6 ms·

>"Since it generally takes around 10 years to develop a drug, they made those choices before a lot of the alternatives had any data to back them up." Its possi

by nonbel 8y ago

>"Since it generally takes around 10 years to develop a drug, they made those choices before a lot of the alternatives had any data to back them up." Its possible to act on a "working hypothesis" without it becoming anything close to "unthinkable" to consider other hypotheses. For some reason this amyloid hypothesis became well accepted and ingrained without good reason. >"The fun thing is that even if you consider the amyloid hypothesis to be false, that doesn't make any of the alternative theories any more likely to be true" This violates Bayes' theorem. >"In fact, the amyloid hypothesis could still be true, but nobody has found the right way to tackle the problem." Why don't the supporters of the amyloid hypothesis make a definitive prediction, then if it doesnt work out the hypothesis can be dropped?

db48x 8y ago
> Its possible to act on a "working hypothesis" without it becoming anything close to "unthinkable" to consider other hypotheses. For some reason this amyloid hypothesis became well accepted and ingrained without good reason. This is my point. It was never literally true that alternatives were "unthinkable"; the word "unthinkable" was a small bit of hyperbole on the part of the author of the quote we have been discussing. Just don't lose sight of the fact that the amyloid hypothesis did and still does have good evidence for it. (In addition to a whole load of new and very expensive evidence against it.)
- nonbel 8y ago
  >"Just don't lose sight of the fact that the amyloid hypothesis did and still does have good evidence for it." What is this evidence? The evidence I know of is quite weak. There are a few correlations here and there, and the "amyloid hypothesis" has never even yielded some kind of quantitative model to be tested. Instead it just keeps getting more and more byzantine".
  - db48x 8y ago
    Yea, the evidence for it was never amazing, in absolute terms. But how much of that is hindsight? The alternative theories have even worse evidence, so the evidence for it was obviously good enough to convince a large number of people to spend a decade or more on it. If you think that the evidence was so weak that everybody who worked on it is crazy, then you're wrong. It'll take years of work before anyone has an alternative theory with as much going for it.
    - nonbel 8y ago
      >"But how much of that is hindsight?" I can't say personally. As commented elsewhere, I first looked into it about 2014 and almost immediately thought this amyloid hypothesis should have died years ago. >"the evidence for it was obviously good enough to convince a large number of people to spend a decade or more on it" This doesnt mean much to me. If thats where the funding is then people will link their research to it. If NIH is about "aging" this year then you change your grant to use "aged rats", or somehow link your study to "climate change" or whatever. >"If you think that the evidence was so weak that everybody who worked on it is crazy, then you're wrong." Crazy, no. Personally, I think it was a near total lack of quantitative thought. Stuff like people teaching/learning "APP and presenilin mutations are linked to Alzheimers", when it was only single digit or less percent of cases that could account for: >"The first mutation causing the familial form of the disease was identified in the amyloid precursor protein (APP) gene on chromosome 21.8 When investigating other families with the familial disease, several additional APP mutations were found. However, these mutations explain only a few familial cases. Instead, mutations in the highly homologous presenilin 1 (PSEN1) and presenilin 2 (PSEN2) genes account for most cases of familial disease.9,10 However, the familial form of the disease is rare, with a prevalence below 0·1%.11" https://www.ncbi.nlm.nih.gov/pubmed/16876668 https://www.ncbi.nlm.nih.gov/pubmed/16876668
      - db48x 8y ago
        >> the evidence for it was obviously good enough to convince a large number of people to spend a decade or more on it > This doesnt mean much to me. If thats where the funding is then people will link their research to it. If NIH is about "aging" this year then you change your grant to use "aged rats", or somehow link your study to "climate change" or whatever. I don't mean doing research with grant money, I mean that hundreds of people at half a dozen or more companies spent billions of dollars of private funding and at least a decade of their lives to develop drugs based on the amyloid hypothesis. That means that the evidence was good enough, or the evidence for any alternatives bad enough, to speculate on. Drug development is always speculative, but it's not crazy or irrational. And it's not a lack of quantitative thought, either; it's speculation. But my purpose here wasn't to defend the evidence (as you seem to want me to do), it was merely to point out that for all this time "the" amyloid hypothesis (again, it's a family of hypotheses because each drug has a different intended mechanism of action and is therefore a different testable hypothesis) has been the only workable theory of how the disease works. Thus someone used the word "unthinkable", which you objected to. He clearly didn't mean it literally, since people have actually thought about alternatives, so it must have been hyperbole. Since he probably agrees with you that the evidence is weak in absolute terms, I don't see why you would object to the hyperbole. I pointed it out because your question (which was "What made it so previously unthinkable?") indicated that you were taking the word literally.
db48x 8y ago
>> The fun thing is that even if you consider the amyloid hypothesis to be false, that doesn't make any of the alternative theories any more likely to be true > This violates Bayes' theorem. Not quite. If I understand the math correctly, this would only be true if we had a complete enumeration of all the other possible theories. Since the real cause of Alzheimer's could be something nobody has thought of yet, disproving a single theory would not add significant evidence in favor of the theories we have enumerated.
- nonbel 8y ago
  Bayes rule is just: Pr(H0|Data) = Pr(H0)Pr(Data|H0)/sum(Pr(H[0:n])Pr(Data|H[0:n])) If the data indicates you lower the probability of any hypothesis 0:n, it means the others all become more likely. There is no reason this has to be by a significant amount though.
  - db48x 8y ago
    Yea. My point was that with no way to fully enumerate the set of hypotheses, you can't do that sum. Thus the change is insignificant.
    - nonbel 8y ago
      You dont need to do the exact sum. Small values in the denominator have negligible effect and can be dropped. Also, "you cant do the sum, therefore its insignificant" makes no sense.
db48x 8y ago
> Why don't the supporters of the amyloid hypothesis make a definitive prediction, then if it doesnt work out the hypothesis can be dropped? Did you read the article? Every drug manufacturer who came up with a therapy made a definitive prediction. One predicted that drug A would break down the plaques and people would recover. Disproven. Another predicted that drug B would slow down plaque formation and people would recover. Disproven. Another predicted that drug C could bind with and disable protein A which normally disables protein B which removes amyloid plaques (and thus drug C would have the effect of letting protein B work longer before getting down-regulated) and that people would recover. Disproven. But we come back to the reason why the someone used the word "unthinkable": a lot of researchers think that whatever the cause of Alzheimer's turns out to be, it will have to explain why amyloid plaques exist in Alzheimer's patients. (If it were just a random coincidence, you would expect the plaques to show up in people without Alzheimer's, but they don't.) So everyone who disproved at great expense some theory about amyloid plaques is going to look for some related theory that they can try, since that won't require that they start from scratch. They can test early treatment, or a combination therapy, or... I think you can see that the brain is complicated enough that "amyloid plaque in the brain causes Alzheimer's" is not quite detailed enough to test. No one has figured out how the brain works well enough to prove that the plaques cause the damage (or even to quantify in objective terms at the neural level what the damage is), and the formation and removal of plaques is a horrible web of interactions between potentially thousands of proteins. Thus every specific, testable theory ends up being quite similar to a lot of other specific, testable theories. Disproving one doesn't disprove them all. So the best way to disprove the amyloid hypothesis might turn out to be finding a cure by some other route. The exiting thing is that people are starting to actually try this, with real money. Perhaps in 10 or 15 years we'll know if they succeeded. Or it might take longer.
- nonbel 8y ago
  >"(If it were just a random coincidence, you would expect the plaques to show up in people without Alzheimer's, but they don't.)" I don't see what "random coincidence" has to do with it, there can be a correlation for a pretty much infinite number of reasons besides the amyloid hypothesis. Anyway this is wrong. It seems to be that ~20-30% of "cognitively normal" elderly have substantial plaques. This increase with age. The earliest reference I found for this info was 1937: >"there is now building data on Aβ deposition in healthy controls that suggest at least 20% and perhaps as much as a third of healthy older adults show significant deposition." https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2844114/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2844114/ >"An attempt was made to estimate the severity of the histologic alterations as a whole, and it was found that severe intellectual impairment was apt to occur in cases showing extensive involvement. However, this parellelism was neither close or constant. There were enough exceptions to make one reject the idea that there could be any simple and direct quantitative relationship between clinical and pathologic changes. For example, in case 4 (Table i), a patient with a paranoid type of senile psychosis, there was only a minimal amount of intellectual impairment though the changes in the brain, particularly with respect to plaques, were very extensive." https://ajp.psychiatryonline.org/doi/pdf/10.1176/ajp.93.4.757 https://ajp.psychiatryonline.org/doi/pdf/10.1176/ajp.93.4.75... >"There is a close relationship between the degree of dementia measured during life and quantitative neuropathological changes found post mortem, with two important exceptions: (1) brains of nondemented elderly patients sometimes contain pathological features in a degree diagnostic of Alzheimer’s disease; and (2) conversely, there are occasional brains of demented subjects that do not contain markers of Alzheimer’s disease or of other disorders known to produce dementia that can be detected by neuropathological examination. The discovery of such exceptions led some neuropathologists in the 1950s to doubt the importance of Alzheimer brain changes as a cause of dementia in the elderly 1161." https://www.ncbi.nlm.nih.gov/pubmed/2897823 https://www.ncbi.nlm.nih.gov/pubmed/2897823 >"Ten subjects whose functional and cognitive performance was in the upper quintile of the nursing home residents, as good as or better than the performance of the upper quintile of residents without brain pathology (control subjects), showed the pathological features of mild Alzheimer's disease, with many neocortical plaques. Plaque counts were 80% of those of demented patients with Alzheimer's disease." https://www.ncbi.nlm.nih.gov/pubmed/2897823 https://www.ncbi.nlm.nih.gov/pubmed/2897823 >"Recent advances in amyloid imaging have made it possible to observe Aβ amyloid accumulation in the patient's brain. As a result, it has been found that there are many normal patients with amyloid deposits, and also AD patients with very few amyloid deposits (Edison et al., 2007; Li et al., 2008). Further, in the brain of elderly non-demented patients, the distribution of senile plaques is sometimes as extensive as that of dementia patients (Davis et al., 1999; Fagan et al., 2009; Price et al., 2009; Chetelat et al., 2013)." https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5797629/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5797629/

by nonbel 8y ago

db48x 8y ago

> Its possible to act on a "working hypothesis" without it becoming anything close to "unthinkable" to consider other hypotheses. For some reason this amyloid hypothesis became well accepted and ingrained without good reason. This is my point. It was never literally true that alternatives were "unthinkable"; the word "unthinkable" was a small bit of hyperbole on the part of the author of the quote we have been discussing. Just don't lose sight of the fact that the amyloid hypothesis did and still does have good evidence for it. (In addition to a whole load of new and very expensive evidence against it.)

nonbel 8y ago
>"Just don't lose sight of the fact that the amyloid hypothesis did and still does have good evidence for it." What is this evidence? The evidence I know of is quite weak. There are a few correlations here and there, and the "amyloid hypothesis" has never even yielded some kind of quantitative model to be tested. Instead it just keeps getting more and more byzantine".
- db48x 8y ago
  Yea, the evidence for it was never amazing, in absolute terms. But how much of that is hindsight? The alternative theories have even worse evidence, so the evidence for it was obviously good enough to convince a large number of people to spend a decade or more on it. If you think that the evidence was so weak that everybody who worked on it is crazy, then you're wrong. It'll take years of work before anyone has an alternative theory with as much going for it.
  - nonbel 8y ago
    >"But how much of that is hindsight?" I can't say personally. As commented elsewhere, I first looked into it about 2014 and almost immediately thought this amyloid hypothesis should have died years ago. >"the evidence for it was obviously good enough to convince a large number of people to spend a decade or more on it" This doesnt mean much to me. If thats where the funding is then people will link their research to it. If NIH is about "aging" this year then you change your grant to use "aged rats", or somehow link your study to "climate change" or whatever. >"If you think that the evidence was so weak that everybody who worked on it is crazy, then you're wrong." Crazy, no. Personally, I think it was a near total lack of quantitative thought. Stuff like people teaching/learning "APP and presenilin mutations are linked to Alzheimers", when it was only single digit or less percent of cases that could account for: >"The first mutation causing the familial form of the disease was identified in the amyloid precursor protein (APP) gene on chromosome 21.8 When investigating other families with the familial disease, several additional APP mutations were found. However, these mutations explain only a few familial cases. Instead, mutations in the highly homologous presenilin 1 (PSEN1) and presenilin 2 (PSEN2) genes account for most cases of familial disease.9,10 However, the familial form of the disease is rare, with a prevalence below 0·1%.11" https://www.ncbi.nlm.nih.gov/pubmed/16876668 https://www.ncbi.nlm.nih.gov/pubmed/16876668
    - db48x 8y ago
      >> the evidence for it was obviously good enough to convince a large number of people to spend a decade or more on it > This doesnt mean much to me. If thats where the funding is then people will link their research to it. If NIH is about "aging" this year then you change your grant to use "aged rats", or somehow link your study to "climate change" or whatever. I don't mean doing research with grant money, I mean that hundreds of people at half a dozen or more companies spent billions of dollars of private funding and at least a decade of their lives to develop drugs based on the amyloid hypothesis. That means that the evidence was good enough, or the evidence for any alternatives bad enough, to speculate on. Drug development is always speculative, but it's not crazy or irrational. And it's not a lack of quantitative thought, either; it's speculation. But my purpose here wasn't to defend the evidence (as you seem to want me to do), it was merely to point out that for all this time "the" amyloid hypothesis (again, it's a family of hypotheses because each drug has a different intended mechanism of action and is therefore a different testable hypothesis) has been the only workable theory of how the disease works. Thus someone used the word "unthinkable", which you objected to. He clearly didn't mean it literally, since people have actually thought about alternatives, so it must have been hyperbole. Since he probably agrees with you that the evidence is weak in absolute terms, I don't see why you would object to the hyperbole. I pointed it out because your question (which was "What made it so previously unthinkable?") indicated that you were taking the word literally.

db48x 8y ago

>> The fun thing is that even if you consider the amyloid hypothesis to be false, that doesn't make any of the alternative theories any more likely to be true > This violates Bayes' theorem. Not quite. If I understand the math correctly, this would only be true if we had a complete enumeration of all the other possible theories. Since the real cause of Alzheimer's could be something nobody has thought of yet, disproving a single theory would not add significant evidence in favor of the theories we have enumerated.

nonbel 8y ago
Bayes rule is just: Pr(H0|Data) = Pr(H0)Pr(Data|H0)/sum(Pr(H[0:n])Pr(Data|H[0:n])) If the data indicates you lower the probability of any hypothesis 0:n, it means the others all become more likely. There is no reason this has to be by a significant amount though.
- db48x 8y ago
  Yea. My point was that with no way to fully enumerate the set of hypotheses, you can't do that sum. Thus the change is insignificant.
  - nonbel 8y ago
    You dont need to do the exact sum. Small values in the denominator have negligible effect and can be dropped. Also, "you cant do the sum, therefore its insignificant" makes no sense.

db48x 8y ago

> Why don't the supporters of the amyloid hypothesis make a definitive prediction, then if it doesnt work out the hypothesis can be dropped? Did you read the article? Every drug manufacturer who came up with a therapy made a definitive prediction. One predicted that drug A would break down the plaques and people would recover. Disproven. Another predicted that drug B would slow down plaque formation and people would recover. Disproven. Another predicted that drug C could bind with and disable protein A which normally disables protein B which removes amyloid plaques (and thus drug C would have the effect of letting protein B work longer before getting down-regulated) and that people would recover. Disproven. But we come back to the reason why the someone used the word "unthinkable": a lot of researchers think that whatever the cause of Alzheimer's turns out to be, it will have to explain why amyloid plaques exist in Alzheimer's patients. (If it were just a random coincidence, you would expect the plaques to show up in people without Alzheimer's, but they don't.) So everyone who disproved at great expense some theory about amyloid plaques is going to look for some related theory that they can try, since that won't require that they start from scratch. They can test early treatment, or a combination therapy, or... I think you can see that the brain is complicated enough that "amyloid plaque in the brain causes Alzheimer's" is not quite detailed enough to test. No one has figured out how the brain works well enough to prove that the plaques cause the damage (or even to quantify in objective terms at the neural level what the damage is), and the formation and removal of plaques is a horrible web of interactions between potentially thousands of proteins. Thus every specific, testable theory ends up being quite similar to a lot of other specific, testable theories. Disproving one doesn't disprove them all. So the best way to disprove the amyloid hypothesis might turn out to be finding a cure by some other route. The exiting thing is that people are starting to actually try this, with real money. Perhaps in 10 or 15 years we'll know if they succeeded. Or it might take longer.

nonbel 8y ago
>"(If it were just a random coincidence, you would expect the plaques to show up in people without Alzheimer's, but they don't.)" I don't see what "random coincidence" has to do with it, there can be a correlation for a pretty much infinite number of reasons besides the amyloid hypothesis. Anyway this is wrong. It seems to be that ~20-30% of "cognitively normal" elderly have substantial plaques. This increase with age. The earliest reference I found for this info was 1937: >"there is now building data on Aβ deposition in healthy controls that suggest at least 20% and perhaps as much as a third of healthy older adults show significant deposition." https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2844114/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2844114/ >"An attempt was made to estimate the severity of the histologic alterations as a whole, and it was found that severe intellectual impairment was apt to occur in cases showing extensive involvement. However, this parellelism was neither close or constant. There were enough exceptions to make one reject the idea that there could be any simple and direct quantitative relationship between clinical and pathologic changes. For example, in case 4 (Table i), a patient with a paranoid type of senile psychosis, there was only a minimal amount of intellectual impairment though the changes in the brain, particularly with respect to plaques, were very extensive." https://ajp.psychiatryonline.org/doi/pdf/10.1176/ajp.93.4.757 https://ajp.psychiatryonline.org/doi/pdf/10.1176/ajp.93.4.75... >"There is a close relationship between the degree of dementia measured during life and quantitative neuropathological changes found post mortem, with two important exceptions: (1) brains of nondemented elderly patients sometimes contain pathological features in a degree diagnostic of Alzheimer’s disease; and (2) conversely, there are occasional brains of demented subjects that do not contain markers of Alzheimer’s disease or of other disorders known to produce dementia that can be detected by neuropathological examination. The discovery of such exceptions led some neuropathologists in the 1950s to doubt the importance of Alzheimer brain changes as a cause of dementia in the elderly 1161." https://www.ncbi.nlm.nih.gov/pubmed/2897823 https://www.ncbi.nlm.nih.gov/pubmed/2897823 >"Ten subjects whose functional and cognitive performance was in the upper quintile of the nursing home residents, as good as or better than the performance of the upper quintile of residents without brain pathology (control subjects), showed the pathological features of mild Alzheimer's disease, with many neocortical plaques. Plaque counts were 80% of those of demented patients with Alzheimer's disease." https://www.ncbi.nlm.nih.gov/pubmed/2897823 https://www.ncbi.nlm.nih.gov/pubmed/2897823 >"Recent advances in amyloid imaging have made it possible to observe Aβ amyloid accumulation in the patient's brain. As a result, it has been found that there are many normal patients with amyloid deposits, and also AD patients with very few amyloid deposits (Edison et al., 2007; Li et al., 2008). Further, in the brain of elderly non-demented patients, the distribution of senile plaques is sometimes as extensive as that of dementia patients (Davis et al., 1999; Fagan et al., 2009; Price et al., 2009; Chetelat et al., 2013)." https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5797629/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5797629/