Introducing: Multilingual Nuns
September 16, 2019 1:25 PM   Subscribe

What multilingual nuns can tell us about dementia A strong ability in languages may help reduce the risk of developing dementia, says a new study. The research examined the health outcomes of 325 Roman Catholic nuns who were members of the Sisters of Notre Dame in the United States. The data was drawn from a larger, internationally recognized study examining the Sisters, known as the Nun Study.

But wait, there's more!

The Nun Study, begun in 1986 with funding by the National Institute on Aging, focuses on a group of 678 American Roman Catholic sisters who are members of the School Sisters of Notre Dame. Studying a relatively homogeneous group (no drug use, little or no alcohol, similar housing and reproductive histories, etc.) minimizes the extraneous variables that may confound other similar research.

"The study shows quite dramatically how pathology alone can often mislead."

"possible correlations between linguistic abilities in early life and the presence of AD pathology with and without clinical manifestations in late life were considered."

What Nuns Are Teaching Us About Alzheimer’s


NYT from 2001: https://www.nytimes.com/2001/05/07/us/nuns-offer-clues-to-alzheimer-s-and-aging.html

Various scholarly articles from SceinceDirect

Proof Positive: Can Heaven Help Us? The Nun Study – Afterlife

Cheerful sisters live longer than unhappy nuns

Atlas Obscura, 2016: The Neurologists Who Fought Alzheimer’s By Studying Nuns’ Brains
posted by bq (21 comments total) 24 users marked this as a favorite
 
"The researchers found that six per cent of the nuns who spoke four or more languages developed dementia, compared to 31 per cent of those who only spoke one. However, knowing two or three languages did not significantly reduce the risk in this study, which differs from some previous research."

Sad trombone noise.
posted by grumpybear69 at 1:46 PM on September 16, 2019 [9 favorites]


Did they name the languages?
posted by Apocryphon at 1:50 PM on September 16, 2019 [1 favorite]


How do they decide what 'speaking' a language means?
posted by os tuberoes at 2:17 PM on September 16, 2019 [1 favorite]


Doesn't seem like there's much to see here:

In terms of the analysis of linguistic ability in this study, it is important to remark that the linguistic scores were available in only 93 out of 678 sisters enrolled in the Nun Study. This explains the availability of these data in only 14 subjects in this study. To circumvent this limitation, we opted to compare the linguistic scores between subjects with or without cognitive deficits. Despite the small number of subjects available, a significant association between higher idea density scores and the absence of cognitive deficits was observed.

The headline data comes from 14 subjects. Only 2 of these were in the really interesting group: people who had Alzheimers-like lesions in the brain but showed no symptoms of cognitive loss. It's pretty easy to get statistical significance when you can fit your entire study group inside a large van.

To be fair to the researchers, the main thing they were looking at was neuron hypertrophy; people with lesions but no symptoms had larger neurons than usual, which could be the result of some kind of compensation. Science media of course took the juciest bit they could find and blew it out of proportion. Same ol' story as always.
posted by echo target at 2:59 PM on September 16, 2019 [5 favorites]


Today, it is known that plaques and tangles are the two most important pathological features of Alzheimer’s disease. However, some stunning results from the Nun Study show that Alzheimer’s is not a yes/no disease. Rather, it is a process…

For example, approximately one third of the sisters whose brains were found to be riddled with Alzheimer’s plaques and tangles at autopsy had shown no symptoms and scored normal results in all mental and physical tests while alive! Though the opposite result was true in other cases; such contradictory results show that there is much more to Alzheimer’s than neurological changes in the brain alone.


This is a big deal, since eliminating or preventing these "plaques and tangles" has been the target of most experimental therapies over the past 25 years. And probably why they've all failed. There seems to be something "spooky" about Alzheimer's, as there is with back pain, which also often fails to correlate with physical presentation.
posted by Modest House at 3:27 PM on September 16, 2019 [9 favorites]


Did they name the languages?

Heptapod B.
posted by GuyZero at 4:00 PM on September 16, 2019 [12 favorites]


I think it's interesting that they studied nuns at all given how little research is done on women in general. That's pretty cool.
posted by bleep at 4:14 PM on September 16, 2019 [9 favorites]


Did they name the languages?

This may be an opportunity to find out what language God speaks.
posted by justsomebodythatyouusedtoknow at 6:36 PM on September 16, 2019


It's pretty easy to get statistical significance when you can fit your entire study group inside a large van.

This is backwards, actually; you’re more likely to be able to distinguish an effect of a given size from background noise if you have a large sample.

(Where’d you see the quoted bit? I couldn’t find it. It looked to me like the linked article above the fold is still paywalled.)
posted by eirias at 6:57 PM on September 16, 2019 [1 favorite]


It's pretty easy to get statistical significance when you can fit your entire study group inside a large van.

No, it's pretty easy to get statistical significance when your sample includes tens of thousands of people. If it's a handful of people even a very very large difference is unlikely to be statistically significant.
posted by If only I had a penguin... at 7:56 PM on September 16, 2019 [3 favorites]


Quoted bit is from the paper. What I meant about the sample size is that if you're not picky about what correlation you're looking for, you can find some significant and interesting correlations in any small group. Say my uncle died of cancer but my aunt didn't; turns out pierced ears prevent cancer, p < 0.05! This is basically what people are talking about when they refer to p-hacking: you do your study, then go see what interesting correlations you can pull out of it.

approximately one third of the sisters whose brains were found to be riddled with Alzheimer’s plaques and tangles at autopsy had shown no symptoms

Don't forget we're talking about seven or eight people here, total.
posted by echo target at 8:01 PM on September 16, 2019 [1 favorite]


Say my uncle died of cancer but my aunt didn't; turns out pierced ears prevent cancer, p < 0.05!

Please show the math where this gives you a p value below .05. It certainly does not. P-hacking tends to work best with large samples. Small samples will give you the large differences but not the P you seek.

Don't forget we're talking about seven or eight people here, total.

We're talking about 325 people here, total.

And I just looked at the article and I don't see where you're getting 7 or 8 people. Here's what I see in the article
Of 325 participants, 33.5% (n = 109) developed dementia at some point during the 11 waves of follow-up (Table 1). Overall, age, number of follow-up assessments, and presence of an APOE ɛ4 allele were significantly associated with dementia. There was no significant association between dementia and multilingualism overall. However, significantly fewer participants speaking four or more languages developed dementia compared to those speaking fewer than four languages (5.9% [1/17] versus 35.1% [108/308]; p = 0.01). Participants who developed dementia were significantly older at baseline than participants remaining dementia-free (mean 83.8 versus 81.7 years; p  LT  0.001) (Table 1). Almost half (47.5%; 28/59) of APOE ɛ4 carriers developed dementia compared to 30.5% (81/266) of non-carriers (p = 0.02).
I think the finding that it requires 4 languages not 2 or 3 made me wonder if there is just a very small linear effect, so it's not going to show up as significant until you hit 4. If that's the case, then this might be statisticaly significant, though not substantively so. I was thinking I'd like to see them model "number of languages" instead of dummies for each number. But then look at Table 2! The odds ratios show it's not linear at all or even monotonic. Nuns who speak 2 or 3 languages have HIGHER odds of dementia than those who speak 4+.

That's very odd and it does suggest something besides "that fourth language makes all the difference!" is going on here.
posted by If only I had a penguin... at 8:11 PM on September 16, 2019 [4 favorites]


Nuns who speak 2 or 3 languages have HIGHER odds of dementia than those who speak 4+.
This should be higher odds than those who speak just one. They have higher odds than those who speak 4+, too, but that's obviously less interesting.
posted by If only I had a penguin... at 8:21 PM on September 16, 2019


We're looking at different articles. This is the article being used in the main link (According to the citation at the bottom of the main link.)
posted by If only I had a penguin... at 8:32 PM on September 16, 2019


Oes-Day Ig-Pay Atin-Lay ount-cay?
posted by theora55 at 8:48 PM on September 16, 2019


Tyas is quoted as: "it makes sense that the extra mental exercise multilinguals would get from speaking four or more languages might help their brains be in better shape", which posits a kind of exercise effect, but I'm thinking that perhaps it's the very ability to acquire 4 languages in the first place that's correlated with the ability to maintain brain health / performance. In particular because the exercise effect seems negligible in people who "only" acquire 3 languages.
posted by dmh at 3:25 AM on September 17, 2019 [2 favorites]


but I'm thinking that perhaps it's the very ability to acquire 4 languages in the first place
Yep, and the ability to acquire 4 language is not just cognitive but social. The opportunity to acquire 4 languages is not uniformly distributed. I know they control for class, but of course these kinds of social differences are very hard to control for completely and there's surely unobserved heterogeneity.
posted by If only I had a penguin... at 4:42 AM on September 17, 2019


Does speaking in tongues count as a language? Just asking for a friend.
posted by mareli at 11:52 AM on September 17, 2019


You're thinking of the pentacostal nuns.
posted by If only I had a penguin... at 12:57 PM on September 17, 2019


I have a sister who is a polyglot and a sister. I sent this article to her.
posted by dances_with_sneetches at 6:10 AM on September 18, 2019 [1 favorite]


Yep, and the ability to acquire 4 language is not just cognitive but social.

But that's why knowing what the four languages are is important. Are they four similar Romance languages? Does knowing both Spanish and Portuguese help, even though they are similar? Given these are Catholic nuns, was Latin one of the languages- even though it wasn't spoken in everyday conversation? Or do you need to know four from separate language families and be fluent enough to speak at any point?
posted by Apocryphon at 12:32 PM on September 19, 2019


« Older What’s interesting about American redstarts?   |   Mother of Dragons(ounds) Newer »


This thread has been archived and is closed to new comments