This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. For attribution, the original author(s), title, publication source (PeerJ Preprints) and either DOI or URL of the article must be cited.
Forming generalizations from previous experiences is a complex skill, which requires a delicate coordination between several basic cognitive abilities. In menacing situations, this ability is called “fear generalization”. It allows humans to predict harmful events and is necessary for survival. Impairments of this ability may lead to overgeneralizations – a phenomenon we know from anxiety disorders. By and large, fear generalization has been studied with one type of experimental paradigm. Stimuli forming a carefully controlled perceptual similarity gradient have been the basis to quantify behavioral and neuronal “fear generalization profiles”. This paradigm has provided fruitful insights into how learnt fear generalizes to perceptually similar events. Yet, a number of findings suggest that fear generalization is more adaptive than predicted by a mechanism which is solely based on perceptual similarity. This is a proposal that aims to bring new perspectives onto fear generalization as a complex, adaptive process. I will investigate the following major hypotheses: (1) Fear generalization can be understood as the optimal result of a Bayesian inference problem. (2) In real-world conditions, fear generalization builds on conceptual knowledge rather than perceptual similarity alone. (3) Brain structures involved in fear generalization can be causally linked to modulate fear responses adaptively. To test these hypotheses, I propose use of tools including fMRI, EEG as well as intracranial electrical stimulation and LFP recordings in presurgical epilepsy patients. With the combination of these tools, the expected findings have the potential to revolutionize our understanding of fear generalization and anxiety disorders.
In this review, I am presenting a roadmap of five milestones that research in fear generalization needs to take into account in order to bring a computational understanding. I provide three empirical directions to achieve these goals.