NPR reports on a study giving volprioc acid to adults and training them on pitch (singing):

Hensch is studying a drug which might allow adults to learn perfect pitch by recreating this critical period in brain development. Hensch says the drug, valprioc acid, allows the brain to absorb new information as easily as it did before age 7.

"It's a mood-stabilizing drug, but we found that it also restores the plasticity of the brain to a juvenile state," Hensch tells NPR's Linda Wertheimer.

Brain plasticity is useful for a whole lot more than learning pitch. As the article notes it would be invaluable for training one's ear to pick up sounds of foreign languages, but also it seems reasonable to this commentator that high levels of plasticity during rationality training or other forms of self-development would result in more transformative results.

New to LessWrong?

New Comment
35 comments, sorted by Click to highlight new comments since: Today at 9:47 PM

http://en.wikipedia.org/wiki/Valproic_acid

The foremost and most severe concern for anyone taking valproic acid is its potential for sudden and severe, possibly fatal, fulminating impairments in liver and impairments of hematopoietic or pancreatic function, especially in those just starting the medication. This particular warning is the first one listed on any drug adverse effect listing when one receives the drug at the pharmacy.

Yeah, messing with valproate as a mind-hack is a potentially Darwinian endeavour. It's nasty shit in many ways, i.e. it's medicine that people don't take without a clear need and regular medical monitoring, and is not to be fooled with.

One problem I see with this kind of study is that valproic acid has a very distinct effect (from personal experience), which makes it easier for participants to determine whether they are in the placebo group. It would be nice if there were an "active placebo" group who took another mood stabilized that is not an HDAC. Also, it would have been nice to see the effect on ability to produce a tone by humming or whistling, given the pitch name.

Some very weak anecdotal evidence in favor of the hypothesis: For a couple months in 2005 I was being treated with valporic acid and, during that time, I took an undergraduate course in topology. In my brief stint as a graduate student (2012), I also took topology and performed much better in this than in any of my other courses, though this could just be due to liking the subject.

24 participants in the study (11 of which took valproic acid), and the linked article hailing the new wonder-drug doesn't even get its name right.

The article is crap but referring to the sample size without considering the baseline success rate is misleading. If, say, the task were to be creating a billion dollar company, and the treated group had even one success, then that would be quite serious evidence for an effect, just because of how rare success is.

As is usus, consider such data to be within expected parameters unless mentioned.

/me waits for Gwern to chime in

Other than accent, is there anything that children learn better than adults? Is there even anything that adults do not learn much faster?

I'm wondering if recovery from blindness, or other sensory impairments, would benefit from such a substance, if used correctly.

Yes, some animals (generally predators) have a critical period for developing sight. Unlike language, this is an easy faculty to test in animals. One could try this drug on cats that had been blindfolded during the critical period (a standard experiment).

Experiments of this general sort have already been done. I haven't read deeply yet, but exploring the biblio it appears that the novelty here is that something that worked in animal models by a semi-well-understood mechanism was demonstrated to also work in humans via mere oral administration of an already FDA-approved substance.

Epigenetic treatments of adult rats promote recovery from visual acuity deficits induced by long-term monocular deprivation

We found that chronic intraperitoneal administration of valproic acid or sodium butyrate (two different histone deacetylases inhibitors) to long-term monocularly deprived adult rats coupled with reverse lid-suturing caused a complete recovery of visual acuity, tested electrophysiologically and behaviorally. Thus, manipulations of the epigenetic machinery can be used to promote functional recovery from early alterations of sensory input in the adult cortex.

I'd bet the learning effect is quite generic but I suspect that most of the things that would be really useful for humans would take some pedagogic intervention, and designing the post-administration training process is probably a non-trivial task. Something I'd bet would probably work is building up great smell/chemistry associations ("this smells like it has a ring with a sulfur in it") but even collecting a lot of different samples and their chemical structures for paired presentation would take a bunch of work.

I think the smell example sort of illustrates some of the other limits... Just as with Anki, a non-obvious hard part is simply figuring out what things are even worth the effort of tracking down, leaving aside whether they should be acquired, retained, and folded deep into one's neural circuitry.

I'd bet the learning effect is quite generic

How generic? to all things said to have a critical period (mainly basic use of the senses)? or more broadly? If the latter, do you have an answer to my original question: what do children learn faster than adults?

[-][anonymous]10y50

Any kind of formative skill development. Adults are great (or can be) at learning new applications of those skills, but not so much at forming entirely new skills. Look at children learning motor skills verses stroke patients, for example.

Do you have a reference discussing the comparison between children and stroke patients?

[-][anonymous]10y40

Personal experience, being both a parent and volunteer at a stroke rehabilitation center.

Would you record your experience?

[-][anonymous]10y10

I would if I had the time. I'll add it to my someday/maybe list a mile long :\ Making a positive singularity by finishing my AGI cones first...

Consider prioritizing goals you can conceivably accomplish over those with most glitz.

[-][anonymous]10y10

AGI is my field, developmental psychology is not. I am actually working on AGI infrastructure day-to-day. I am not writing papers in developmental psychology, stroke rehabilitation, or nurology.

I'm not even clear what you're saying. Should I be avoiding doing things that have any possibility of glitz associated with them? Am I incapable of accomplishing anything others might find meaningful?

Languages in general.

Maybe for first languages, but it's hard to do that experiment. It's widely held false belief that children are good at learning second languages. wikipedia has the hilarious quote:

Certainly, older learners of a second language rarely achieve the native-like fluency that younger learners display, despite often progressing faster than children in the initial stages. This is generally accepted as evidence supporting the [critical period hypothesis].

[-][anonymous]10y10

Why is that hilarious? It is in line with the studies I've read. I'm on my phone so it is a little hard to find cites. With focused study and the proper learning environment adults achieve fluency faster, but it us rare that they ever achieve a native like accent. Children with exposure under the age of 11 typically have no trouble in the other hand

Children with exposure under the age of 11 typically have no trouble [achieving a native-like accent] in the other hand

I started learning French at 10, and Spanish and Italian at 12, so this seems to predict I would have a native-like accent in French but not in Spanish or Italian. As a matter of fact, I do think my accent is slightly better in French, but I don't think the difference is large enough to count as "native-like" versus "non-native-like", and furthermore I suspect it has more to do with patterns of study and practice well after the ages in question (e.g. some systematic training in French phonetics at age 16) than with anything that went on in my brain during the first year.

[-][anonymous]10y10

1) Anecdotal evidence doesn't really mean anything. Everyone develops slightly differently - 11 years isn't a hard wired rule. These are rough averages. I expect that in the data there were some nine year olds that had difficulty assimilating, and some 13 year olds that had no trouble.

2) What the early exposure does is for example train your brain on distinguishing and producing phonemes which would otherwise literally be imperceptible otherwise.

3) French, Spanish, and Italian are very similar languages, in structure, idioms, and phonetics. It is not uncommon for adult learners of these languages coming from related backgrounds to develop a near-native accent within a realistic study regimen. This is because there are not that many perceptual hurdles between these European languages - if you've learnt one of these languages as a child, you have most of the mental machinery necessary to handle a native accent in one of the others. IIRC these studies are often more about Vietnamese or Chinese kids learning English or vice versa, where the problems are much, much greater and the prior of adult success very low.

The quote is about fluency, not accent. The first sentence is correct. The second sentence is the hilarious part; I read it as "linguists are generally idiots." It is harder for me to survey linguists than the literature, but in my limited attempt, it appears correct.

It's widely held false belief that children are good at learning second languages.

I am not sure it's false, that may depend on how do you define "children".

I personally know a couple of kids who were forced to learn a different language around the ages of 4-5. The process went much faster and easier than with adults.

[-][anonymous]10y120

To expand on Douglas_Knight's answer, if it seems counter intuitive that is because children spend basically all their waking hours in a learning environment for nearly two decades of their life. Most of that time is spent learning or using in some way one or more languages.

The typical adult method for learning a second langauge, on the other hand, is to spend an hour or two a week in a classroom or with a tutor. No wonder it doesn't work as well.

To compare apples-to-apples, consider for example the Monterey Naval Postgraduate School which trains American soldiers and intelligence officers. Using a full-immersion, 24/7 learning environment they are able to take adult learners from zero to near practical fluency in months to years (depending on the difficulty of the language). Similar results are reported with Peace Core volunteers, for example, at least those which find themselves in a fully non-English environment.

No, it doesn't depend on how you define "children." People get continually better at learning second languages, up at least to age 16. For every aspect of language (except accent) that people have measured, older people learn faster. Forget your anecdotes and read the literature.

If you're appealing to the literature, it would be good form to provide references.

Would this include humans less than four years of age?

I've only seen one study that included 3 year-olds. And it didn't include many, so it didn't break them out and only concluded that the range of 3-5 did worse than 6-7.

From the original study:

Importantly, this result was not due to a general change in cognitive function, but rather a specific effect on a sensory task associated with a critical-period.

[-][anonymous]10y00

Experiments of this general sort have already been done. I haven't read deeply yet, but exploring the biblio it appears that the novelty here is that something that worked in animal models by a semi-well-understood mechanism was demonstrated to also work in humans via mere oral administration of an already FDA-approved substance.

Epigenetic treatments of adult rats promote recovery from visual acuity deficits induced by long-term monocular deprivation

We found that chronic intraperitoneal administration of valproic acid or sodium butyrate (two different histone deacetylases inhibitors) to long-term monocularly deprived adult rats coupled with reverse lid-suturing caused a complete recovery of visual acuity, tested electrophysiologically and behaviorally. Thus, manipulations of the epigenetic machinery can be used to promote functional recovery from early alterations of sensory input in the adult cortex.

I'd bet the learning effect is quite generic but I suspect that most of the things that would be really useful for humans would take some pedagogic intervention, and designing the post-administration training process is probably a non-trivial task. Something I'd bet would probably work is building up great smell/chemistry associations ("this smells like it has a ring with a sulfur in it") but even collecting a lot of different samples and their chemical structures for paired presentation would take a bunch of work.

I think the smell example sort of illustrates some of the other limits... Just as with Anki, a non-obvious hard part is simply figuring out what things are even worth the effort of tracking down, leaving aside whether they should be acquired and retained.

[This comment is no longer endorsed by its author]Reply

I was hoping for a minute that this would be able to truncate the horrid use of auto-tune but then I remembered I mostly, luckily, stay away from the radio regardless of its use.

Would love to see a larger scale study with more specifics done with this especially with regards to language acquisition.