Today's post features Christopher Mole, Associate Professor of Philosophy, and Chair of the Programme in Cognitive Systems at the University of British Columbia. In this post he outlines a recent debate concerning the causes of intrusive memory in post traumatic stress.
Christopher has published on this topic in his papers "A methodological flaw in 'The neural basis of flashback formation: the impact of viewing trauma'" (Psychological Medicine, 46(8), p. 1785) and "Causes and correlates of intrusive memory: a response to Clark, MacKay, Holmes and Bourne." (Psychological Medicine, 46(15), p. 3255).
People often witness events that cause post-traumatic stress in others, without suffering from such stress themselves. Because of that, it has been supposed that the explanation of PTSD should not be sought in the experience of the traumatizing event, but in disruptions to the processes that occur in the days and weeks after it: processes of grieving, coping, and coming to terms.
That supposition is questionable, and an explanation of the fact that only some witnesses to an event will be traumatized by it does not require it. That fact could instead be explained in lots of other ways. Some disgusting event might traumatize Jones, but not Smith, because Jones has just eaten. Some violent event might traumatize Smith, but not Jones, because Smith better understands its consequences. And so on, for any number of other reasons; some to do with differences of situation, some with differences of personality, some with cognitive factors, and some with emotional ones. There is no need to suppose that the untraumatized person has better coping capabilities than the traumatized one.
The several factors that might contribute to post-traumatic stress at the time of the traumatizing event have seldom been studied, partly because the assumption that PTSD is a disorder of subsequent coping has long been the orthodoxy, and partly because such peri-traumatic factors are hard to observe in the lab. A series of experiments has begun to rectify this. The experiments use fMRI to observe the brains of people who are watching films of events such as eye-surgeries and road traffic accidents, and who are therefore undergoing something analogous to a traumatizing experience.
These participants are then asked to keep a diary, which enables the researchers to ascertain which of the upsetting films are recollected intrusively. Such experiments enable us to identify factors at the time of a traumatizing experience that contribute to its PTSD-like recollection. They have made a crucial contribution to our understanding of the ways in which peri-traumatic factors might contribute to PTSD. (A review of this paradigm can be found here.)
These experiments are undoubtedly of the first importance, but I have a qualm about one way in which their results have been interpreted. To bring the qualm into view, notice that, whereas some of the possible peri-traumatic contributors to PTSD are emotional (Smith is frightened), some are not (Jones has just eaten). One might naïvely suppose that it is always the emotional factors doing the work. Jones having eaten contributes to his trauma only to the extent that it makes him susceptible to an emotion; in this case disgust. Smith’s understanding of the violence contributes to her trauma, only to the extent that it makes her susceptible to an emotion; in this case fear. According to this naïve view, the immediate causes of trauma-symptoms are emotional through and through. That view may ultimately prove to be mistaken. My qualm is that it has been dismissed too quickly. The data that are thought to make trouble for it do not.
What those data show is that, if we are trying to use fMRI to predict whether an experience is being processed in a way that is likely to lead to PTSD-like symptoms, then useful information on which to base our prediction can be found in the activation pattern of brain regions that are traditionally implicated in the non-emotional processing of language, memory, and cognition (see here). Areas that are traditionally implicated in emotion do also carry predictively-useful information. The machine-learning methods that are used to locate these information-bearing patterns do not allow for quantifiable comparisons as to which areas contain more information, relative to their size, but what is clear is that patterns of activation in non-emotional regions predict intrusive memory.
This is a precarious basis for an inference about the causes of the intrusive memories whose occurrence is predicted. In a couple of short pieces I have tried to draw out the reasons for this (see the readings below), but the basic point can be indicated with an analogy: When online advertisers are trying to predict whether you might soon be having a baby, they find useful information on which to base their predictions in the patterns of your internet searches. Those internet searches are not the cause of the baby’s occurrence. The lesson to draw from this is that the richest sources of non-intrusively gatherable information about a thing are not typically the causes of that thing. We should therefore be wary of supposing that we identify the causes of PTSD-like symptoms by identifying sources of information about the occurrence of those symptoms.
With that in mind, I think that questions about the role of emotion in the cause of PTSD should be regarded as open. This is not to deny the importance of the work that has brought these questions to light, and that has drawn our attention to the contributions made to PTSD at the time of the traumatizing experience.
Christopher has published on this topic in his papers "A methodological flaw in 'The neural basis of flashback formation: the impact of viewing trauma'" (Psychological Medicine, 46(8), p. 1785) and "Causes and correlates of intrusive memory: a response to Clark, MacKay, Holmes and Bourne." (Psychological Medicine, 46(15), p. 3255).
People often witness events that cause post-traumatic stress in others, without suffering from such stress themselves. Because of that, it has been supposed that the explanation of PTSD should not be sought in the experience of the traumatizing event, but in disruptions to the processes that occur in the days and weeks after it: processes of grieving, coping, and coming to terms.
That supposition is questionable, and an explanation of the fact that only some witnesses to an event will be traumatized by it does not require it. That fact could instead be explained in lots of other ways. Some disgusting event might traumatize Jones, but not Smith, because Jones has just eaten. Some violent event might traumatize Smith, but not Jones, because Smith better understands its consequences. And so on, for any number of other reasons; some to do with differences of situation, some with differences of personality, some with cognitive factors, and some with emotional ones. There is no need to suppose that the untraumatized person has better coping capabilities than the traumatized one.
The several factors that might contribute to post-traumatic stress at the time of the traumatizing event have seldom been studied, partly because the assumption that PTSD is a disorder of subsequent coping has long been the orthodoxy, and partly because such peri-traumatic factors are hard to observe in the lab. A series of experiments has begun to rectify this. The experiments use fMRI to observe the brains of people who are watching films of events such as eye-surgeries and road traffic accidents, and who are therefore undergoing something analogous to a traumatizing experience.
These participants are then asked to keep a diary, which enables the researchers to ascertain which of the upsetting films are recollected intrusively. Such experiments enable us to identify factors at the time of a traumatizing experience that contribute to its PTSD-like recollection. They have made a crucial contribution to our understanding of the ways in which peri-traumatic factors might contribute to PTSD. (A review of this paradigm can be found here.)
These experiments are undoubtedly of the first importance, but I have a qualm about one way in which their results have been interpreted. To bring the qualm into view, notice that, whereas some of the possible peri-traumatic contributors to PTSD are emotional (Smith is frightened), some are not (Jones has just eaten). One might naïvely suppose that it is always the emotional factors doing the work. Jones having eaten contributes to his trauma only to the extent that it makes him susceptible to an emotion; in this case disgust. Smith’s understanding of the violence contributes to her trauma, only to the extent that it makes her susceptible to an emotion; in this case fear. According to this naïve view, the immediate causes of trauma-symptoms are emotional through and through. That view may ultimately prove to be mistaken. My qualm is that it has been dismissed too quickly. The data that are thought to make trouble for it do not.
What those data show is that, if we are trying to use fMRI to predict whether an experience is being processed in a way that is likely to lead to PTSD-like symptoms, then useful information on which to base our prediction can be found in the activation pattern of brain regions that are traditionally implicated in the non-emotional processing of language, memory, and cognition (see here). Areas that are traditionally implicated in emotion do also carry predictively-useful information. The machine-learning methods that are used to locate these information-bearing patterns do not allow for quantifiable comparisons as to which areas contain more information, relative to their size, but what is clear is that patterns of activation in non-emotional regions predict intrusive memory.
This is a precarious basis for an inference about the causes of the intrusive memories whose occurrence is predicted. In a couple of short pieces I have tried to draw out the reasons for this (see the readings below), but the basic point can be indicated with an analogy: When online advertisers are trying to predict whether you might soon be having a baby, they find useful information on which to base their predictions in the patterns of your internet searches. Those internet searches are not the cause of the baby’s occurrence. The lesson to draw from this is that the richest sources of non-intrusively gatherable information about a thing are not typically the causes of that thing. We should therefore be wary of supposing that we identify the causes of PTSD-like symptoms by identifying sources of information about the occurrence of those symptoms.
With that in mind, I think that questions about the role of emotion in the cause of PTSD should be regarded as open. This is not to deny the importance of the work that has brought these questions to light, and that has drawn our attention to the contributions made to PTSD at the time of the traumatizing experience.