Center for Cognitive Neuroscience, Duke University, Box 90999, Durham, NC, 27708-0999, USA.
Centre for Youth Substance Abuse Research, University of Queensland, St Lucia, Australia.
Cogn Affect Behav Neurosci. 2020 Oct;20(5):983-997. doi: 10.3758/s13415-020-00814-4.
Extinction learning is a primary means by which conditioned associations to threats are controlled and is a model system for emotion dysregulation in anxiety disorders. Recent work has called for new approaches to track extinction-related changes in conditioned stimulus (CS) representations. We applied a multivariate analysis to previously -collected functional magnetic resonance imaging data on extinction learning, in which healthy young adult participants (N = 43; 21 males, 22 females) encountered dynamic snake and spider CSs while passively navigating 3D virtual environments. We used representational similarity analysis to compare voxel-wise activation t-statistic maps for the shock-reinforced CS (CS+) from the late phase of fear acquisition to the early and late phases of extinction learning within subjects. These patterns became more dissimilar from early to late extinction in a priori regions of interest: subgenual and dorsal anterior cingulate gyrus, amygdala and hippocampus. A whole-brain searchlight analysis revealed similar findings in the insula, mid-cingulate cortex, ventrolateral prefrontal cortex, somatosensory cortex, cerebellum, and visual cortex. High state anxiety attenuated extinction-related changes to the CS+ patterning in the amygdala, which suggests an enduring threat representation. None of these effects generalized to an unreinforced control cue, nor were they evident in traditional univariate analyses. Our approach extends previous neuroimaging work by emphasizing how evoked neural patterns change from late acquisition through phases of extinction learning, including those in brain regions not traditionally implicated in animal models. Finally, the findings provide additional support for a role of the amygdala in anxiety-related persistence of conditioned fears.
灭绝学习是控制条件性威胁关联的主要手段，也是焦虑障碍中情绪失调的模型系统。最近的工作呼吁采用新方法来跟踪与灭绝相关的条件刺激（CS）表示的变化。我们应用了一种多元分析方法来分析以前收集的关于灭绝学习的功能磁共振成像数据，其中健康的年轻成年参与者（N=43；21 名男性，22 名女性）在被动导航 3D 虚拟环境时遇到了动态蛇和蜘蛛 CS。我们使用表示相似性分析来比较主体内从恐惧获得的后期到灭绝学习的早期和晚期阶段的冲击强化 CS（CS+）的体素激活 t 统计映射。这些模式在先前确定的兴趣区域中从早期到晚期灭绝变得越来越不相似：前扣带皮质下和背侧前部、杏仁核和海马体。全脑搜索灯分析在岛叶、中扣带皮质、腹外侧前额叶皮质、体感皮质、小脑和视觉皮质中发现了类似的发现。高状态焦虑减弱了杏仁核中与 CS+模式相关的灭绝变化，这表明存在持久的威胁表示。这些影响都没有扩展到未强化的控制线索，在传统的单变量分析中也没有明显。我们的方法通过强调从晚期获得到灭绝学习阶段的诱发神经模式如何变化，包括传统上与动物模型无关的大脑区域，扩展了以前的神经影像学工作。最后，这些发现为杏仁核在与焦虑相关的条件恐惧持续存在中的作用提供了额外的支持。
