Remembering the Past is Like Imagining the Future

By Sean Carroll | April 14, 2009 8:32 am

Because of the growth of entropy, we have a very different epistemic access to the past than to the future. In retrodicting the past, we have recourse to “memories” and “records,” which we can take as mostly-reliable indicators of events that actually happened. But when it comes to the future, the best we can do is extrapolate, without nearly the reliability that we have in reconstructing the past.

However — the human brain, as most readers of this blog probably know, was not intelligently designed. It’s doesn’t have the high-level structure of a computer program, where all the processes are carefully planned to achieve some goal. (The lower-level structures share the mechanical features of any other physical system, but that’s of little help here.) Evolution nudges the genome in useful directions, but it can only work with the raw materials it’s given; it doesn’t have the luxury of starting from scratch. So over and over in biological organisms, we find features that were originally developed for one purpose being re-engineered for something else.

As it turns out, the way that the human brain goes about the task of “remembering the past” is actually very similar to how it goes about “imagining the future.” Deep down, these are activities with very different functions and outcomes — predicting the future is a lot less reliable, for one thing. But in both cases, the brain goes through more or less the same routine.


That’s what Daniel Schacter at Harvard and his friends have discovered, by doing functional MRI studies of brains subjected to different kinds of cues. (Science News report, Nature review article, Charlie Rose interview.) Subjects are inserted gently into the giant magnetic field, then asked to either conjure up a memory or imagine a future scenario about some particular cue-word. What you see is that the same sites in the brain light up in both cases. The brain on the left in this image is remembering the past — on the right, it’s concocting an imaginary scenario about the future.


Further confirmation comes from studies of amnesiacs, who famously can’t remember the past. But if you ask the right questions, you find that they also have significant problems imagining their own future.

We tend to assume that the brain must be like a computer — when we want to access a memory, we simply pull up a “file” stored somewhere on the brain’s hard drive, and take a look at its contents. But that’s not it at all. Schacter believes that pieces of data relevant to any particular memory — times, images, sounds — are stored piecemeal in different parts of the brain. When we want to “remember” something, another part of the brain assembles these pieces into a (hopefully) coherent picture. It’s like running a new simulation every time you need a memory, and it’s the same thing we do when we try to imagine some event in the future.

Everyone has heard that memories can be unreliable, but many of us don’t appreciate the extent to which that is true. It’s not the case that “real” memories are stored once and for all deep in the darkest recesses of the brain, and it’s just a matter of digging them up. False memories — conjured from any number of sources, from gradual embellishment to direct suggestion by others — seem precisely as vivid and real to us as accurate memories do. For a good reason: the brain uses the same tools to construct the memory from the available raw materials. A novel and a history book look the same on the printed page.

  • Aaron Sheldon

    That is brilliant.

    I wonder if we will one day be able to test why we experience the present sequential, or rather image the part of the brain that is responsible for sequential experience.

    There are good odds that there is a structure in the brain responsible for this, the evidence coming from the anecdotal reports of the use of hallucinatory drugs.

    Good luck getting ethics approval for those experiments though!

  • greg

    In Hume’s his Treatise of Human Nature, he suggests that memory and imagination are the two methods in which our brain processes the same set of information which we’ve gathered from our experiences. I just started reading this, so I’m not sure if the rest of his discussion of memory and imagination lines up with this experimental data, but I was struck by the coincidence of my having just started reading this and your posting the article discussing scientific examination of the same thing Hume was writing about.

  • tyler
  • Peter Coles

    Both images show what the brain is doing when the subject’s head is stuck in an MRI machine….not surprisingly, it’s the same response.

  • Martin

    You’d think they would have noticed by now if *all* instructions to people with their heads stuck in MRI scanners gave the same result…

  • Brian

    “So over and over in biological organisms, we find features that were originally developed for one purpose being re-engineered for something else.”

    At first I accidentally read the above sentence with an extra colon inserted, like so: “So over and over in biological organisms, we find features that were originally developed for one purpose: being re-engineered for something else.” The original is probably more factually accurate, but this version is nonetheless oddly compelling.

  • Counterfly

    Sean: despite the size of your hammer, be careful not to view everything as a nail. I feel (as a physicist) that lately you have been pushing the arrow of time schema into realms of dubious applicability.

  • Peter Coles

    Martin, I hope you’re right but I’d like to know how they did the control. An activity with your brain that doesn’t involve either memory or prediction? Hmmm.

  • Yuri Danoyan

    If the Universe is cyclic, no difference between the past and the future.Remember Plato philosophy….

  • sh

    “So over and over in biological organisms, we find features that were originally developed for one purpose being re-engineered for something else.”
    Point taken, but don’t you mean something more like, “… we find features that originally survived and propagated because they were useful for one purpose being re-utilized for something else.”

  • Jasph

    Poets and philosophers really do get there before scientists do. I think it was Gaston Bachelard, or someone like him, who said, “Memory and imagination come from the same place.”

    I guess it figures. Memory, after all, tends to be made of images.

  • Ryan

    “However — the human brain, as most readers of this blog probably know, was not intelligently designed. It’s doesn’t have the high-level structure of a computer program, where all the processes are carefully planned to achieve some goal.”

    This seems a little dismissive of the capability of the human brain, and seems to assume that ‘carefully planning’ each function of a control system is the best solution. Rather than viewing these results as evidence that the brain is ignorantly using the same hardware to solve two fundamentally different problems, I see this as evidence that the algorithm of consciousness is not very well understood.

  • Gabriel

    Your last paragraph’s conclusions have some serious implications, Sean – and they also explain a lot.

  • Pingback: Science Tuesday and Tea Party Eve…. « blueollie()

  • Follower

    Counterfly said: “Sean: despite the size of your hammer, be careful not to view everything as a nail. I feel (as a physicist) that lately you have been pushing the arrow of time schema into realms of dubious applicability.”

    Maybe, but isn’t that a nice change from the way most physicists studiously ignore the AOT issues? For example, how people can imagine that the black hole information problem can be solved without understanding the AOT is inexplicable to me.

  • Jennifer West

    I guess the brain knows which input files it draws from to do the simulation. Even if the input files cannot be uncovered by those studying the brain, the brain likely knows whether it was sight/sound/smell/feel/taste, and thus can recognize the difference between true and false memories. Unless the false memory was planted using the same set of senses.

  • daisyrose

    I wonder if this brain” similarity” for the future and past could effect : say – buying Insurance or ones future ability to pay back a loan?

  • Andy Lawrence

    Fascinating. Possibly one of the first actually new things about the AOT that I have seen for a long time. This may be asking a lot, but what we would really want is a time series. Maybe as you go through a series of mental states, the track of locations, as opposed to the mean location, will be different between remembering and imagining ?

  • Kaleberg

    If you ever do any machine planning or machine learning work you find yourself developing some kind of canonical data structure for expressing your plans that allows your system to anticipate sensor input and results. You’ll find the same data structures and analysis useful for assessing how a plan executed in the past. You’ll also find that it helps to use the same structures when in the middle of executing a plan, modulo the need for rapid responses and offloaded execution.

    In fact, it turns out that having the same kind of representations for past, present and future makes it easier to plan for the future, remember the past and figure out what you are doing now and how well it is working.

    I’m not at all surprised that the human brain seems to use the same brain regions for rehearsing and reasoning about future actions, remembering and analyzing past actions and, most likely, keeping track of what is going on. (You might experiment with the last by having the subject follow a story or physical walkthrough in an MRI and prompting them with questions about the story as it unfolds e.g. what happened first?, where?, what is likely to happen next?) The existence of mirror neurons that track both the performance, or anticipated performance of some action, and the observation of the same actions suggests that this past-present-future symmetry is also symmetric with regards to the actor, oneself, one imagined or one observed.

  • Jean-Paul Billon

    I agree with kaleberg. AI machine Planning future actions is quite similar to extracting a coherent view of past chunks of memory. In fact, remembering is just like finding an optimal coherent trajectory in a phase space with few fixed and very partial records of stimulus. Imagining the future is the same optimal trajectory computation with a few variable “what if ” event records extrapolated form our past experience. The only thing that makes the difference to our consciousness is a kind of tag “has truly been recorded” versus “might happen”. If this tag is not present, past and future are confused and we have delusional personalities. In brief, remembering and planning are structurally rather similar, and I guess that remembering is just a feature that came to species who are able to planning.

  • Pingback: Links With Your Tea Party at Tête-à-Tête-Tête()

  • John P

    I am skeptical about these functional MRI studies. It seems like quite a leap from different pretty pictures of brain activity to the conclusion that we have learned something useful about this enormously complex piece of matter. There is activity in the brain that does not show up on the scans. Analogies using “files” and other computer science concepts are laughable. The human brain has arisen via evolution through eons of time. Its structure is completely independent of our projection of design concepts of modern silicon-based computer hardware onto it. A bit more humility and awareness of our ignorance is in order here.

    “This is a very gross technique,” says critic Steven Faux , who heads the psychology department at Drake University. “It’s like a blurry photo–better than no photo but still blurry, with real limitations that are too often overlooked. It’s very easy to overextend [the value of] this technology.”

  • Fred

    Explains a lot about Joe Biden

  • Pingback: My Brain Has a Mind of Its Own « Either Here or There()

  • Ankur Dnyanmote

    Its appears that the common center-piece involved in both activities (remembering past as well as imagining future) operates on information stored in different parts of the rest of the brain. This information HAS to be a memory if it is part of the brain. Whether it is a real memory or imagined is not really a concern. What is interesting is that the common center-piece that is orange in the above pictures conjures up “information” in response to the external cues. To imagine a future this region of the brain must require inputs of information. (And this is true also if it has to remember the past). The difference is that in the former case the brain CREATES experiences that have not happened (and may not happen) while in the latter case the brain CREATES experiences that HAVE happened (or may not have happened). The paradoxical aspect of this is our subjective treatment of a PAST and a FUTURE. In reality both these are epiphenomena resulting from our perception of time as a sequential experience.

  • Sili
  • Pingback: Things I kinda meant to blog about « A Fistful of Science()

  • Pingback: AAAS 2010 | Cosmic Variance | Discover Magazine()

  • Pingback: Remembering the Past is Like Imagining the Future « Fire EXIT()

  • Pingback: Neurologie du Temps « Dr. Goulu()

  • Pingback: K and J Investigations and Case Management » Remembering the Past While Thinking About the Future()


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

About Sean Carroll

Sean Carroll is a Senior Research Associate in the Department of Physics at the California Institute of Technology. His research interests include theoretical aspects of cosmology, field theory, and gravitation. His most recent book is The Particle at the End of the Universe, about the Large Hadron Collider and the search for the Higgs boson. Here are some of his favorite blog posts, home page, and email: carroll [at] .


See More

Collapse bottom bar