This is third in a series.  See parts one and two.

Not all memories are good, obviously.  I’ve done plenty of stupid things, blurted out plenty of awkward remarks in conversations, that I’d prefer to forget.  And those are harmless.  They might make me flush and feel retroactively embarrassed if I think of them at night, but, big deal.

Other people have seen far worse things than me.  Their memories, instead of minor self-consciousness, lead to post-traumatic stress disorder (PTSD): sleeplessness, headaches, stress biomarkers suggestive of shortened lifespan, proclivity toward substance abuse, etc.

PTSD is bad news.  Very bad news.  There’s no data yet on what causes it — why does one person subject to a horrible event pull through fine, but another contracts the nightmares?  There’s simply too much we don’t understand about the brain.

Mostly this essay is going to be about memory erasure — if we could consistently & specifically delete memories, that’d go a long way toward curing PTSD.  And I have a minor ulterior motive for slapping up a few scientific references for memory erasure: because misplaced memories are essential for the plot of The Ramayana, I tucked the concept into my book.  When I first wrote those passages, they were moderately speculative, but in the intervening years our scientific understanding has actually caught up appreciably.

But first I wanted to give a small shout-out for the film Renegade.  Despite being a terrible cowboy film, Renegade is also the best depiction I’ve seen of using psychedelics to treat PTSD.  Which maybe sounds a little odd, but there are reasons to expect why it might work (as opposed to, say, attempting to use psychedelics to cure autism, which is every bit as ill-reasoned and abusive as it sounds).

Cognitive behavioral therapy (CBT) seems to be the leading non-pharmaceutical strategy to treat depression, sleep disorders, and other crummy brain states including PTSD.  The basic idea is that thoughts are reinforcing.  Thoughts you’ve had once are easier to have again, which is why studying works, why practice makes perfect.  But people can get stuck in ruts, looping through the same negative thoughts over and over.  CBT aims to replace those ruts with new virtuous cycles of helpful thoughts.

And that’s why psychedelics might make CBT more effective: the therapy will reinforce itself once it gets going, but actually jolting someone out of their initial rut seems difficult.  Mind-altering substances might provide a window of time during which a new cycle of thoughts can be more readily inoculated into someone’s brain.

This is speculative, of course.  FDA restrictions have made it difficult to conduct research using psychedelics.  Not many studies have been done so far, and a lot of our nation’s more illustrious research groups haven’t been involved; if your research is already going well, investigating psychedelics probably seems like a lot of hassle for little payoff.

But, let’s say you were a film director who’d been given a reasonably high budget to create an adaptation of a French comic book but instead used much of that money to take ayuhuasca hundreds of times in Peru … would you care that the ideas are speculative?  I think not.  Jan Kounen, who directed Renegade, certainly didn’t.  The scaffolding film is a cheesy western, but the central premise deals with overcoming the curse of a traumatic memory.  And, sure, the trippy CGI fractal swirls are often vaguely reminiscent of screensavers, but I appreciate the director’s ambitious attempt to depict visually what was happening in his protagonist’s mind as he reconciled himself with his past.  I think Renegade does a better job of depicting these unarticulated inner states than, say, Eternal Sunshine of the Spotless Mind, Inception, or Dark City.

Even if psychedelics could help people with PTSD, though, it seems inconceivable that the drugs alone would be effective — the only reasonable mechanism I can come up with is that the drugs might make someone more receptive to CBT.  And the therapy takes time, takes a lot of concentrated effort from both patient and therapist, and probably seems like it’s not accomplishing anything for most of the initial sessions.  About a third of people with PTSD who have participated in CBT research studies drop out of treatment, and in general practice this number might be even higher.

It’d be nice if we could address the underlying traumatic memories directly.  Delete them specifically from someone’s brain.

And we can’t, obviously.  As per my previous post in this series, we can’t identify memories based on brain structure alone.  That rules out opening up someone’s head and attempting to physically ablate the offending thoughts… although it’s quite clear that physical disruption could remove memories.  Here’s the first paragraph of Henry Marsh’s Do No Harm:

I often have to cut into the brain and it is something that I hate doing.  With a pair of diathermy forceps I coagulate the beautiful and intricate red blood vessels that lie on the brain’s shining surface.  I cut into it with a small scalpel and make a hole through which I push with a fine sucker — as the brain has the consistency of jelly a sucker is the brain surgeon’s principal tool.  I look down my operating microscope, feeling my way downwards through the soft white substance of the brain, searching for the tumour.  The idea that my sucker is moving through thought itself, through emotion and reason, that memories, dreams and reflections should consist of jelly, is simply too strange to understand.  All I can see in front of me is matter.  Yet I know that if I stray into the wrong area, into what neurosurgeons call eloquent brain, I will be faced by a damaged and disabled patient when I go round to the Recovery Ward after the operation to see what I have achieved.

We could delete memories, but given our current understanding of the brain we would be wrecking them at random.

What else do we know about erasing memories?

My favorite model is that memories, when used, seem to be replaced — details that weren’t included in an initial act of remembrance are often lost forever, which could indicate that a new memory of that remembrance does something like overwriting it.  And this would explain why our most vivid recollections are of things we seldom think about — as in Proust, a rare smell or sound or physical sensation can spur our sharpest memories.

Conversely, repeated remembrance seems able to dim a memory.  But there’s a trick — that whole problem with mental ruts.  Typically remembrance dims the initial memory, but encodes new information.  Whatever we think about while remembering becomes the memory (in part this seems to be why eyewitness testimony is so often wrong — poor police practice can easily allow a new, wrong memory to be encoded and reinforced before a trial).

To erase a memory, it would have to be recalled but not replaced.

One method for this is termed “thought substitution” — actively trying to make your mnemonic records incorrect.  If you try your darnedest to think about something else while remembering a traumatic event, you might be able to replace the traumatic memory with an innocuous thought.

…although (and this isn’t addressed in the paper I cited above), it seems possible that you would instead link horrible emotions to the previously innocuous replacement thought you tried to overwrite your trauma with.  I’m not sure whether there’s any reason to expect the transfer to be unidirectional.

And, right, I first began researching memory because it is integral to the Ramayana.  There’s an incident of brutal sexual violence (perpetrated by the heroes!!), which, right, often results in PTSD.  And there are less horrific instances of memory loss throughout the myth — the central protagonist has to forget his divine origin in order to be sufficiently human (i.e., weak) to slay the erudite vegetarian ruler of Lanka, and the heroic monkey has to forget his powers so that he doesn’t destroy sacred temples in his exuberant rampages.

So I had to put memory erasure into my book, despite this being beyond the ken of contemporary science.  My assumption was, if memories are accessed at a time when there is pharmaceutical destabilization of firing synapses or even just inhibition of synaptic connection reinforcement, those memories might well be diminished.  The biggest problem, from a pharmaceutical perspective, is one of targeting: epigenetic phosphorylation seems to be essential to maintain long-term memories, and inhibition of the phosphorylase that maintains them (PKMζ) seems to erase memories, but we can’t deliver an inhibitor specifically to potentiated synapses.  Memory erasure is still science fiction — but doesn’t seem to be unreachably far away.