Hacker News new | past | comments | ask | show | jobs | submit login
Differences between expert and novice brains in mice: study (cshl.edu)
213 points by known 19 days ago | hide | past | web | favorite | 61 comments

I wrote a literature review about this [1] in 2015 I think. I mostly look at support from fMRI studies in Shogi players (Japanese chess). While it was a school assignment, I took it very seriously as I knew the review itself would have very serious implications on my personal life.


Main question: how can one develop an effective intuition?

The development of an effective intuition relies on: a high validity environment, a high amount of practice and a high interoceptive awareness. The first two elements account for intuitive answers being right. The third element accounts for the gut feeling being felt by the person. As a result the brain changes, and integrates certain brain areas more into the default mode network. The idea that if one has high activation in a brain area a lot, then the structure will be changed.

High validity environment: rules are easily understandable and deterministic (think chess), or you have a lot of practice moments (think poker). Understanding psychological patients would be an example of a low validity environment.

[1] https://melvinroest.github.io/articles/intuition.pdf

"As a result the brain changes, and integrates certain brain areas more into the default mode network. The idea that if one has high activation in a brain area a lot, then the structure will be changed."

In your review you cite cross-sectional research, not a longitudinal, within-subject effect of change with study/practice/learning. Therefore your conclusions should not be about change, but on group differences. Those difference might have arisen from the training, or pre-existed in those individuals that were more prone to study/practice/learn/meditate.

Post-Note: Sorry. As a developmental neuroscientist, you hit on a pet-peeve. Cheers.

Fair enough, in group differences it's tough to be certain about causality. It's interesting that my teacher never gave this as feedback. Well, that just shows that when it comes to HN, people simply care more about science (as my teacher was a neuroscience researcher).

Personally, I do think the brain is plastic in this sense though. I haven't looked deep enough into it to be a 100% certain. Nevertheless, from all the research summaries I read, I gather that whenever a skill is practiced in a focused manner for 1 to 2 hours, neural plasticity occurs.

I wish there was a literature review around this idea. But I got the feeling that 4 years ago the literature was mostly on clinical patients. Normal people and neural plasticity seems to be researched less well.

Is intuition always a literal gut feeling (or rather, an embodied feeling, in the robotics sense), such that having no interoceptive awareness would be a hindrance to developing it?

I always figured people were speaking mostly metaphorically about "gut feelings." Certainly, some kinds of intuitions (e.g. ones involving disgust or fear) are held onto as "physiological state" (i.e. written through motor-cortex commands; read back through pattern-detection in embodied cognition.) But I would expect most intuitions to present as e.g. sudden emotional states, or other weird purely-mental qualia.

I'm biased, from my own literature review I gathered that one could strengthen their intuitive signal via meditation, which is basically feeling your body (not only your gut though).

Intuitive feelings that I acted on were all body-related in my experience. In most cases, these intuitive feelings always had to do with assessing who another person is. My intuition about that seems to be pretty accurate.

I think there's a lot more happening in your body than you might realize. I've noticed that myself by practicing mindfulness. However, I don't think it rules out your idea that some intuitions are purely mental.

By the way the word emotion is a tough one, as I believe there is always a bodily component part to an emotion. I subscribe to the James-Lange theory [1] and the appraisal theory. In part, because I experience emotions in those ways before I even knew about the names and knew that these theories existed.

In short how I view emotions is process-based:

- An event occurs (thought, feeling, someone talks to you)

- This automatically gets processed by you whether you want it or not (though when it's automatic, obviously you're not using your full cognitive resources on it)

- This automatic processing concludes into two things:

1. You get thoughts about it

2. You get feelings about it (feelings in your body)

Note: thoughts and feelings are immediately re-triggering this process. So it gets meta pretty quick as you can have feelings about feelings about thoughts about feelings about <mix_and_match_as_long_as_you'd_like>

During my meditation retreat I realized that feelings come slightly earlier than thoughts. At least that's how I experienced it and I happened to be laser focused on observing my feelings and thoughts.

Coincidentally, that's also why I believe I have no free will as my feelings give rise to my thoughts, even when I simply observe my feelings. I don't choose that these feelings give rise to thoughts (and there were times where I really didn't want them to give rise to thoughts), therefore there's no free will. But that's another story :D

[1] There's probably a more advanced theory of this one as it is pretty old. For practical purposes, this theory was good enough for my understanding on what emotions are.

Love this mindset; far too many people think of the System 1 part of our brains as things to fear and reduce as much as possible. Training it is vastly superior.

An expert having learned a magnitude more has magnitudes more experience than a novice. This education and experience has taught him (or equally her) how much there is left to learn and how little we really know. In our field, we have learned humility... no, not really learned, so much as have had it beaten into us so harshly that there's no ignoring it. It becomes a part of us.

Conversely, this lack of experience and humility and the bubbling exuberance and over abundance of confidence that comes with inexperience tempts people to walk a path that someone with more experience dare never venture. It's these paths that help move us forward into the unknown.

Our confidence comes from different places. An expert's from the confidence that they know their place in the world. They've learned their reality, they've learned to build tools to deal with and capitalize on that reality. Their experience has taught them how to survive and, if they're lucky, prosper. While a novice's confidence is often borne out of idealism, hope, arrogance, ego and sheer audacity.

There is not only room for this in the world but, I believe, a need for it. That we need each other for humanity to move forward.

Not sure about the top part. I know several programmers, mathematicians, etc. who think they are geniuses and other fields are trivial.

Some people I guess never grow out of their ego... perhaps it's my naivete, but I'd like to think they're the exception rather than the rule.

What comes first, the limited perspective or the ego?

So basically pruning the neural network makes it faster. Isn’t this effectively what we are doing since birth? As babies we just have a bunch of semi-random connections, and then we prune away as we learn.

And in AI it’s basically the same. The fewer number of hidden nodes in your neural network the better, as long as there are enough nodes to accomplish the task.

Except the mind is creating new neural connections throughout life. The idea that neuroplasticity is a baby's only phenomenon is a common myth. In fact, we can see new connections being formed whenever new skills and memories are formed in the elderly, and everyone has the capacity to make new connections at any time. Sure, the RATE of those connections being made is lower over time, we are constantly decelerating in our ability to add new skills.

Making new connections doesn't preclude pruning old ones (or pruning new ones for that matter). And in fact the brain is doing both all the time.

I just didn't think it was relevant to this discussion to go deeply into the details of neurogenesis.

Babies are born with the capacity to learn language, motor skills, etc. The brain is very structured at birth for these capacities. So no, it isn't random, but probabilistic. Meaning the structure of the brain and connections are probabilistically formed via evolution. Or is this what you meant by semi in semi-random?

In ML, think of evolution as hyper-parameter tuning. Most ML models have a defined structure but are tuned via learning. Pruning is probably such a tuning strategy in the brain.

Yes that’s what I meant by semi-random. I was trying to keep it as layman as possible. :)

Something... something... Kolmogorov complexity... lost in thought.

A really good book on this is "Pragmatic Thinking and Learning: Refactoring Your Wetware" by Andy Hunt (of the Pragmatic Programmer series of books). I particularly like the explanation of the Dreyfus model of skill progression: Novice, Advanced Beginner, Competent, Proficient, and Expert.

I've heard some very vague grumblings about the Dreyfus model but never worked up the motivation to actually dig into them.

But the first thing I noticed is that here's a guy writing a paper about skill progression and he hasn't encapsulated the act of writing a paper about skill progression as a progression of its own.

I suspect it's more like Novice, Advanced Beginner, Competent, Proficient, Expert, Artist, Philosopher.

I like the four stages of competence (https://en.wikipedia.org/wiki/Four_stages_of_competence) which reads:

1. Unconscious incompetence The individual does not understand or know how to do something and does not necessarily recognize the deficit. They may deny the usefulness of the skill. The individual must recognize their own incompetence, and the value of the new skill, before moving on to the next stage. The length of time an individual spends in this stage depends on the strength of the stimulus to learn.[5]

2. Conscious incompetence Though the individual does not understand or know how to do something, they recognize the deficit, as well as the value of a new skill in addressing the deficit. The making of mistakes can be integral to the learning process at this stage.

3. Conscious competence The individual understands or knows how to do something. However, demonstrating the skill or knowledge requires concentration. It may be broken down into steps, and there is heavy conscious involvement in executing the new skill.[5]

4. Unconscious competence The individual has had so much practice with a skill that it has become "second nature" and can be performed easily. As a result, the skill can be performed while executing another task. The individual may be able to teach it to others, depending upon how and when it was learned.

> 4. Unconscious competence

This can end up being an uncomfortable position when people ask you to explain exactly why you made a decision, and even though you probably could explain it with some effort, it's difficult because you made your choice based on experience-derived intuition. Being at the conscious competence stage may actually require less work in general, unless one gets to work with other unconscious competent people.

I have this a bit with the Object Oriented Programming principles.

The code structure just needs to feel right. It's hard to throw principles at it to show which structure is better. To me it always feels like there are so many aspects to take into account, that cannot just be expressed by priciples.

Some code feels right, and some doesn't. Sometimes extending code "falls into place" and confirms that your previous reasoning was the right one. And other times it just remains a complex mess ;).

This is why language tests like TOEFL and IELTS have 2 levels of mastery above native language speaker.

When you're a native English speaker, you're in the unconscious mastery phase for the most part. Things feel right or wrong.

The 2 levels above are being able to explain why. That's when you're at the "able to teach English" level.

I think the same applies to code. You can be very very good at something, but true mastery is being able to impart that knowledge on others as well.

> You can be very very good at something, but true mastery is being able to impart that knowledge on others as well.

Is that true though? How do we know that the ability to teach is not a unique skill in and of itself rather than just an extension of one's "mastery"?

Anecdotally, I've engaged with "masters" of various fields as different as blacksmithing to theoretical physics. Of these "masters", I've encountered good teachers and bad. But I've not personally felt any correlation between their mastery of the subject (i.e. their ability to produce results in the area) vs. their ability to teach the area.

Anecdotal evidence from the boxing gym follows.

I've had various trainers in boxing. Most of them were damn good boxers. One of the best was a real talent, been boxing since he was 10. He never went very far with competing, but ho boy did he look great at the gym.

When explaining how to do something, he really struggled. Most of the time he could show us what he does, but not explain how he does it, why he does it, or sometimes not even realize what exactly he does until he tries it. Then he's like "Oh yeah I guess I do take a little step there, you're right"

Then there's a trainer who is a little less talented himself, competed at a pretty decent level as an adult, but decided to pursue training instead. He can tell us what to do and why we should do it. Also how to think about strategies and whatnot. He's great but we need a lot of repetition and trial and error.

Then there was the lady who is an 8 time professional world champion on Muy Thai. She came to our gym to become a better boxer and to get free membership she offered to teach.

Dude, that was on another level. She could not only explain why we should do something, but also how. She'd look at your punch and be like "No no, move your wrist 5 degrees this way" and damn did it immediately make an impact.

3 months of working with her created more boxing improvement than 3+ years of other trainers has. It was phenomenal.

The difference was that she wasn't just talented, she also had a deep understanding of exactly what you have to do, why you have to do it, and how you have to do it. Her advice was measured in half-inch adjustments to your technique.

Teaching might be a separate skill, but true mastery is being able to break down exactly how you do what you do and why.

Oh and better yet: she explained how it should feel when we get it right. So not only would we know what to do, but also how to tell if we did it right.

In Pirsig's Zen and The Art of Motorcycle Maintenance he discusses mastery of writing and language as beyond that of robotic regurgitation of language rules. In fact, most languages have many exceptions and inconsistencies among their rules, and being a good writer is not defined by them, and even can invove stretching or breaking them.

I also have taught English to novice speakers and sometimes am struck by such questions they receive, sometimes being all of: obtuse, irrelevant and pedantic at the same time. But that is their primary objective: to sell tests and teaching materials to their tests, not direct competency.

I think it makes sense, because if you know the right choice wihout being able to explain why, you also can’t easily realise if this case is in fact an exception to the usual case from your experience.

> because if you know the right choice wihout being able to explain why

But teaching isn't simply being able to explain why something is the way it is. It's more than that.

I could explain to you why planets move the way they do by using the language of tensor fields built up out of tensor products mapping vectors to their duals. It would be a thoroughly accurate definition of why we see the movement we see but it would not necessarily be a good way to teach. Particularly if the recipient of the teaching wasn't familiar with the formalism of general relativity.

It's this aspect, that teaching is more than simply understanding that makes me think it potentially could be orthogonal to mastery of a subject.

I also think the mistakes you make are worse when you get to that level. You don't make many at that point, but the ones you do make tend to come out of nowhere and surprise you a lot more because you're not expecting to make them. Whereas at that conscious competent level you're still paying more attention to every little thing. Or there's those silly little mistakes you make just because you're not really thinking about the process and working unconsciously. Those things that you really should know better, haven't done for years but overlooked that one time because, you're just going for it.

This book is an excellent examination of this concept: https://www.amazon.com/Mastering-Mountain-Skills-Brian-Lopes...

The author had almost no mountain biking skills and a friend who was one of the best in the world. The book was built up during the author’s advancement through the stages by observing the expert, who was largely incapable of communicating exactly how he was doing what he was doing.

That's sometimes the case with unconscious competence, but it's not implied by the definition.

The line "The individual may be able to teach it to others" addresses precisely this point. Having unconscious competence is orthogonal to being able to explain what you are doing.

Explaining your decisions is a skill in itself. When you are working in a team it is very valuable.

i imagine another factor to why explaining actions taken would be harder for a unconscious competent person is that the skill of explaining that action is a different one than performing it. but, in the same manner, that skill of explanation also has its own 4 stages of (un)conscious ignorance/competence. interesting stuff for sure

I'd gladly add another concept: how to quickly navigate into unknown territory. Experts might not know everything, but they can analyze a new situation quickly due to accumulation of concepts and overcoming new situations over the years.

ps: and ultimately, when I compare my younger self to me right now (in music) .. the only difference, is a subtle blend of patience and sensitivity. When you don't know you try small things without waiting too long or digging too deep. A newcomer may often go paralysed or stuck on the wrong intuition.

This model has always had the ring of truth to it, to me.

One amusing state you can add when you get old is something like this:

5) Semi-conscious semi-competence

The individual used to be an actual expert in something, but so many years have passed that they are only semi-competent. Yet they retain the manners and affectations of someone who is a true expert. This state tends to confuse and alarm others who are themselves competent and can’t figure out exactly what you are.

This reminds me of a saying we had in high school:

Freshmen know not, and know not that they know not.

Sophomores know not, and know that they know not.

Juniors know, and know not that they know.

Seniors know, and know that they know.

Then you graduate, grow older, perhaps go on to university, grow older still, perhaps marry and have kids, continue to grow older, and if you're lucky...

...you realize that you still "know not that you know not, know that you know not, know not that you know, and know that you know" about most everything.

Thus becoming humble and wise...?

Nope it just stops at know that you know, and then you are a brilliant leader with unconscious competence. People will go on national television to explain how you are so brilliant that you don’t even need to think about it to make the best and most great decisions.

The version of this I first heard in school was:

He knows not and knows not that he knows not is fool. Shun him.

He who knows not and knows that he knows not is a student. Teach him.

He who knows and knows not that he knows is asleep. Wake him.

He who knows and knows that he knows is wise. Follow him.

Someone said that the wise (person) knows that they know nothing.

I'd prefer to say wisdom is knowing what you know and what you don't, and it's much harder and more valuable than you might think.

Does this follow for the journey in math?

* https://terrytao.wordpress.com/career-advice/theres-more-to-...

Well, then there is a fifth stage where an experienced person does something a person in stage 2-3 would probably not, because they have enough experience to be overconfident. Thinking of skydiving accidents.

When I was young, I said to God, 'God, tell me the mystery of the universe.' But God answered, 'That knowledge is for me alone.' So I said, 'God, tell me the mystery of the peanut.' Then God said, 'Well George, that's more nearly your size.' And he told me. -George Washington Carver

The joke's on God. To truly get to the mystery of the peanut you have to go through the mystery of the universe!

This is pretty close to truth. And if all laws of physics at at work within the peanut, then all laws of the universe would be unlocked. Quite clever.

Love it. Reads like a country version of the Serenity Prayer. "God, grant me the serenity to accept the things I cannot change, Courage to change the things I can, And wisdom to know the difference."

“An expert is a person who has made all the mistakes that can be made in a very narrow field.”

― Niels Bohr

Ha. This is me now. Except I don't always know how to stop them from happening. I'm just less surprised when they happen.

an expert is someone who knows more and more about less and less, until finally he knows everything about nothing.

But more seriously from Suzuki:

the beginner's mind has many possibilities, the expert's mind has few.

From the wonderful book "Zen mind, beginner's mind"

Important note for those unfamiliar with Suzuki: this quote is usually understood to be praising the beginner's mind, which is full of possibilities and without prejudice,[0] and pointing out the weakness of the expert's mind.

[0] https://en.m.wikipedia.org/wiki/Shoshin

This is way to coherent to have been written by Bohr.

What a great perspective!

For a laymen like myself, I fail to see how this is anything more than additional evidence for Pavlovian response. Are you an "expert" when you learn that a specific pattern of actions results in a reward? That's all that's going on here -- lick the correct water spout according to the pattern of flashing lights, get a reward.

How about an MRI of a newbie in flight school and an MRI of a seasoned pilot. What does that show in terms of neural development and associated pathways? What does the newbie's brain look like after extensive training? Can you measure neural pathways over many years and see how they grow or change?

I'm not an expert in ML or mice intelligence, but it sounds like some of the concepts have a mapping in ML. The author describes a faster decision making process happening in advance of the actual decision. This sounds like the development of a LSTM, where the contextual inputs leading up to a decision are encoded in the network. The author also describes an inhibitory network. This reminds me of a GAN, where the adversarial network is discriminating on whether or not the decision will be a good one based on its knowledge.

So let me get this, if too many neurons firing then it means it is a novice's brain and less neurons firing (targeted firing) means an expert, is that right?

When given a problem, an expert might be able to solve it by recalling the solution to another problem he solved earlier, while a novice has to actually solve the problem.

"In the novice's mind there are mang possibilities; in the expert's mind there are few."

That's absolutely the case for a novice chess player vs expert.

A novice will see so many pieces. So many possibilities. They have no idea how to choose between them. And there are so many, they'll fail to see an excellent move, or that one of their pieces is about to be killed.

An expert sees far fewer moves, and so is able to look much further ahead.

Maybe the title should indicate this is about an "expert" mouse.

I'm imagining Splinter from ninja turtles.

Splinter is a rat, though. For expert mice, you probably can't beat these guys:


Is the rat aware he's being used for a scientific experiment for human entertainment and a stepping stone to the mass enslavement of other species?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact