Muscle memory – an exercise in carriage wit

Last week I became embroiled in a conversation about ‘muscle memory’ via the medium of Twitter. I don’t really do pithy at the best of time so the conversation was, for me, deeply unsatisfactory and I spent the rest of the day thinking about it. So I thought I would try and take those original posts and put together the argument I was trying to make without the 140 character constraint.

The conversation was sparked by a query from @OllieOrange2: “Is muscle memory a real thing? Surely only the brain can remember things.”

I responded to this along with @ManYanaEd who posted: “It is a silly phrase. The brain encodes memory, not muscles.” He later clarified: “Silly in the sense that it implies memory in a muscle.”

It is a flawed phrase but it is not silly. The phrase ‘muscle memory’ accurately represents people’s subjective experience of learnt processes (and for this reason alone, I think we should take it seriously). A good example would be my elder son who is learning to play the piano. He has mastered a couple of tunes to such an extent that he can bash them out without really concentrating on them (one of them being the Tetris theme tune). His fingers seem to do the playing without much conscious input. (For this reason the tune has been banned from music practice time, because he really isn’t thinking.) The key here is ‘conscious input’ once a process has been learnt then control of that process gets handed over from cortical to sub-cortical regions of the brain such as the basal ganglia and the cerebellum, to which we have no conscious access. Hence the feeling that our hands (or whichever limbs) are doing this without us. (There is some interesting and detailed discussion of this in the Neuroscientists Talk Shop podcast interview with Paul Colombo.) I suspect that @ManYanaEd would not disagree with this account.

The flaw in the phrase is that as a metaphor it draws people to think along certain lines which are misleading. So, as @OllieOrange2 observed: “I’ve definitely heard people talk about it as though the muscle is remembering the movement.” The metaphor takes people too far.

@ManYanaEd then replied: “They do and it is utterly untrue. There are no neurons in our muscles.” He later commented: “One would need a structure, neurons, to store memory. Those only exist in our brains.” This is where I started to diverge further from @ManYanaEd’s position. There are no neurons in our muscles but there are plenty of neurons projecting into our limbs and onto our muscles. This, after all, is what the Peripheral Nervous System (PNS) is. The neurons of the PNS are part of the systems that control the muscles, detect the position of our limbs, and detect aspects of the environment and any indications of damage to our bodies. So although there may be no neurons in the muscles themselves, there are neurons all around them.

This led me to an interesting question to which I don’t know the answer… Neuron-level studies of memory often focus on long-term potentiation (LTP) of neurons as an analogue for memory. This is basically where stable, long-term changes in the number of synapses between neurons, or the strength of the synaptic connections encode certain, simple forms of learning. (There are a whole host of fascinating mechanisms for changing the strength of a connection!) Typically these focus on neurons within the brain / central nervous system. But, if LTP can happen within the brain, couldn’t it happen within the PNS? Certainly some reflex actions are managed at the level of the spinal cord before the neuronal signal goes up into the brain. Why not look for LTP at this level to improve stimulus response times? (Someone has probably done the work – I’ll add it to my list of things to research!)

This brings me to another interesting topic lurking within this discussion – what sort of memory are we talking about? To go back to @OllieOrange2’s original question – we are talking about memory for learnt mechanical processes. So, to re-phrase, we are talking about stimulus-response patterns linked to certain social and physical environmental cues; in other words, tacit knowledge. (Harry Collins’ book, ‘Tacit and explicit knowledge’ is a really thorough account of tacit knowledge & how it is embedded within a social context .)

I suspect that the metaphor of memory was (perhaps) pushing @ManYanaEd in a particular direction. Towards the end of the conversation he commented that memories were something that “one had conscious access to”. It seems that he is thinking about the various forms of declarative memory, however there are various forms of non-declarative memory (such are memory for learnt processes or priming) which are not accessible to conscious thought and are probably not encoded in any linguistic form.

Actually the discussion was pushing me in a very different direction which I tried, and evidently failed, to express.

Clearly, for @ManYanaEd, memory was information, stored in neurons in the brain. If that is your working definition of memory then that is pretty much the end of the conversation. As @ManYanaEd consistently pointed out there are no neurons in muscles so the muscles could not contain memory. But what if this is an insufficient account of memory?

The neuronal account of memory is a bit ‘leaky’… A classic example of this is a diary. I keep a note of my appointments in a diary. One day, I forget what my appointments for the day are so I look in my diary and there is the record. So where is the memory, in my head or in the diary? As my younger son pointed out, it is possible that the memory was still in my head and that the diary merely acted as a prompt to recall it. This is a fair point. But, on the other hand, in my work as a Placement Manager, I have to keep track of many threads of activity and I do this through a pair of multi-coloured spreadsheets. I am not persuaded that I actually remembered all (if any) of that information – the memory was in the file. For some people, the idea that memory can be encoded outside the head will come as no surprise – it is an element of material culture studies and cognitive archaeology. My point here is that a broader, extra-cranial account of what memory is might be more fruitful. The diary example merely serves to show that a broader account is both possible and (in some cases) necessary.

The discussion led me to think about Shapiro’s comments about ‘cognitive systems’ in his book ‘Embodied Cognition‘. In a cognitive system, objects, whilst not doing any thinking themselves, are necessary parts of a thinking process. (For example, I draw pictures and jot down notes a lot when talking through a problem so the pen and paper become part of my cognitive system.) Could you then have a ‘memory system’? Certainly diaries, rosaries and monuments would fit in such a category. Could the body constitute a sort of memory system? Recall that, in this case, memory = the long-term storage of changed and now non-conscious physical/mechanical stimulus-response patterns linked to certain social and physical environmental cues. Is it possible that the body (as opposed to the brain) is part of such a memory system? Muscles are changed by gymnastics training, as my daughter pointed out to me. Again, I’m not saying that muscles contain neuronal memories. Instead, I’m suggesting that long-term changes to musculature are a necessary part of learning/memorising those patterns and as such could be incorporated into a broader idea of memory.

If your focus is on declarative, semantic memory (i.e. networks of factual knowledge and how to handle it) this might seem a bit pointless. But I want a broader account of the ways that people work in the world so that I can figure out how object handling and reminiscence fits into that. For example, incorporating the body into this account helps when thinking about how people encounter once-familiar objects in later life.

In the end, my intention is not to question the neural understanding of muscle memory, nor am I trying to say that the ‘popular’ understanding is correct. Rather my intention is to suggest that conceiving of memory, not as the storage of information, but as the embodiment of learnt processes might prove a more fruitful avenue of questioning. For all the constraints of a Twitter conversation it was a provocative one, which I appreciate.


About Bruce Davenport

Research associate at Newcastle University. Previously a museum educator and researcher.
This entry was posted in Cognition, memory, Motor control, object handling. Bookmark the permalink.

4 Responses to Muscle memory – an exercise in carriage wit

  1. ollieorange2 says:

    Wow! Very interesting.

  2. @ManYanaEd has asked me (@suzyg001) on Twitter if this post makes sense. @ManYanaEd has pointed out that he and I have been at cross purposes in previous discussions because we were using different definitions of learning, and I agree with you that one’s definition of memory could determine whether or not one thinks ‘muscle memory’ is a silly phrase.

    I would agree with @ManYanaEd that the phrase is unhelpful because it implies that muscle has a memory, which it can’t according to most people’s definition of memory, because muscle doesn’t have the requisite structures to remember things. Having said that, it’s also true that ‘muscle memory’ is a good description of people’s subjective experience of the automation of sequences of movements involved in activities such as walking, driving or playing the piano.

    I think the clue to this might lie in myelination – the formation of a fatty myelin sheath around the nerve fibre. Myelination speeds up the transmission of electrical impulses along nerve fibres and the fibres become myelinated following a particular sequence during development. Because of this sequence, it’s often assumed myelination is genetically determined, but I haven’t yet found any reason to suppose it doesn’t proceed according to the frequency with which nerve fibres are activated. Myelination increases rapidly after birth, for example, as sensory processing increases. Fields (2005) makes the point that myelination might be the result of learning rather than biologically pre-determined.

    In short, it’s possible that sequences of movements that we use repeatedly become so strongly associated with each other and so rapidly executed, that they become automated and preconscious, rather than being handed over from the cortex to lower-level brain areas.

  3. Hi Sue,
    Thanks for your comments. I’ve come across work on myelination of neurons as it relates to glabrous and non-glabrous skin (which is helpful when thinking about different forms of touch) but not in relation to the frequency of activation. It’s an interesting idea. Recent work on the glial cells seems to show that they are much more significant for the function of neurons than was previously thought and have an active role in supporting LTP so what you’re suggesting sounds reasonable to me.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s