Overgeneralized fear (OGF), or indiscriminate fear responses to signals of threat and nonthreat, is a well-studied cognitive mechanism in human anxiety. Anxiety-related OGF has been studied primarily through fear-learning paradigms and conceptualized as overly exaggerated learning of cues signaling imminent threat. However, the role of safety learning in OGF has not only received much less empirical attention but has been fundamentally conceptualized as learning about the absence of threat rather than the presence of safety. As a result, the relative contributions of exaggerated fear learning and weakened safety learning to anxiety-related OGF remain poorly understood, as do the potentially unique biological and behavioral underpinnings of safety learning. The present review outlines these gaps by, first, summarizing animal and human research on safety learning related to anxiety and OGF. Second, we outline innovations in methods to tease apart unique biological and behavioral contributions of safety learning to OGF. Lastly, we describe clinical and treatment implications of this framework for translational research relevant to human anxiety.

过度泛化恐惧 (OGF)，或对威胁和非威胁信号的不分青红皂白的恐惧反应，是人类焦虑的一种经过充分研究的认知机制。与焦虑相关的 OGF 主要通过恐惧学习范​​式进行研究，并被概念化为过度夸大学习暗示迫在眉睫的威胁的线索。然而，安全学习在 OGF 中的作用不仅受到较少的实证关注，而且从根本上被概念化为学习威胁的不存在而不是安全的存在。因此，人们对夸大的恐惧学习和减弱的安全学习对焦虑相关 OGF 的相对贡献仍然知之甚少，安全学习的潜在独特生物学和行为基础也是如此。本审查概述了这些差距，首先，总结了与焦虑和 OGF 相关的安全学习的动物和人类研究。其次，我们概述了方法上的创新，以梳理安全学习对 OGF 的独特生物学和行为贡献。最后，我们描述了该框架对与人类焦虑相关的转化研究的临床和治疗意义。