6 votes

Evidence for bias of genetic ancestry in resting state functional MRI

Posted by Algernon_Asimov
Tags: biology, neurology

Conference paper: Evidence For Bias Of Genetic Ancestry In Resting State Functional MRI
 [blocked]

Preprint (not peer-reviewed): Evidence for Bias of Genetic Ancestry in Resting State Functional MRI
 [not blocked]

Someone posted this on Reddit. It purports to be a study which shows that it is possible to identify a person's genetic ancestry (in other words, their "race") by observing their brain activity.

Thereby, we demonstrated that genetic ancestry is encoded in the functional connectivity pattern of the brain at rest. We hypothesize that these observed differences are a result of known ethnicity-related variations in head and brain morphology

This feels problematic, in that it gives support to the racist idea that different "races" think differently. But I don't know enough myself to believe this study or debunk it. I present it for more knowledgeable people than myself to dissect and discuss.

8 comments

  1. [5]
    JakeTheDog
    Link
    I don't see why this must be "problematic", aside from hijacking by racists (like the milk-drinking thing, which is laughable now because there are plenty of non-white cultures that are...

    I don't see why this must be "problematic", aside from hijacking by racists (like the milk-drinking thing, which is laughable now because there are plenty of non-white cultures that are lactose-tolerant). But in a mature conversation...

    In theory, this is completely plausible. Why wouldn't there be differences in brain function, on the basis of genetics?
    From the paper, the authors state that it's not necessarily differences in the neurology, let alone mode of thinking, but rather the physiology:

    However, we hypothesize that the observed differences are not based on true neuronal differences but that they originate from differences in head and brain morphology as reported in [8], [9]. These morphological differences may be carried forward through the standard rs-fMRI processing pipeline and affect the inferred functional connectivity. In addition, rs-fMRI connectivity is based on correlations between blood-oxygen-level dependent (BOLD) signal time series at rest. Thus, it is conceivable that genetic differences contributing to blood circulation, perfusion and elasticity of the vascular system may modify BOLD dynamics. This is exemplified by reports identifying ethnicity as independent risk factors for cardiovascular disease [14] and intracranial artery tortuosity [15]. In addition, brain hemodynamic responses are known to be heritable traits [16].

    Regarding the topic of thinking style: of course in practice, this is going to require a lot of validation and heavy-duty statistics, precisely because of the confounding variables (a la chopstick gene skybrian mentioned). I would even go so far as to argue that it may not be possible with modern techniques; can we really separate the variables of culture, ancestry and way of thinking?

    9 votes
    1. [2]
      skybrian
      Link Parent
      Yeah, who knows? The "chopsticks gene" thing is a caricature, but on the other hand, diet, exercise, and health care are cultural and may have effects on the brain (doesn't everything you do have...

      Yeah, who knows? The "chopsticks gene" thing is a caricature, but on the other hand, diet, exercise, and health care are cultural and may have effects on the brain (doesn't everything you do have effects on the brain?), so we're still not out of the woods.

      1 vote
      1. JakeTheDog
        Link Parent
        Oh god I wish this were true... it happens way too often to be a mere caricature. Just take a look at any -omics field; metabolomics, transcriptomics (RNA) and microbiome research especially....

        The "chopsticks gene" thing is a caricature

        Oh god I wish this were true... it happens way too often to be a mere caricature. Just take a look at any -omics field; metabolomics, transcriptomics (RNA) and microbiome research especially. Proteomics and genomics are more disciplined these days, but only because of anal pedants (i.e. statisticians) compelling these fields to self-correct.

        3 votes
    2. [2]
      Algernon_Asimov
      Link Parent
      That's exactly why I think it's problematic: because this sort of finding can and will be hijacked by racists.

      I don't see why this must be "problematic", aside from hijacking by racists

      That's exactly why I think it's problematic: because this sort of finding can and will be hijacked by racists.

      1 vote
      1. JakeTheDog
        Link Parent
        Sure, but these days, with all the noise and misrepresented science on the internet, just about anything can be problematic for one group or another. As the example with lactose tolerance (who...

        Sure, but these days, with all the noise and misrepresented science on the internet, just about anything can be problematic for one group or another. As the example with lactose tolerance (who would have thought!). Give me anything related to human differences and I will give you a racist argument.

        I'm not saying you're wrong. I'm saying at best it's a meaningless or value-less word. At worst it frames the topic as being inherently bad (like research on IQ) and skews our interpretation. In this case—and in the case of IQ—there is a lot of potential to improve our understanding of cognitive diseases; maybe certain ancestral backgrounds are more/less prone to certain mental illnesses.

        I can't find an appropriate antonym for problematic...

        4 votes
  2. [2]
    skybrian
    Link
    I don't know and it's not my area, but the first thing I'd worry about is some kind of "chopsticks gene" problem. Genetic ancestry and culture are often correlated.

    I don't know and it's not my area, but the first thing I'd worry about is some kind of "chopsticks gene" problem. Genetic ancestry and culture are often correlated.

    8 votes
  3. spit-evil-olive-tips
    Link
    From their conclusion in the preprint PDF (https://www.biorxiv.org/content/biorxiv/early/2018/10/11/440776.full.pdf): So notably, they're not saying (as the racists who are trying to misinterpret...

    From their conclusion in the preprint PDF (https://www.biorxiv.org/content/biorxiv/early/2018/10/11/440776.full.pdf):

    This work exemplifies that in the domain of rs-fMRI analysis there is a need to consider genetic ancestry as a confound in the analysis.

    So notably, they're not saying (as the racists who are trying to misinterpret this study would claim) that you can stick someone in an MRI machine, run a scan, and determine their ancestry. Instead, they're raising the possibility that it could act as a confounding variable and that other studies that use rs-fMRI may need to take this into account when designing their experiments:

    Our results indicate that genetic ancestry is a serious bias that modifies estimated brain connectivity and may mask genuine differences or may introduce spurious differences in rs-fMRI analyses between groups, e.g., a disease group and a control group.

    They also say:

    The exact origin of these apparent connectivity differences between continental ancestries remains elusive at the moment. However, we hypothesize that the observed differences are not based on true neuronal differences but that they originate from differences in head and brain morphology as reported in [8, 9].

    In other words, they don't think the differences they observe are due to any differences in thinking (neuronal differences) but rather just differences in skull size and shape that are detectable if you train a machine learning model on enough MRI scan results.

    2 votes