Lane et al. note that there is no universally accepted account of how psychotherapy works. They draw on neurological and psychological data to develop a theory about the relationships between memory, emotion, and semantic structures; they then put this theory to work in explaining how psychotherapy works and how it can be made more effective. Because it draws on biological, psychological, and social elements, Lane et al.'s account of psychotherapeutic effectiveness can be deemed a “multilevel” explanation.
Understanding and treating mental illness is recognised by many psychiatrists to require consideration of biological, psychological, and social perspectives; yet organizing these perspectives into useful, coherent explanations remains fraught with difficulty. Some, such as Christopher Frith, hold that multilevel explanations are, in an important sense, unintelligible and uninformative (Frith Reference Frith1992, p. 26). Others, such as Marmot (Reference Marmot and White2005, p. 53) and Ghaemi (Reference Ghaemi2010, p. 58ff), view multilevel explanations as impracticably complex, believing that for practical purposes it is necessary to constrain one's thinking to a single level when trying to understand a medical phenomenon.
Despite these concerns about multilevel explanations, Lane et al.'s account of psychotherapeutic effectiveness provides a case study of the usefulness of multilevel explanation. I identify three specific ways in Lane et al. advance understanding of psychotherapeutic effectiveness by drawing on multilevel insights. I then argue that whether a given explanation is useful depends largely on what purpose one intends the explanation to serve, and that this point has not been adequately recognised in psychiatry.
First, Lane et al. note that attention to one level can lead us to revise beliefs about the composition of another level. For example, they note that memory reconsolidation – which involves changes in recalled memories – is behaviourally similar to so-called extinction, in which a new memory overrides an old one. Extinct memories, unlike earlier versions of reconsolidated memories, can reappear over time. Lane et al.'s account of psychotherapy depends upon reconsolidation and extinction being distinct psychological processes. The claim that they are psychologically distinct is supported by their biological differences: Lane et al. note that the two processes differ at the cellular/molecular level. As such, a multilevel explanation of these processes advances understanding of them.
Second, sometimes there is no single-level explanation for why a given phenomenon exists. Lane et al. answer the question of why our memories admit of revision through reconsolidation by appealing to Klein et al.'s (Reference Klein, Cosmides, Tooby and Chance2002) argument that this feature is adaptive because it enables us to update existing knowledge in light of new information. Appealing to adaptiveness in this way explains a psychological phenomenon (the mutability of memory through reconsolidation) in biological (evolutionary) terms. Confining ourselves to the psychological level makes it hard to see how this feature of memory could be advantageous; indeed, the flashbulb memory literature in psychology conceives it chiefly in terms of a vulnerability to error, as Lane et al. note. Multilevel explanation of memory reconsolidation, then, helps us understand it better.
Third, psychiatry is a goal-directed enterprise: Its goal is to prevent, cure, and/or manage mental illness. A multilevel understanding of how a desired effect occurs can reveal new ways of achieving that effect, thereby opening up new possible treatment routes. Lane et al. draw on their biologically informed understanding of the psychological processes underlying psychotherapeutic success to suggest ways of pharmacologically bringing about the results of successful psychotherapy. Whilst psychotherapy is one way of effecting desirable memory modification, pharmacologically manipulating patients' emotional responses during recollection could be another. Lane et al. note that efforts to develop such treatments are already under way: Propranolol, a beta-adrenergic antagonist, has been used to block the formation (or strengthening) of traumatic memories in patients with (or at risk of) post-traumatic stress disorder. Further multilevel insights into the biology of psychotherapy could reveal new opportunities for pharmacological intervention.
Lane et al. make use of these and other multilevel explanations in understanding psychotherapeutic effectiveness and considering how its effectiveness can be improved. Reflecting on their article reveals not only that multilevel explanations can be useful, but also that the usefulness of a multilevel explanation – or, indeed, any explanation – depends largely on what one wants from it. Frith's complaint that the multilevel explanation “alien thoughts are caused by inappropriate firing of dopamine neurones … is clearly inadequate” might be reasonable given his wish to learn about “the nature of hallucinations” and “the role of dopamine neurons within the physiological domain” (Frith Reference Frith1992, p. 26). Yet the explanation that Frith finds inadequate is useful if our aim is, instead, discovering whether pharmacologically manipulating dopamine neurons is likely to be an effective way to control alien thoughts. Similarly, Marmot's and Ghaemi's complaints that multilevel explanations are best avoided if we are to avoid becoming “paralysed by complexity” (Marmot Reference Marmot and White2005, p. 53) is reasonable in cases where one's aims are most effectively realised by considering only single-level explanations, but not in cases – like that of Lane et al.'s attempt to explain how psychotherapy works – where achieving one's aims requires consideration of multilevel factors.
I remarked above that combining the perspectives of different levels into coherent explanations of mental illness is a difficult task. Although impressive work has been done – particularly by Kendler (e.g., Reference Kendler2008; Reference Kendler2012) and by Kendler and Campbell (e.g., Reference Kendler and Campbell2009; Reference Kendler and Campbell2014) – to demonstrate the need for multilevel explanations in psychiatry and to consider how they are best formulated, the question of what makes a good explanation for a given purpose has been ignored. We know, for example, that some cases of depression are best explained primarily in terms of psychosocial factors such as bereavement, and that others are best explained primarily in terms of biological factors such as abnormal brain activity – but what factors in general determine whether and when attention to one or another level, or to multiple levels, is explanatorily more appropriate and useful, is an issue that requires further investigation.
You have
Access
Lane et al. note that there is no universally accepted account of how psychotherapy works. They draw on neurological and psychological data to develop a theory about the relationships between memory, emotion, and semantic structures; they then put this theory to work in explaining how psychotherapy works and how it can be made more effective. Because it draws on biological, psychological, and social elements, Lane et al.'s account of psychotherapeutic effectiveness can be deemed a “multilevel” explanation.
Understanding and treating mental illness is recognised by many psychiatrists to require consideration of biological, psychological, and social perspectives; yet organizing these perspectives into useful, coherent explanations remains fraught with difficulty. Some, such as Christopher Frith, hold that multilevel explanations are, in an important sense, unintelligible and uninformative (Frith Reference Frith1992, p. 26). Others, such as Marmot (Reference Marmot and White2005, p. 53) and Ghaemi (Reference Ghaemi2010, p. 58ff), view multilevel explanations as impracticably complex, believing that for practical purposes it is necessary to constrain one's thinking to a single level when trying to understand a medical phenomenon.
Despite these concerns about multilevel explanations, Lane et al.'s account of psychotherapeutic effectiveness provides a case study of the usefulness of multilevel explanation. I identify three specific ways in Lane et al. advance understanding of psychotherapeutic effectiveness by drawing on multilevel insights. I then argue that whether a given explanation is useful depends largely on what purpose one intends the explanation to serve, and that this point has not been adequately recognised in psychiatry.
First, Lane et al. note that attention to one level can lead us to revise beliefs about the composition of another level. For example, they note that memory reconsolidation – which involves changes in recalled memories – is behaviourally similar to so-called extinction, in which a new memory overrides an old one. Extinct memories, unlike earlier versions of reconsolidated memories, can reappear over time. Lane et al.'s account of psychotherapy depends upon reconsolidation and extinction being distinct psychological processes. The claim that they are psychologically distinct is supported by their biological differences: Lane et al. note that the two processes differ at the cellular/molecular level. As such, a multilevel explanation of these processes advances understanding of them.
Second, sometimes there is no single-level explanation for why a given phenomenon exists. Lane et al. answer the question of why our memories admit of revision through reconsolidation by appealing to Klein et al.'s (Reference Klein, Cosmides, Tooby and Chance2002) argument that this feature is adaptive because it enables us to update existing knowledge in light of new information. Appealing to adaptiveness in this way explains a psychological phenomenon (the mutability of memory through reconsolidation) in biological (evolutionary) terms. Confining ourselves to the psychological level makes it hard to see how this feature of memory could be advantageous; indeed, the flashbulb memory literature in psychology conceives it chiefly in terms of a vulnerability to error, as Lane et al. note. Multilevel explanation of memory reconsolidation, then, helps us understand it better.
Third, psychiatry is a goal-directed enterprise: Its goal is to prevent, cure, and/or manage mental illness. A multilevel understanding of how a desired effect occurs can reveal new ways of achieving that effect, thereby opening up new possible treatment routes. Lane et al. draw on their biologically informed understanding of the psychological processes underlying psychotherapeutic success to suggest ways of pharmacologically bringing about the results of successful psychotherapy. Whilst psychotherapy is one way of effecting desirable memory modification, pharmacologically manipulating patients' emotional responses during recollection could be another. Lane et al. note that efforts to develop such treatments are already under way: Propranolol, a beta-adrenergic antagonist, has been used to block the formation (or strengthening) of traumatic memories in patients with (or at risk of) post-traumatic stress disorder. Further multilevel insights into the biology of psychotherapy could reveal new opportunities for pharmacological intervention.
Lane et al. make use of these and other multilevel explanations in understanding psychotherapeutic effectiveness and considering how its effectiveness can be improved. Reflecting on their article reveals not only that multilevel explanations can be useful, but also that the usefulness of a multilevel explanation – or, indeed, any explanation – depends largely on what one wants from it. Frith's complaint that the multilevel explanation “alien thoughts are caused by inappropriate firing of dopamine neurones … is clearly inadequate” might be reasonable given his wish to learn about “the nature of hallucinations” and “the role of dopamine neurons within the physiological domain” (Frith Reference Frith1992, p. 26). Yet the explanation that Frith finds inadequate is useful if our aim is, instead, discovering whether pharmacologically manipulating dopamine neurons is likely to be an effective way to control alien thoughts. Similarly, Marmot's and Ghaemi's complaints that multilevel explanations are best avoided if we are to avoid becoming “paralysed by complexity” (Marmot Reference Marmot and White2005, p. 53) is reasonable in cases where one's aims are most effectively realised by considering only single-level explanations, but not in cases – like that of Lane et al.'s attempt to explain how psychotherapy works – where achieving one's aims requires consideration of multilevel factors.
I remarked above that combining the perspectives of different levels into coherent explanations of mental illness is a difficult task. Although impressive work has been done – particularly by Kendler (e.g., Reference Kendler2008; Reference Kendler2012) and by Kendler and Campbell (e.g., Reference Kendler and Campbell2009; Reference Kendler and Campbell2014) – to demonstrate the need for multilevel explanations in psychiatry and to consider how they are best formulated, the question of what makes a good explanation for a given purpose has been ignored. We know, for example, that some cases of depression are best explained primarily in terms of psychosocial factors such as bereavement, and that others are best explained primarily in terms of biological factors such as abnormal brain activity – but what factors in general determine whether and when attention to one or another level, or to multiple levels, is explanatorily more appropriate and useful, is an issue that requires further investigation.
ACKNOWLEDGMENT
This work was produced with the support of the Oxford Loebel Lectures and Research Programme.