I can't remember which question it was, but in the past here, it was explained to me that memory is actively stored in the brain, rather than "physically". Meaning, memories are made up of the brain's activity, not its physical cell structure.

I'm not sure I have that right though. I mean, that, for one thing seem inefficient and overly complex. It seems like a vast amount of energy would be required to store so much information in the form of activity patterns. Maybe I misunderstood. So I'm asking. If my brain lost all of its energy (activity) for a moment (for example, one second), would my memory be gone too?


Memories are dynamic; that is not equivalent to "memories are evanescent". Cellular structure and composition controls memory, not just a continuous passage of signals. This point is already discussed in the comments.

Loss of energy (in the form of ATP) would cause memory deterioration and in some cases neuronal cell death too— strokes, which happen due to lack of blood supply to brain (and hence unavailability of nutrients and oxygen) can cause amnesia.

But one second will practically not affect anything.


Summary Working memory on a current task could be distorted, but I cannot find a reason to suggest sudden loss of electrical information would permanently alter short or long term memory.

Hopefully a neuroscientist can answer with a more electro-chemical oriented answer, but fundamentally I think the experts aren't sure what a memory actually is (there are only a handful of theories each that are understudied or flawed). This is also a fairly hypothetically question. Baring these points in mind, treat not only this, but all answers skeptically and take with a pinch of salt!

LTP A recent review by Baudry et al covers the modern biochemical view on long-term potentiation (LTP) of the hippocampus that is probably responsible for things like learning and memory. It is a complex protein formation involving BDNF, calpain, and synaptic modifications. If hypothetically you became 'brain dead' for a moment, this way of looking at memory probably wouldn't take a hit.

Engrams Something I have less of a grasp on is Richard Semon's Engram, but was more heavily studied by Richard Thompson. Even the most modern studies describe this kind of brain mapping as chemical and cellular rather than electrical. So again perhaps nothing other than the unstored would be lost.

Working Memory loss This rather technical article discusses primate prefrontal cortex neurone firing as a persistent stream needed for a "working memory". I am out of my depth here, but it seems like some "forming" memories may be lost if there was momentary activity loss, but I don't see why the underlying biochemical state that cause this effect would be lost, so very little information would be directly lost.

Memory loss. Diseases that cause memory loss involve tissue and protein changes along with other physiological changes. Memory loss induced by trauma is likely caused by swelling where the neurones are damaged by swelling or the trauma itself on a cellular level. It seems to be the case that the electrical information isn't the important factor to memory loss.

Note: I think memory would be lost indirectly, perhaps removed in a similar way as when you wake up from a dream or in cases of trauma. I am also ignoring that there would be a reasonable probability that this would induce seizure, stroke, or other bad health effects!


I found the answer to my question here, however not linked to a duplicate question, so I'll quote:

Memories are represented in the brain as patterns of firing neurons. Let's say, for simplicity, that each neuron can either be on or off. Since there are 86 billion neurons in the brain, we can experience 2^86bn possible brain states. That's a lot.

Of course there are a lot of simplifications here--most notably, there is a lot of structure to the brain, and so i'm not suggesting you can store 2^86bn memories--but the logic is the same. Even with only 1,000 neurons, we can store a ridiculous amount of information (2^1000). - Jeff

And from the comments, I understand that this quote is not totally accurate, as a brain state is not a memory, and we just simply dont know enough about memory to do this kind of calculation, but the start of this quote is the key to the question:

Memories are represented in the brain as patterns of firing neurons.

Meaning yes, if the brain lost activity for a moment, its stored information would likely deteriorate tremendously, if not completely. That is, if my cited information (and interpretation of its effect) is correct.

  • $\begingroup$ @Jeff I quoted you. $\endgroup$
    – J.Todd
    Dec 5 '14 at 2:47
  • $\begingroup$ I cannot find any academic literature about memories being "firing neurones". Could you cite this please? I am quite interested! $\endgroup$
    – James
    Dec 5 '14 at 3:03
  • $\begingroup$ @GoodGravy This is only information that I'm citing from another answer. It could be incorrect. $\endgroup$
    – J.Todd
    Dec 5 '14 at 3:05
  • $\begingroup$ I found an article about persistent firing in the prefrontal cortex (responsible for conscious thought), and as a part of the study they found persistent firing, but this wasn't really to do with short or long term memory as far as I can deduce, but rather "working memory". As described in my answer, this firing is still the result of chemical cellular states, so might not take as big a hit as it would first seem. $\endgroup$
    – James
    Dec 5 '14 at 3:13
  • $\begingroup$ There is no basis for this conclusion: "Meaning yes, if the brain lost activity for a moment, its stored information would likely deteriorate tremendously, if not completely. That is, if my cited information is correct." Since the neurons themselves are still extant and unchanged, as soon as they started firing again, the memories would all be there. $\endgroup$ Dec 5 '14 at 3:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.