Can a snippet of hair from an infant reveal whether that child will have or develop autism? On the US television show "Today" 5 January 2023, Evan Bush reported on research about diagnoses of autism. He discussed studies about systematic analyses of the chemical composition of children's hair.

The idea is that the chemicals in children's hair can reflect environmental toxins to which the children have been exposed during their infancy and toddlerhood. Those chemicals might be associated with higher chances of autism. The story is wrapped in the importance of early diagnosis. Early identifiation is helpful!

I imagine that the question about hair assessment is causing quite a stir in the autism community. Indeed, the Kaiser Health News, the New York Post (Kato, 2023), and other sources have run stories about the development, even though Laura Dattaro (2022) described preliminary evidence in January 2022. Such reports can be galvanizing, electrifying, and I suspect that this story will be widely discussed. I caution readers, though, to take a calm, reflective view of this story. Let me explain.

First, however, here's the "story," as represented in Mr. Bush's lede

''Researchers have developed a first-of-its-kind test for autism that they say can find markers of risk in a single strand of hair, an innovation that might help clinicians identify it in young children before they miss developmental milestones.

The test — which is still in the early stages of development by the startup LinusBio and a ways from federal approval — is a diagnostic aid, meant to assist clinicians in identifying autism but not to be relied on alone. Because hair catalogs a history of exposures to metals and other substances, the technology uses an algorithm to analyze it for patterns of particular metals the researchers say are associated with autism.

Mr. Bush's report is appropriately cautious. He noted that "they say" and the innovation "might help." At the least, those are weasel words, but they are appropriately cautious.

Does the research merit certainty?

I read the original research. Here's my take.

Autin et al. (2022) reported that they had found a way to predict diagnoses of autism based on bio-metabolic analyses of young children's hair. The research is not conclusive. One can't send a snippet of her toddler's hair to a lab and get a diagnosis. At best, it's just a prediction.

The big idea here is that measurement of one variable at one time will accurately predict a subsequent outcome. This ides is fundamental concept in measurement. I want to return to it later in this post. First, though, here's the abstract of the article:

''Autism spectrum disorder (ASD) is a neurodevelopmental condition diagnosed in approximately 2% of children. Reliance on the emergence of clinically observable behavioral patterns only delays the mean age of diagnosis to approximately 4 years. However, neural pathways critical to language and social functions develop during infancy, and current diagnostic protocols miss the age when therapy would be most effective. We developed non-invasive ASD biomarkers using mass spectrometry analyses of elemental metabolism in single hair strands, coupled with machine learning. We undertook a national prospective study in Japan, where hair samples were collected at 1 month and clinical diagnosis was undertaken at 4 years. Next, we analyzed a national sample of Swedish twins and, in our third study, participants from a specialist ASD center in the US. In a blinded analysis, a predictive algorithm detected ASD risk as early as 1 month with 96.4% sensitivity, 75.4% specificity, and 81.4% accuracy (n = 486; 175 cases). These findings emphasize that the dynamics in elemental metabolism are systemically dysregulated in autism, and these signatures can be detected and leveraged in hair samples to predict the emergence of ASD as early as 1 month of age.

I ain't so sure about the conclusions. I have concerns about the conceptual basis of the study and the scientific source. Here are those concerns:

1. The fundamental concept of the study is that children's hair reflects exposures over a recent time to environmental toxins, toxins that might be associated with later autism. If one gets snippets from kids with long, long hair, do different snippets—those closer to the head vs. those closer to the end—reveals exposure to toxins?

   1. If the snippet of hair shows exposure to certain environmental toxins, then the probability of a later diagnosis of autism is elevated. It's reasonable that the hair may reflect exposure to toxins. How much do scientists know about those toxins and autism? I'd say only a little...it'd be good to have more analyses from scientists who study environmental toxins and autism. Here's the opportunity for broad replication; we need a body of evidence.

   2. Many other experts about causes of autism examine genetic problems (not just simple inheritance, but also changes in genetic material among eggs or sperm). It's unlikely that children's hair will capture such variables. That is, just assuming that the analyses of toxins reflected in hair will pick up autism accurately is a bit of a problem is autism is caused by something else. Thus analysis of hair will need additional study.

2. What is the source of this report? The original report: seems to me to be a bit sketchy.

   1. First, the publisher appears to be one that publishes a broad spectrum of papers [https://www.mdpi.com/about/journals]. That's not bad. Reputable publishers (Sage, Wiley, APA, etc.) do so. MDPI claims to publish good journals. And it claims to be "peer reviewed," but when I checked it out, there were not many reasons to laud the journal’s faithful readers.

   2. Second, I couldn’t find much evidence that the journal promoted open science. That is, I’d like to find fundamental data from study. Can I check the facts of the study?

Summary

So, I recommend caution with regard to this finding. I recommend calmly reflecting on studies such as this one. It's a wonderful possible advance, but it may be a red herring, it may be predicated on wobbly assumptions, and it

Let's just go slow, not run out recommending haircuts and baggies of those hairs being sent to labs.

Sources

Austin, C., Curtin, P., Arora, M., Reichengerg, A., Curtin, A, Iwas-Shimada, M., Wrigth, R. O., Wright, R. J., Gemnelius, K. L., Isaksson, J., Bolte, S, & Nakayama, S. F. (2022). Elemental dynamics in hair accurately predict future autism spectrum disorder diagnosis: An international multi-center study. [Some Publisher That Uses Way Proprietary Identification], 11, 23. https://www.mdpi.com/2077-0383/11/23/7154/html#2]

Dattaro, L. (2022). FDA cites hair-based autism diagnostic aid as ‘breakthrough.’ https://www.spectrumnews.org/news/fda-cites-hair-based-autism-diagnostic-aid-as-breakthrough/

Kaiser Health News. (2023). KHN morning briefing. https://khn.org/morning-breakout/first-of-its-kind-autism-test-could-help-with-earlier-diagnosis/

Kato, B. (2023). ‘Groundbreaking’ autism test detects disorder in a single strand of hair. https://nypost.com/2023/01/06/groundbreaking-autism-test-detects-disorder-in-single-strand-of-hair/