Facing up to Faces: Changing biases in face perception to improve emotional processing in mental health

在日常生活中，准确探测和识别他人情绪的能力是极其重要的。即使在这一能力上的微小差异，如对某些情绪的相对不敏感，或对其他情绪的高度敏感，都可能导致问题。例如，抑郁症患者往往对快乐的表达不那么敏感，而对悲伤的表达更敏感。这实际上可能会导致他们对世界的看法扭曲，特别是对其他人的看法发生扭曲，从而导致他们抑郁的发生或持续。例如，如果某人对他人的快乐不那么敏感，他或她可能没有注意到别人对他们的积极回应，也没有对此做出回应，从而错失了积极社交的机会。我们已经开发出一种技术，可以微妙地改变这些情绪识别偏差。模棱两可、喜忧参半的面孔可以用来判断一个人何时不再看到一种情绪，而开始看到另一种情绪。利用这些信息，这个人可以被“训练”更倾向于从那些模棱两可的脸上看到快乐，而不是悲伤。这个培训过程只需要大约20分钟，可以在笔记本电脑上完成，原则上也可以通过网站、平板电脑或智能手机完成。在几天的重复训练中，情绪识别会发生显著的变化，因此个人更倾向于在别人的脸上看到快乐而不是悲伤。在这里，我们提出了一项研究，以回答两个问题：这一过程是否改善了人们的情绪，以及这一过程涉及哪些大脑区域？我们的试点工作非常有希望。在一组被选为高度抑郁症状的人中，四天的训练在两周后导致积极情绪的显著改善。我们想招募更大的样本，以提供更明确的结果，并收集更长时间的六周随访期的情绪数据，以确定这些影响的持久性。这将使我们在这种干预的发展中迈出关键的下一步，即从概念验证转移到临床翻译。与此同时，我们希望进行一项规模较小的研究，该研究将使用相同的训练程序，但随后将使用大脑扫描技术来确定哪些大脑区域会受到这种训练程序的影响。最后，我们将对经历过抑郁症的人进行一小部分焦点小组的研究，以找出他们是否认为像我们这样的自动训练程序可以接受抑郁症的治疗。我们的研究结果可能会产生一种新型的抑郁症治疗方法，可以与现有的治疗方法如认知行为疗法或药物一起使用。它具有相对简单和非侵入性的优势，原则上可以通过互联网或智能手机远程交付。此外，我们将使用的技术原则上可以应用于其他情绪识别似乎起作用的精神健康问题，如品行障碍、精神分裂症和成瘾。

The ability to accurately detect and recognize emotion in other people is extremely important in everyday life. Even slight differences in the ability to do this, such as a relative insensitivity to some emotions, or heightened sensitivity to other emotions, can cause problems. For example, people with depression tend to be less sensitive to expressions of happiness, and more sensitive to expressions of sadness. This may actually contribute to the onset or maintenance of their depression, by causing distortions in how they perceive the world, and in particular other people. If someone is less sensitive to happiness in others, for example, he or she may not notice when someone is responding positively to them, and fail to reciprocate this, thereby missing out on the opportunity for a positive social interaction.We have developed a technique which allows us to subtly modify these emotion recognition biases. Faces which are ambiguous, and consist of a blend of happiness and sadness, can be used to determine when a person stops seeing one emotion and starts seeing another. Using this information, that individual can then be "trained" to be more inclined to see happiness in those ambiguous faces rather than sadness. This training procedure takes only about 20 minutes, and can be done on a laptop or, in principle, via a website or on a tablet computer or smartphone. Repeated training over a few days can produce a marked change in emotion recognition, so that the individual becomes more inclined to see happiness over sadness in other people's faces. Here we are proposing a study to answer two questions: Does this procedure improve people's mood, and what brain regions are involved in this process?Our pilot work is very promising. In a group of people selected for high levels of depressive symptoms, four days of training resulted in a noticeable improvement in positive mood two weeks later. We would like to recruit a larger sample, to provide more definitive results, and collect mood data at a longer follow-up period of six weeks, to establish the longevity of these effects. This would allow us make the critical next step in the development of this intervention, which is to move from proof of concept to clinical translation. In parallel, we would like to run a smaller study which will use the same training procedure but will then use brain scanning technology to identify which brain regions are influenced by this training procedure. Finally, we will run a small focus group of people who have experienced depression, in order to find out whether they feel that an automated training procedure such as ours would be an acceptable treatment for depression.The results of our study have the potential to give rise to a new type of treatment for depression, which could be used in conjunction with existing treatments such as cognitive behavioural therapy or medication. It has the advantage of being relatively simple and non-invasive, and in principle can be delivered remotely via the internet or a smartphone. In addition, the techniques we will be using can in principle be applied to other mental health problems where emotion recognition appears to play a role, such as conduct disorder, schizophrenia and addiction.