Department of Neurobiology, Weizmann Institute of Science, Rehovot, Israel.
Nat Neurosci. 2011 Jun;14(6):791-6. doi: 10.1038/nn.2802. Epub 2011 May 8.
Animal studies of discriminative fear conditioning traditionally use stimuli that are distant in physical features and thus easily distinguished perceptually. Independently, human studies have shown that training mostly improves discrimination thresholds. We found that aversive learning actually induced an increase in discrimination thresholds in humans and that subjective aversion during conditioning predicted the individual threshold change. This counterintuitive performance deterioration occurred when using odors or sounds as aversive reinforcers and was not a result of attentional distraction or decision bias. In contrast, positive reinforcement or mere exposure induced the typically reported decrease in thresholds. Our findings indicate that aversive outcomes induce wider stimulus generalization by modulating perceptual thresholds, suggesting the engagement of low-level mechanisms. We suggest that for risk- or loss-related stimuli, less specificity could be a benefit, as it invokes the same mechanisms that respond quickly and efficiently in the face of danger.
动物辨别性恐惧条件反射的研究传统上使用物理特征上差异较大的刺激，因此很容易在感知上区分。独立的人类研究表明，训练主要提高了辨别阈值。我们发现，厌恶学习实际上会导致人类的辨别阈值增加，而条件作用期间的主观厌恶会预测个体阈值变化。当使用气味或声音作为厌恶强化物时，会出现这种违反直觉的性能恶化，而不是由于注意力分散或决策偏差所致。相比之下，正强化或仅仅暴露会导致通常报道的阈值降低。我们的研究结果表明，厌恶的结果通过调节感知阈值引起更广泛的刺激泛化，这表明了低水平机制的参与。我们认为，对于与风险或损失相关的刺激，较低的特异性可能是一种优势，因为它引发了在面对危险时快速而有效地做出反应的相同机制。
