In anxiety disorders, fear does not remain specific to the stimulus that was originally linked to danger but generalizes to a broad set of stimuli, resulting in a snowballing of complaints. Even though generalization is a well-established phenomenon, the precise mechanisms underlying the spreading of fear remain unclear. The project aims to elucidate these mechanisms, building on our previous work, demonstrating that fear generalization can be experimentally induced through Pavlovian learning processes and that both intra- and interindividual variations in perception strongly affect fear generalization via various manners.The aim of the current project is to develop a computational model of response generalization that disentangles the distinct contribution of perceptual, learning and memory processes both in healthy volunteers and anxiety patients. These effects will be studied at both the self-report level as well as the psychophysiological level (e.g., skin conductance). Apart from the development, the task is to design innovative and effective empirical tests of the model to rigorously test and validate it before implementing it in a clinical context.