Le renforcement signalé chez l'animal : renforcement positif
Institution:
Montpellier 3Disciplines:
Directors:
Abstract EN:
The signaled reinforcement is a typical pavlovina conditioning procedure in which the reinforcer is preceded by the presentation of a signal. In instrumental conditioning with signaled reinforcement, the reinforced response lend to the presentation of a stimulus which predicts the reinforcer. Such an experimental paradigm represents an excellent test of the associative processes which develop between responses, signal and reinforcer and, therefore, of the theoretical models describing those processes in animals. The experimental results confirm the validity of recent models of conditioning (e. G. The wagner-rescorla model) as opposed to the traditional s-r interpretations. The results also suggest some constraints the future models should support, especially in relation to the functioning of neural networks or to inferential information processing.
Abstract FR:
Le terme de renforcement signale designe des procedures de conditionnement dans lesquelles l'agent de renforcement est regulierement precede d'un signal. Caracteristiques des contingences pavloviennes, ces procedures peuvent etre egalement employees dans un contexte instrumental ou les reponses renforcees entrainent la presentation du signal qui predit le renforceteur. Un tel paradigme experimental constitue alors un test crucial des processus associatifs reliant la reponse et le signal au renforcateur et, par consequent, des modeles qui tentent de decrire ces processus chez l'animal. Les resultats experimentaux permettent de confirmer la pertinence des modeles recents du conditionnement (e. G. Le modele de wagner et rescorla), par rapport aux modeles s-r qui les ont precedes, et revelent egalement les contraintes auxquelles devront repondre les futurs modeles cognitifs du conditionnement associatif, inspires du fonctionnement des reseaux de neurones formels ou des processus inferentiels du traitement de l'information.