Implanted memories... or are they?

(An improved and updated version of this essay has appeared in my book Brief Peeks Beyond. The version below is kept for legacy purposes.)

Are memories stored in the brain as physical traces?

This week several people sent me a link to an article just published on Scientific American. The title promised something extraordinary: "The Era of Memory Engineering Has Arrived: How neuroscientists can call up and change a memory." That certainly sparked my curiosity, so I decided to read it tonight, while sipping some cold Riesling in my garden, trying to cool off from the heat of the day. In the end, I was indeed amused, but not for the reasons I thought I would be... allow me to elaborate.

The article starts with references to science fiction films in which the hero, at some point, realises that his memories were all implanted by 'evil' scientists. None of the past he remembers actually happened, but was artificially synthesised and inserted into his head. I immediately thought of Total Recall (the original), where people could go to a shop called 'Recall' and order custom-made memories of holidays, adventures, heated romances, or what not, without actually having to live through any of that. The author goes on to suggest that cutting edge work done at MIT is comparable to these amazing sci-fi scenarios. He writes: "...these scientists have captured specific memories in mice, altered them, and shown that the mice behave in accord with these new, false, implanted memories. The era of memory engineering is upon us, and naturally, there are big implications for basic science and, perhaps someday, human health and society." Wow, really? Have we been able to synthesise and implant memories like in Total Recall? I mean, no need for entire narratives... if even a simple memory (say, of switching on the lights) could have been synthesised and implanted, it would be very significant not only for science, but for philosophy as well.

You see, if it were possible to synthesise and implant memories that way, it would imply that we knew exactly what memories were, as well as where and how they were encoded in the brain. So the author's opening lines in the Scientific American article caught my attention. However, things weren't as they seemed...

When you read the rest of the article critically, and with attention, here is what you discover:

1. No memories were synthesised at all.
2. What was actually done was this: they found a way to measure and record the pattern of brain activity in mice when they were placed in a certain environment, and then they managed to 'reactivate' that same pattern of brain activity later on, in another environment.
3. When they re-activated the pattern in another environment, they gave the mice electric shocks, scaring the heck out of them.
4. When they put the mice back in the original environment, without the shocks, the mice were paralyzed with fear.

That's it. Now, let's look at what this actually means.

Both the experiences of the environment and of the electric shocks weren't 'implanted' memories. They actually happened. They actually shocked the mice. They actually placed the mice in that environment. All the experiment accomplished was to create an association between the original environment and the electric shock, without needing to actually make the two happen together, as in classical conditioning. So there is a sense in which one could perhaps say that the 'memory of the association' was 'implanted,' but that's totally different from what the article suggests in the beginning. It has very little, if anything, to do with Total Recall-type memory implants. No memory was synthesised at all, not even a very tiny simple memory. All experiences involved were actual experiences of the mice. They just tricked the mice into linking one real experience to another real experience. This is rather a cognitive link than a memory. They 'implanted' association, conditioning, not phenomenological or experiential memories.

You might say that, by reactivating a certain pattern of brain activity, the scientists artificially created recall. This is true, but it doesn't address the important question of what memory is or where it's stored. You see, experiences correlate with brain activation patterns; we know that. So if you induce a certain brain activation pattern in mice and associate that with a shock, it's no surprise that the shock will be cognitively linked to any future experience that triggers the same brain activation pattern. But that's not the question. The question is, when I close my eyes and remember my dead father, how the heck do I know what exact pattern of brain activity to bring back to my brain? Where is the information stored that allows me to reconstruct that pattern? In the MIT experiment, the scientists created their own storage mechanism by genetically modifying the mice to grow molecular switches in each neuron activated when the mice were placed in their original environment. Only the activated neurons grew the switches, so the distribution of the switches recorded the neural correlates of the original experience. The scientists could then turn these neurons back on later, using light. Of course, this doesn't explain how mice remember things when they haven't been engineered to grow these switches! The experiment explains exactly nothing about the mechanism of memory storage simply because it by-passes it altogether! It was the scientists who recorded and stored the information, and then used this information to create a pattern of brain activity, not the mice. How do the mice do it when there are no scientists to record, store, and re-launch the information in their brains?

The experiment also says nothing new about the nature of conscious experience. That experience is correlated with certain patterns of brain activity is very old news. That they could create an association between two events by activating their respective patterns together in the brain is also no news, since this has been shown by classical conditioning since the time of good-old Pavlov. The only novelty (and, make no mistake, it's amazing and important, just not in the way the article portrays it) is the scientists' amazing ability to record and then re-activate a particular pattern of brain activity. Kudos to MIT! This might have important far-future applications in e.g. new treatments of brain diseases and perhaps even in training & education. But don't expect a 'Recall' shop near you any time soon. Don't expect a solution to the 'hard problem of consciousness' any time soon. And don't expect an answer to the nature and location of memory any time soon. None of this was addressed, not even in principle, by this work.

To me, the key value of this Scientific American article was the fact that it powerfully illustrated how, through hysteria, lack of critical judgment, and naive enthusiasm, a false idea can be hyped by the mainstream media to the point of not only looking entirely plausible, but even certain. How many well-meaning people out there, who briefly read this article, won't be thinking now: 'Wow, it's a done deal... memory and consciousness really are all in the brain'? Can we even blame them? When one is already predisposed to finding confirmation of a certain idea, it's very easy to find it if one isn't critical of what one is looking at. We tend to find what we are looking for, even if it's not there. This article powerfully illustrates it. Here is an experiment that says exactly nothing about the nature of memory or our ability to synthesise memories, yet it's hyped precisely as such.

Since you're reading this post in my blog, this time you are at least thinking a little more critically about this specific article. But how many other similar articles have you read before? How many of your implicit beliefs today have been subtly created this way? It's scary, isn't it?