Transferência de aprendizagem - Uso Conceitual Em Análise Comportamental e Neurociências

Abstract

Learning transfer (LT) comprises a series of phenomena related to the maintenance of behavioral repertoire in contexts other than acquisition. In behavioral analysis, this concept is related, at least, to the phenomena of irradiation of responses, operant induction, learning-set, functional generalization and generalization of stimuli, with special emphasis on the latter, recurrently adopted as a criterion of success in interventions and behavioral technologies. Despite their widespread use, there are conceptual problems with the generalization of stimuli, since their definition can be superimposed on those of resurgence and behavioral variability, and the phenomenon is often analyzed much more as a byproduct of stimulus control in the establishment of an operant than as an independent phenomenon. In the Neuroscientific literature, the re-presentation of learning in other contexts is commonly associated with different memory phenomena, correlated to distinct learning patterns, such as perceptual learning, learning stimuli-responses, motor learning and relational learning. The present review was based on the assumption that, by focusing on neurophysiological variables correlated to behavioral phenomena from the analysis of the same functional relations in experimental context it is possible to expand the understanding about a specific behavioral process. However, it was necessary for a concept to be consistently defined in behavioral analysis before looking for related mechanisms in neurosciences. In this sense, to review conceptual use in behavioral analysis (radical behaviorist) and neuroscience (biological mechanic), this research was based on the literature of the two areas produced between 2005 and 2015 to differentiate and compare the types of L.T as described in the Behavioral Analysis, in order to synthesize a concept with fewer conceptual errors and, from this definition, to identify neurophysiological mechanisms related to the same phenomenon, from the neuroscientific literature. In order to do this, the method used was the search in journals classified as A1 (maximum quality) by the QUALIS scale of the Coordination of Improvement of Higher Education Personnel (from portuguese, CAPES), a system that classified and made available these same journals for access. The key terms used in behavioral analysis were generalization, learning transfer, functional generalization, stimulus generalization, induction, irradiation, and learning-set. In neurosciences, they were perceptual learning, stimulus-response learning, motor learning, and relational learning. Two sections of analysis were proposed. In the first one, types of learning transfer in Behavior Analysis were differentiated by the 1) identification of the conceptual use in the selected research and 2) identification of the potential overlap or complementation of the defined concept. For the second section, the goal was to differentiate between types of transference of learning in neurosciences from the behavioral definitions, through the description of the correlated mechanisms found in the literature. In behavioral analysis, eight articles were found, all of which described transference of learning as generalization of stimuli and with the same concept (and associated errors) found in the established literature. As a result, the neuroscience data consisted of a presentation of how the area described the phenomena of AT during the period considered. There were 25 reviews in this area, analyzed in three categories (associative learning, perception and memory, relational learning, learning stimulus-response and memory, and synaptic plasticity). In the reviews that described the phenomena of generalization of stimuli, the use of the concept was analogous to that of the behavioral analysis. However, it was clear that the respondent dimensions of the learning processes receive special attention in the neuroscientific literature consulted, being described as belonging to the broader field of associative learning, which involves more complex respondent phenomena than the classic conditioning often described in behavioral analysis. From the literature reviewed, we derive three conclusions: 1) AT has been commonly represented as generalization of stimuli (operant and respondent) in the two literature, which share similar concepts and conceptual problems; 2) the associative learning literature describes more complex and detailed respondent phenomena than those commonly found in behavioral analysis, and such literature emerged as fundamental within different neuroscientific researches and this can be a point of methodological contact between behavioral and neuroscientific areas; 3) Other revisions, focusing on conceptual refinement, are necessary for the phenomenon of generalization of stimuli, emphasizing that one possibility would be the refinement of discrimination, a phenomenon directly related to TA as described in the recent behavioral literature.