Bizarre Things Purported to Cause Autism: Glyphosate in Pesticides

EXECUTIVE SUMMARY: This is the first of two posts dealing with the claim by MIT researcher Stephanie Seneff that a chemical found in Monsanto's RoundUp herbicide is implicated in the increase in autism prevalence rates in recent years.

This post does not address the claim itself, but evaluates whether Dr. Seneff is a credible source.

She does appear to have credibility and standing in the field of computer science -- particularly the subfield of natural language processing -- and has a lengthy publication history and an impressive array of citations within that field. But in public health, she mostly seems relegated to fringe conferences, the Internet and one not particularly selective journal.
______________________________________________________________________________________________________________________

Sometimes I wonder if I should even be bothering to write about some of these autism hypotheses, because they're so fringe, and most people haven't even heard them, much less believe them to be true, but since I mostly write about these for my own amusement, and to satisfy my collector's instinct, the feeling soon passes.

Anyway, on Tumblr I came across a link to this article on some bogus "alternative health" website, about a senior research scientist at MIT who's been going around giving presentations (at conferences of dubious repute, like AutismOne and this anti-GMO symposium hosted by this day spa/alternative medical clinic in Groton, Massachusetts) about how pesticides --- specifically, pesticides containing glyphosate --- are turning everyone autistic.

This was one of the ads in the sidebar of that article

I will address the specifics of Dr. Seneff's claims in another post, because right now I see a golden opportunity to talk about source evaluation, and I'm taking it.

Every news article I could find on this topic had a headline along the lines of, "MIT researcher says chemical in RoundUp linked to autism." That's true as far as it goes (Stephanie Seneff is a researcher at MIT, and that is indeed what she says), but it also has the unfortunate effect of giving the reader the impression that MIT says those things, when really it's just her.

The closer you look at her background and publication history, the more red flags you see.

Her position at MIT is Senior Research Scientist at the MIT Computer Science and Artificial Intelligence Laboratory; she's a computer programmer, Jim, not a doctor!

Her advanced degrees are in electrical engineering and computer science, and her undergraduate degree --- the closest she gets to having any background whatsoever in biology or medicine --- is in biophysics.

Biophysics is a perfectly fine field of study --- it's an interdisciplinary melding of biology and physics --- but it's very broad, and it seems to me like it isn't so much a specialty in itself as it is a collection of specific research topics that require study from both biological and physics perspectives. 

Parts of it deal with atomic-level interactions between molecules: figuring out, in exacting detail, how any particular grouping of molecules fits together; how an enzyme or receptor forms a complex with its target molecule; how proteins arrive at, and maintain, their three-dimensional shape. Parts of it deal with electrostatic interactions between those same atoms and molecules, as in the study of ion channels, or the difference in electrostatic potential that exists across cell membranes.

Parts of it are more purely mathematical or computational, dealing with mathematical modeling and computer simulations.

That actually all sounds super interesting and cool, but the point is that most of those areas of inquiry are things that would benefit from investigation by very carefully chosen teams of experts in diverse fields, not one person, and it also seems to me like a bachelor's degree in biophysics (where the option exists -- it looks like biophysics is more often a graduate degree program) involves a fairly solid grounding in basic physics, chemistry, and molecular/cellular biology (as distinct from organismal or population biology, I mean) and then your choice of advanced courses in a very wide range of topics.

Here are some lists of degree requirements for a B.S. in biophysics from various universities that offer it, if you want to look at them yourself: Johns Hopkins University (PDF); York University in Canada (PDF); Arizona State University; Wake Forest University

The point of all the foregoing is just to show that someone could have an undergraduate degree in biophysics that is weighted more heavily toward the physics or chemistry ends of things, with only cursory attention paid to cell biology and little or none to metabolism or physiology. 

Anyway, it looks like most of her research over the years has been concerned with speech and language, and improving computer recognition of human speech. Her Ph.D. thesis, according to this webpage, was a computer model of how the human brain processes language.

She seems to have only pivoted to medical research in recent years.

What's worse, she seems to be writing about a very wide range of unrelated topics in medicine: the heart, the brain, the gut, epigenetics, nutrition, toxicology, epidemiology ... how much can one person understand of so many disparate fields, especially when that person is trained primarily in computer science and has only just (2011-2014) begun to publish about any of them?

Also, when you look at her CV, you notice a striking change in the nature of her citations, corresponding with the change in subject matter.

Following her pivot to writing about public-health issues, more and more of her writing is either self-published (on her website) or published in a single journal, Entropy, which has been called a "pay to play" journal -- one that will publish whatever you send them, regardless of merit, as long as you pay the fee.

For comparison, when she was writing about natural language processing stuff she would be getting published in peer-reviewed journals* published by prestigious academic and professional organizations like IEEE (the Institute of Electrical and Electronics Engineers), the Acoustical Society of America, the European Association for Signal Processing, the International Speech Communication Association, and the Association for Computational Linguistics.

The same holds true for her speaking engagements. 

In the past, she's spoken about natural language processing stuff at international conferences in various fields relating to linguistics and computer science --- the International Conference on Spoken Language Processing in 1990, 1992, 1994, 1996, 2000, 2002, 2004, 2006, and 2010; the Conference of the International Speech Communication Association in 2007, 2009, 2010 and 2011; the European Conference on Speech Communication and Technology in 1991, 1993, 1995, 1997, 2001, 2003 and 2005; the International Conference on Computational Linguistics in 1996; the Conference on Empirical Methods in Natural Language Processing in 2009; the Special Interest Group on Discourse and Dialogue in 2010; and the IEEE International Conference on Acoustics, Speech, and Signal Processing in 1991, 1992, 1994, 2000, 2008 and 2012 --- whereas her presentations about glyphosate seem to be given mostly to small gatherings of laypeople or at crank conferences, like the Weston A. Price Foundation's Wise Traditions Conference. (She's spoken at five of those!)

Given all of this, the logical thing to do is take whatever she has to say about medicine with a huge grain of salt.

* This is as good a place as any to point out that peer review isn't 100% effective at screening out dodgy science; even The Lancet managed to let Andrew Wakefield's fraudulent research slip past their vetting process. And just recently there's been a huge scandal over a "peer review ring" through which a few authors were able to fabricate favorable reviews of their submissions to make sure they would be published.