared to incompatible) were only found given valence-congruence of prime and target thereby indicating that only evaluatively congruent primes are activated sufficiently to interfere with the target response. ERP correlates (N2, lateralized readiness potential) corroborate the behavioral findings. Implications for the representation of evaluation are discussed.
PS_1.064 - Comparing effects of orientation and structure in implicit learning of serial actions with three most recent methods
Pasquali, A. 1, 2 , Gaillard, V. 1 & Cleeremans, A. 1
1 Consciousness, Cognition, & Computation Group. Université Libre de Bruxelles. Brussels, Belgium.
2 Neurogenics Research Unit. Adam-NX. Solaro, France.
We hereby examine the influence of orientation to learn and stimulus structure on performance in the serial reaction time task - implicit learning paradigm involving sequentially structured visual stimuli. In six conditions inducing different interactions of these two factors, we assess the nature of the knowledge acquired by participants. Using the methods of PDP - comparison of transfer contrasts with post-test recognition scores-, ANCOVA - comparing reaction times and their covariance with the material's probability structure - and SDT - analysis of confidence judgements in recognition -, all three analyses converge and confirm weak explicit learning compensated by implicit influences when the probability of the material is intermediate between deterministic and pseudorandom. However, effects of orientation for that same material condition are only captured by the second and third methods, showing that participants explicitly engaged in the learning process effectively acquire a large amount of explicit knowledge, but of poor quality. Thus implicit learning would automatically occur in relation with the material structure, whereas explicit learning would depend on both the orientation to learn and an ability to match the material structure with the coherent mental models allowing control of the acquired knowledge.
PS_1.065 - Cognitive control of sequential knowledge in children and children with autism
Pichon, C. , Kissine, M. & Destrebecqz, A.
Free University of Brussels
The card-sorting test is a well-known test of cognitive control. In this task, subjects have to classify cards according to one of two dimensions (i.e., either the color or the shape of the stimuli). Based on explicit instructions, children are unable to switch rules before the age of 4. Recently, Bremner et al. (2007) found that 2-years-old children acquire the ability to manipulate and inhibit mental representations when learning is implicit. These results suggest that inhibition seems to develop in early stages of life. Despite the proven difficulty of autistic children with executive functions (including inhibition), implicit learning remains largely unexplored in this population. Hence, in the present study, we trained children with a deterministic sequence of six elements. We compared the ability of normal and autistic children to learn flexibly and to reproduce a visual sequence (inclusion condition) or to produce a different sequence (exclusion condition). Our results show that children with autism are able to learn a visual sequence and also to inhibit the production of the first sequence. These results are discussed regarding the potential importance of implicit learning for cognitive control in autism.
PS_1.066 - Perceptual sequence learning without eye movements?
Coomans, D. , Deroost, N. , Van den Bussche, E. & Soetens, E.
Department of Cognitive Psychology. Vrije Universiteit Brussel. Brussels, Belgium.
We examined whether pure perceptual sequence learning occurred in a serial reaction time task in which eye movements were avoided. Participants had to respond to the identity of a target letter pair (“XO” required a left button press, “OX” a right button press,) appearing in one of four locations between similar distractor letter pairs (“YQ” or “QY”). While target identity, and therefore manual responses, changed randomly, target location followed a fixed sequence. To avoid eye movements as much as possible, (1) the letter pairs appeared around a fixation cross with a small visual angle and (2) stimulus duration was only 100 ms, a period too short to allow eye movements. At the end of the training, a block in which the trained fixed sequence was replaced by a new fixed sequence was inserted to be able to assess learning. The results showed that the perceptual location sequence was learned in this experiment, as participants responded slower in this ‘new sequence’ block than in the surrounding ‘old sequence’ blocks. Consequently, pure perceptual sequence learning does not rely on eye movements and cannot be considered as a sort of motor learning.