Forget the hype: how close are we to a 'forgetting pill'?

The neuralyzer from Men in Black

I've been a little disconcerted by the recent appearance in the popular science press of a number of articles seeming to claim that we're just around the corner from being able to erase painful or traumatic memories.  For example:

Wired: The Forgetting Pill Erases Painful Memories Forever by Jonah Lehrer

Scientific American: Erasing Painful Memories: Drug and Behavioral Therapies Will Help Us Forget Toxic Thoughts by Jerry Adler

The articles are beautifully written, full of interesting and thought-provoking questions, and obviously the product of a great deal of work.  I think good science writing is really important and greatly value the work that writers like Jonah Lehrer and Jerry Adler do. However, I can't understand how these very clever, usually marvellous writers make the huge leap in this instance from the (albeit in themselves fascinating) findings in animal models to the putative selective erasure of the complex, multidimensional, highly interconnected ensemble of neural representations that constitutes a single human autobiographical memory.

This matters because many thousands of people suffer enormous anguish every day with the dreadful effects of post-traumatic stress or related conditions, and may have their hopes raised that a "forgetting pill" is just around the corner.  It seems to me that this hype isn't justified based on current knowledge, although as this isn’t my area of specialist expertise, maybe I’m missing something.  I had an interesting email conversation with Jonah Lehrer in which he was characteristically open to a number of my (hopefully constructive) criticisms.  However, to find out whether I might have misunderstood the science, I asked someone who is an expert in this area, Dr Amy Milton from the University of Cambridge, to set things straight.  Here’s her view:

Dr Amy Milton

When I describe my research to anyone, the conversation invariably ends up leading to the film, “Eternal Sunshine of the Spotless Mind”. My research – focused on the neurochemical mechanisms by which memories persist in the brain and can be modified after their initial storage – has been developed not only to understand fundamentally how memory works, but also with the intention to translate this understanding into new forms of treatment for psychiatric disorders like post-traumatic stress disorder (PTSD) and drug addiction. The hypothesis itself is simple: we know that maladaptive memories (cue-fear memories in PTSD, cue-drug memories in addiction) contribute to the persistence of these disorders, and enhance the risk of relapse following a symptom-free period. So, if we can weaken or erase these maladaptive memories, can we reduce the risk of relapse? This type of pharmaceutically-enhanced behavioural therapy has the potential to revolutionise treatment of psychiatric disorders, and it is often reported excitedly by the media in terms of ‘forgetting pills’. But to what extent is the hype supported by the research?

The first point to consider is that memory is not a single entity; even putting aside the differences between short- and long-term memory, long-term memory can itself be subdivided. ‘Explicit’ or ‘declarative’ memories, such as for events (episodic) and facts (semantic) are consciously remembered and their content can be articulated, unlike ‘implicit’ or ‘non-declarative’ memories, such as emotional or procedural memories, which cannot. Of course, for any single event, an individual might form multiple memories; for example, an episodic memory of having a car accident, and an implicit emotional memory that leads to fear of getting into a car. The fact that memories form independently, depending upon different brain structures, may also mean that they can be modified independently. Thus, a ‘forgetting pill’ would not necessarily cause amnesia for all memories of an experience.

Secondly, it is important to consider whether our putative ‘forgetting pill’ would be designed to prevent the initial storage of the memory in the brain (the ‘consolidation’ process) or its persistence: this is likely to be achieved through targeting the process of ‘reconsolidation’. Briefly, when a memory is recalled under certain behavioural conditions (‘reactivated’), it returns to a state where it can be updated before restabilising in its new form, overwriting the original memory. Thus, by giving amnestic agents at reactivation, it may be possible to weaken or erase old, well-established memories. Though there has been work investigating whether disrupting memory consolidation might be used to treat PTSD, these types of treatment are faced with the difficult task of racing the closing ‘consolidation window’. Memories take only a few hours to form within the brain, and are fully consolidated within 24 hours; thus, a ‘forgetting pill’ that aimed to block the consolidation of cue-fear memories in PTSD would have to be taken shortly after a traumatic event. This raises both practical and ethical issues, particularly as the individual may be undergoing other medical treatment, and would have to give informed consent in an acutely traumatised state. Targeting reconsolidation may avoid these issues; as even old, well-established memories can have their reconsolidation disrupted, this means that the patient can be referred for treatment only if they go on to develop the disorder (and not all individuals exposed to a traumatic event do develop PTSD). Therapies based on the disruption of reconsolidation would also potentially allow treatment of disorders like drug addiction, which have a less clearly defined onset than many PTSD cases.

Many studies using animal models of both PTSD and drug addiction have shown that it is possible to disrupt the reconsolidation of both cue-fear and cue-drug memories, and so to reduce anxiety and relapse risk respectively.  However, while emotional memories can be readily assessed in animals, it is more difficult to assess episodic-like memories. Declarative memory can be more easily assessed in human studies, such as those conducted with PTSD patients. However, although these studies appear to have disrupted the reconsolidation of the emotional cue-fear memory, the effects on the declarative memory of the trauma were not reported. A study conducted with healthy individuals who had been trained to fear spider pictures may be informative, however. When the spider-fear memory was retrieved in conjunction with the amnestic agent (the non-selective beta-blocker, propranolol), the subjects showed a reduction in physiological anxiety measures, but still maintained the declarative knowledge that the spider picture had previously been paired with an aversive shock. Thus, in these studies, it seems that only the reconsolidation of the emotional memory, not the declarative memory, was disrupted. This is not to say that the episodic memory does not reconsolidate, only that its reconsolidation was not affected using this procedure. Of course, this may not matter; from a therapeutic perspective, it may be that weakening the maladaptive emotional memories that contribute to PTSD and addiction is sufficient to allow the patient to live a normal life and, indeed, may leave the patient population in a state similar to those who do not go on to develop the disorders despite similar experiences.

In my opinion, memory-disrupting treatments have real potential to enhance clinical outcomes when administered alongside carefully tailored behavioural therapy. There is strong evidence that pharmaceutical treatment, given with memory reactivation, can disrupt the maladaptive emotional memory that contributes to persistent anxiety in PTSD patient populations. There is encouraging evidence from animal models of drug addiction that similar sorts of combined behavioural and pharmaceutical therapy could be used to disrupt cue-drug memories, reducing the risk of relapse in the long term. It is also safe to say that episodic memories can be interfered with at memory retrieval, suggesting that they do reconsolidate, but to date there is little evidence that these event memories can be ‘erased’ in a human population. So, the answer to how close we are to a ‘forgetting pill’: ultimately, it depends upon what you are trying to forget.

References
Brunet A, Orr SP, Tremblay J, Robertson K, Nader K, & Pitman RK (2008). Effect of post-retrieval propranolol on psychophysiologic responding during subsequent script-driven traumatic imagery in post-traumatic stress disorder. Journal of psychiatric research, 42 (6), 503-6 PMID: 17588604
Kindt M, Soeter M, & Vervliet B (2009). Beyond extinction: erasing human fear responses and preventing the return of fear. Nature Neuroscience, 12 (3), 256-8 PMID: 19219038
Milton AL, & Everitt BJ (2012). The persistence of maladaptive memory: addiction, drug memories and anti-relapse treatments. Neuroscience and biobehavioral reviews, 36 (4), 1119-39 PMID: 22285426
Steckler T, & Risbrough V (2012). Pharmacological treatment of PTSD - established and new approaches. Neuropharmacology, 62 (2), 617-27 PMID: 21736888