Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Averaging is a convenient fiction of neuroscience

thetransmitter.org

Post reply on HN
Print view View original

41–50 of 110 posts

Re: Averaging is a convenient fiction of neuroscience

#41

Post by etrautmann »

jtrueb wrote: post #6
glial wrote: post #3

All models are convenient fictions. I heard a neuroscientist once describe averaging as a low-pass filter. People know it hides high-frequency dynamics. But unless you have a way to interpret the high-frequency signal, it looks an awful lot like noise.

It is a low-pass filter in the frequency domain with a roll-off that is not smooth. I quite like [1] as a quick reference. https://www.analog.com/media/en/technical-documentation/dsp-...

Not the OP but we're talking about different things here. Much of the concern about averaging is about averaging across trials. Smoothing a spike train over time isn't really the issue that this thread is concerned with, since that's just averaging successive samples within some small window.

Re: Averaging is a convenient fiction of neuroscience

#42

Post by etrautmann »

glial wrote: post #3

All models are convenient fictions. I heard a neuroscientist once describe averaging as a low-pass filter. People know it hides high-frequency dynamics. But unless you have a way to interpret the high-frequency signal, it looks an awful lot like noise.

This is broadly speaking not correct. If you average together a bunch of trials with variable timing, then the result can tend to wash out higher frequency components (which you might not have realized were in the data), but trial averaging is not a low pass filter at all. There are some nice methods to recover temporal structure that changes across trials prior to averaging, like:

https://www.sciencedirect.com/science/article/pii/S089662731...

Re: Averaging is a convenient fiction of neuroscience

#43

Post by _yb2s »

llm_trw wrote: post #19
_yb2s wrote: post #7

I've become increasingly convinced that the idea of averaging is one of the biggest obstacles to understanding things... it contains the insidious trap of feeling/sounding "rigorous" and "quantitative" while making huge assumptions that are extremely inappropriate for most real world situations. Once I started noticing this I can't stop seeing this almost everywhere- almost every news article, scientific paper, etc.…

>In reality the modal life expectancy of adults (most common age of death other than 0) has been pretty stable in the 70s-80s range for most of human history- the low average was almost entirely due to infant mortality. This is even wronger than what you critique. For every period in history that we have good data for people had a half-life - a period in which you'd expect half of all people to die: https://batinthea…

What you said in no way conflicts with what I said. For example, if people have dangerous lives in a way that is unconnected to age they will tend to not live long, yet the modal life expectancy due to the additional mortality of actual old age can still be quite old.

What I was saying is in reference to this study I read long ago, which suggests a modal life span of about 72 years of age in the paleolithic: https://onlinelibrary.wiley.com/doi/10.1111/j.1728-4457.2007...

Re: Averaging is a convenient fiction of neuroscience

#44

Post by _yb2s »

bosma wrote: post #18
_yb2s wrote: post #7

I've become increasingly convinced that the idea of averaging is one of the biggest obstacles to understanding things... it contains the insidious trap of feeling/sounding "rigorous" and "quantitative" while making huge assumptions that are extremely inappropriate for most real world situations. Once I started noticing this I can't stop seeing this almost everywhere- almost every news article, scientific paper, etc.…

> In reality the modal life expectancy of adults (most common age of death other than 0) has been pretty stable in the 70s-80s range for most of human history- the low average was almost entirely due to infant mortality. This isn't true: https://ourworldindata.org/its-not-just-about-child-mortalit...

What I was saying is in reference to this study, which suggests a modal life span of about 72 years of age in the paleolithic: https://onlinelibrary.wiley.com/doi/10.1111/j.1728-4457.2007...

Re: Averaging is a convenient fiction of neuroscience

#45

Post by _yb2s »

JumpCrisscross wrote: post #40
_yb2s wrote: post #7

I've become increasingly convinced that the idea of averaging is one of the biggest obstacles to understanding things... it contains the insidious trap of feeling/sounding "rigorous" and "quantitative" while making huge assumptions that are extremely inappropriate for most real world situations. Once I started noticing this I can't stop seeing this almost everywhere- almost every news article, scientific paper, etc.…

Your beef appears to be with simple averages, not averaging per se . The average for life expectancy is the mean of the Gompertz distribution [1]. Specifically, one that is "left skewed and with a flattened slope for ages under 50 years for men and 60 years for women," which proceeds to become "more right skewed and leptokurtic" as age increases [2]. So a simple average in the 55 domain it would overestimate it. Whic…

What I was saying is in reference to this study, which suggests a modal life span of about 72 years of age in the paleolithic: https://onlinelibrary.wiley.com/doi/10.1111/j.1728-4457.2007...

Life spans were abnormally short in medieval Europe, even compared to ancient humans.

Re: Averaging is a convenient fiction of neuroscience

#46

Post by dalmo3 »

_yb2s wrote: post #7

I've become increasingly convinced that the idea of averaging is one of the biggest obstacles to understanding things... it contains the insidious trap of feeling/sounding "rigorous" and "quantitative" while making huge assumptions that are extremely inappropriate for most real world situations. Once I started noticing this I can't stop seeing this almost everywhere- almost every news article, scientific paper, etc.…

I came to the same realisation about a decade ago, after being a "science" enthusiast growing up. As you said once you see it you can't unsee it. Most of science is just a scam. The exceptions are those fields backed up by real world engineering. All of social and most of biological sciences are worse than useless, they are outright dangerous.

Re: Averaging is a convenient fiction of neuroscience

#47

Post by usednoise4sale »

bosma wrote: post #18
_yb2s wrote: post #7

I've become increasingly convinced that the idea of averaging is one of the biggest obstacles to understanding things... it contains the insidious trap of feeling/sounding "rigorous" and "quantitative" while making huge assumptions that are extremely inappropriate for most real world situations. Once I started noticing this I can't stop seeing this almost everywhere- almost every news article, scientific paper, etc.…

> In reality the modal life expectancy of adults (most common age of death other than 0) has been pretty stable in the 70s-80s range for most of human history- the low average was almost entirely due to infant mortality. This isn't true: https://ourworldindata.org/its-not-just-about-child-mortalit...

Our world in data pushes a clear agenda, and it isn't really to be trusted.

Consider: https://ourworldindata.org/child-mortality-in-the-past

From this article: "Researchers also collected data about hunter-gatherer societies. The 17 different societies include paleolithic and modern-day hunter-gatherers and the mortality rate was high in all of them. On average, 49% of all children died.[5]"

This is cited as coming from: https://www.sciencedirect.com/science/article/abs/pii/S10905...

Which categorically states: "Unfortunately there simply is not enough direct paleodemographic archaeological data to make definite claims about the global patterns of infant and child mortality rates of our Paleolithic hunter–gatherer ancestors."

The author of the Our World in Data piece seemingly intentionally conflates the proxy with actual archaeological evidence of the actual child mortality rates. Given the clear warning in the cited article about making definite claims, I cannot read the deception any other way.

After seeing this error, I do not know how you could possibly trust anything they have to say on the matter.

Re: Averaging is a convenient fiction of neuroscience

#48

Post by hifromwork »

dalmo3 wrote: post #46
_yb2s wrote: post #7

I've become increasingly convinced that the idea of averaging is one of the biggest obstacles to understanding things... it contains the insidious trap of feeling/sounding "rigorous" and "quantitative" while making huge assumptions that are extremely inappropriate for most real world situations. Once I started noticing this I can't stop seeing this almost everywhere- almost every news article, scientific paper, etc.…

I came to the same realisation about a decade ago, after being a "science" enthusiast growing up. As you said once you see it you can't unsee it. Most of science is just a scam. The exceptions are those fields backed up by real world engineering. All of social and most of biological sciences are worse than useless, they are outright dangerous.

>Most of science is just a scam

Most of the scientists are not scammers. So if you believe that all science disciplines other than engineering are wrong[1], you should use another word that doesn't suggest researcher malice.

[1]Which is a very strong statement, because you claim to be an expert in all science disciplines at once.

Re: Averaging is a convenient fiction of neuroscience

#49

Post by llm_trw »

_yb2s wrote: post #43
llm_trw wrote: post #19

Earlier quoted context omitted.

>In reality the modal life expectancy of adults (most common age of death other than 0) has been pretty stable in the 70s-80s range for most of human history- the low average was almost entirely due to infant mortality. This is even wronger than what you critique. For every period in history that we have good data for people had a half-life - a period in which you'd expect half of all people to die: https://batinthea…

What you said in no way conflicts with what I said. For example, if people have dangerous lives in a way that is unconnected to age they will tend to not live long, yet the modal life expectancy due to the additional mortality of actual old age can still be quite old. What I was saying is in reference to this study I read long ago, which suggests a modal life span of about 72 years of age in the paleolithic: https://…

>What you said in no way conflicts with what I said.

Neither would a world with 88 people with the following death schedule.

    | Age     | Deaths/Year |
    |---------+-------------|
    | 0 to 5  |           2 |
    | 5 to 70 |           1 |
    | 70      |           3 |
I'm sure that the 3 people who make it to 70 are very happy they are in the highest mode of the distribution. The 85 who did not may have something to say about how meaningful using the mode of a distribution is.

Re: Averaging is a convenient fiction of neuroscience

#50

Post by wodenokoto »

llm_trw wrote: post #19
_yb2s wrote: post #7

I've become increasingly convinced that the idea of averaging is one of the biggest obstacles to understanding things... it contains the insidious trap of feeling/sounding "rigorous" and "quantitative" while making huge assumptions that are extremely inappropriate for most real world situations. Once I started noticing this I can't stop seeing this almost everywhere- almost every news article, scientific paper, etc.…

>In reality the modal life expectancy of adults (most common age of death other than 0) has been pretty stable in the 70s-80s range for most of human history- the low average was almost entirely due to infant mortality. This is even wronger than what you critique. For every period in history that we have good data for people had a half-life - a period in which you'd expect half of all people to die: https://batinthea…

Parent is talking about mode, so the most common number. Let’s build a distribution that satisfy both.

Oldest age is set 80, mode is set to 75. We start by distribution one percentage point to each age. Then add an extra to the mode and all the remaining to zero.

We now have a distribution where the most common age of death, other than zero be in 70-80, and more than half the population die before they reach 40.

Post reply on HN