Statistics in fMRI studies: mere voodoo?

Statistics in fMRI studies: mere voodoo?

“Do you think the media are partly responsible for sensationalizing the findings of social neuroscience? And how can the media do a better job of reporting on brain scanning data?

Ed Vul: In general, I would advocate a bit more skepticism on the part of reporters, with respect to all scientific findings. I think reporters generally try to write up conclusions in slightly grander terms than the scientists used originally. What they may not realize is that scientists themselves have often oversold the implications of their findings a bit. You put these things together and you can end up with really overblown coverage. (On the other hand, perhaps if this advice were followed, science columns would end up dull and unread, so perhaps I should withdraw the suggestion.).”

This is from an interview of Ed Vul, a graduate student with an inquisitive mind.

Ed Vul

He has an article in press which caused big waves in the small community of social neuroscientists and neuroeconomists. In this paper, he makes a strong critique of the statistical methods used to correlate a behavioral trait with a particular brain region – which is the bread and butter of fMRI studies. For the interested readers, here is the exchange in chronological order:

Ed Vul and al., article in press: Voodoo Correlations in Social Neuroscience

Original Rebuttal by incriminated scientists (including Tania Singer, a neuroeconomist at Zurich), new version of their rebuttal – and they work on an article-length version as well.

Rejoinder by Ed Vul (to the first version of the rebuttal)

Interview of Ed Vul for Scientific American

My bet on the final issue of this debate, for what it is worth? From the quick look I had on the papers, it seems that “regression to the mean” is a central issue in this statistical debate. And ah! if there is one topic where nobody agrees on (among and between statisticians, biologists, and economists), this is this one.* So in my humble opinion, the debate is ripe for taking a turn that is very common in these cases: “It always ends in statistics“. Participants will retort with increasingly sophisticated and intractable analytical refinements, obfuscating the core issue that draw a large audience to the debate in the first place.

* See an article by Stephen Stigler on the topic

Read Full Post »

Big hopes

A recurring aspect of the discussions in social neuroscience is the big, indeed massive, hopes generated by the field. Social neuroscientists claim they will one day help cure diseases, handicaps, improve memory, learning capacity,  etc.

To see a perfect example of this rethoric, have a look at this short video by the University of Reading. I just have to say: it worked on me, I was amazed. Have a look by yourself:

.

[from a post by Sylvain on this very good French forum on neuroscience, Ovule Neuroscience.]

Read Full Post »

The “pop” genre in science

Carwash companies want to monitor your brain

Pop music has a distant cousin: pop science. I first encountered the “pop” adjective before “science” when reading about ethology, the continental tradition in the study of animal behavior. The revival of the notion of instinct (Konrad Lorenz), and the flourishing of animal studies in natural conditions (apes in particular) led to the publication of books claiming that humans had deeply-rooted instinctual behaviors, after all. Culture would just be a superficial layer sliding on top of our strong biological nature.  Those books became massive best sellers: The Naked Ape by British zoologist and surrealist painter Desmond Morris is a representative example. Published in 1967, it had sold 8 millions copies by 1979. To compare, the Selfish Gene by Richard Dawkins has sold “just” over a million copies in 30 years.

Pop-ethology often includes among distinct features: a taste for human-animal analogies in any form (“humans are just apes” or “apes are remarkably human”), a penchant for conservative orientation (to choose from: homosexuality is a behavior caused by urban overcrowding, gender inequalities perform an adaptive function, …), a romanticizing of the scientific endeavor, and a glorification of untouched nature.

I had ended by thinking that pop-ethology was distinctively unique, a genre caused and circumscribed by human / higher primate analogies, in the 60s and 70s. But I come to realize that it could be better understood as just an instance of a broader pop-genre that renews itself regularly. And pop-neuroscience is the new trend, it seems.

The same way pop-ethology was primarily a literary genre, but was also expressed in documentaries and cinema movies, one has a variety of pop-neuroscience media. My favorite so far is a the website of a carwash company trying to make sense of how focusing on the reptile brain can boost its sales.

I think there are “areas of development” for the pop genre in neuroscience, and I see at least two: evolutionary claims, and over-optimistic mapping.

Evolutionary claims is when carwashers try to bypass your consciousness and speak directly to your reptilian brain, to make you choose extra options (“calming fear in the customer will make him/her choose the shiny polish at $8.”). This kind of pop-neuroscience is very much like pop-ethology used to be: that’s still the old “Humans are just like apes” brand argument, except that you climb a step further down on the evolutionary ladder, from apes to reptiles.

The second is over-optimistic mapping. Cognitive neuroscientists have even a name for it: grandmother cells. Briefly, this is the belief that high order social phenomena (like the seeing of a grandmother’s face) have a one-to-one relation with the firing of single neurons (the neuron for recognizing grandmothers). How nice would it be for cashwashers if they could find the stimulus that triggers the firing of the “washing my car frantically” neuron! Ah, dreams, fantastic dreams…

The trick is, some single neurons do fire for quite complex phenomena, like with the visualization of faces or hand-shaped stimuli. But from there to the “grandmother cell”, there is still a gap. Pop-neuroscience loves to bridge the gap, by claiming regularly that this or that brain’s area is responsible for this or that complex social function.

Anyway, I have not yet any big conclusion on that, it just strikes me how strong the pop-genre seems to be in social neuroscience. To the point of overshadowing the rest?

Read Full Post »

The new science without a name

The Baptism of Christ, Andrea del Verrocchio and Leonardo da Vinci, 1472-1745

In speech and writing, it occurred to me that I could simply not name the scientific field I was studying. “The neuro- approach in social science” is a clumsy definition for a field that has not been been labeled yet. For now, we just have the above definition, or a repetitive list: neuro-marketing, neuro-management, neuro-linguistics, neuro-politics, etc…

“Naming” might be seen as a trivial thing, not closely as interesting as the substantive subject that labels designate.  But it can almost always reveal things. To mention something I know well, “sociobiology”, the name in itself had an interesting impact. Edward Wilson said that he chose it in reference to an existing sub-field in animal social behavioral studies, and hence it would be a sufficiently known term to designate his own work in social behavior, Sociobiology (Wilson 1975).

As it turned out, nobody knew about the obscure sub-field Wilson referred to, but some critics said that this name did remind them of something: the German and nazi “social-biologie”. Critics used this to support their claim that Wilson’s book was Social Darwinism in new clothes. With such a negative association, a number of biologists a priori sympathetic to Wilson’s research either refused to be called sociobiologists, or if they accepted the label, were indicted for that. Today, “sociobiology” is used in some very specialized articles about the social behavior of monkeys, but that’s about all.

This episode is not forced to happen again, even if that it is still a very open possibility [for an on-going feud about Social Darwinism in economics, – “I am not!”, “Yes you are, and a fascist one!” -, look at here. I’ll blog it soon in the Playground).

So… what for a good name for the field? “Cognitive neuroscience”? To me, it is more a designation of basic capacities such as language, vision or memory, not for a choice between 2 lotteries or attitudes on the financial market. “Neuroeconomics”? This is more or less the direction taken for now, but with an understandable reluctance from other social scientists to be taken under the imperialist covering arm of economics. So far, neuroeconomists I have met are themselves keener to be considered as students of “decision making”. That, I think, showcases their strong relationship with psychology, something they do not want to give up.

The game is then open: “neuro-social science?” “social neuroscience?” “brainomics”? My own preference goes for social neuroscience. A label will appear soon anyway, as journals have to named, and short titles for grant application have to be found. Just wait for a year or two, that’s my guess.

Read Full Post »