L’évaluation des processus de la mémoire de reconnaissance

Recognition memory is the ability to judge whether a given stimulus has previously been encountered. Firstly modelled as a simple detection process, it rapidly became clear that it could rely on the contribution of two processes: familiarity and recollection. Some authors however consi- der that a model based on a unique process could be sufficient to account for most of the observed data. This theoretical opposition between one-process and dual-process models has fuelled a lengthy debate lasting over several decades: numerous models have been proposed, while a host of experimental approaches was developed. In this synthe- sis article, we summarize the theoretical and methodological background necessary to psychologists who would like to investigate recognition memory in their patients. In the first part, we present main models of recognition memory: the historical model based on the Signal Detection Theory (SDT), dual-process models (the Atkinson’s model, and the Yonelinas’ Dual-Process Signal Detection [DPSD] model) relying on the addition of a threshold process to model recollection, concurrent one-process models (Equal-Variance Signal Detection [EVSD] and Unequal-Variance Signal Detection [USVD] model), and the most recent Continuous Dual-Process (CDP) model. In the second part, we propose a review of the main experimental paradigms available today, presenting their specificities and limits. Classical paradigms (i.e., Yes/No and Delayed Matching to Sample paradigm) are briefly described. Emphasis is put on paradigms allowing estimations of familiarity and recollection: source memory tests, Process Dissociation Procedure (PDP), Remember- Know (RK) procedure, Receiver Operating Characteristic (ROC) paradigm and the Speed and Accuracy Boosting (SAB) procedure. How to compute familiarity and recollection is explained clearly for each paradigm, and formulas as well as references are reported. In the third part, limits and difficulties using these paradigms in patients are summarized, as well as related observations from our personal experiences.