Tag Archives: Mark Walters

Computer algorithm links facial masculinity to autism

This is certainly an interesting study, and I can’t see any obvious problem with the way it was done, but as with any study of autism, I believe questions about the validity of the diagnosis of autism must raise questions about the validity of any study of people (adults or children) who have been given that diagnosis.

Is autism a coherent, consistent, clearly-defined, clearly-delineated, natural category that explains purported cases better than alternative forms of diagnosis such as medical, genetic or sensory diagnostic categories? I doubt it. Let’s be clear; autism is nothing more than a multi-faceted description of behaviours, none of them unique to autism, and some quite common among people who have intellectual or sensory disability. There’s no biology, medicine or psychophysics in the core definitions of autism. I know of no validated, objective test designed to measure any of the sensory aspects of autism. Sure thing, autism is associated with countless congenital and genetic disorders, but the scientific validity of those categories doesn’t rub-off onto autism as a scientific category.

I’m a skeptic about the category of autism and I also have questions about diagnostic processes relating to autism and related disorders. We know that children who are purely and solely cases of prosopagnosia can be misdiagnosed with autism, and the literature on gifted and talented children includes many claims that the same can happen to G&T kids. I suspect that intelligence levels are a confounding factor in many studies that are supposed to explore autism or a broader autism phenotype, and I question whether the trend of identifying children as autistic when in the past they might have been identified as intellectually disabled was the great step forward that it is supposed to have been. There’s also the fact that the “testosterone theory of autism” has been around for many years now and has been widely popularized. It is certainly possible that parents and clinicians have been influenced to expect to see “autistic” behaviour in children who are perceived as more masculine than their peers, due to facial appearance or other traits. This conceivably could have a flow-on effect of increasing the chances that a boy or girl with masculine features might be identified as autistic, and this could be behind the effect found in this study.

These kinds of doubts are why in this blog I have never explored autism in terms of facial phenotypes or in terms of face perception deficits in any depth or with much interest. It’s not that I don’t see a problem or problems in these cases. I do, but I believe it is probable that one day in the distant future scientists will look back on the history of the sciences of the mind and wonder why we spent so much time and money researching autism, a concept that was a long, dark, gold-paved dead-end in the journey of scientific progress, while disability remained a constant issue.

Computer algorithm links facial masculinity to autism.  25 August 2017.


Hypermasculinised facial morphology in boys and girls with Autism Spectrum Disorder and its association with symptomatology.
Diana Weiting Tan, Syed Zulqarnain Gilani, Murray T. Maybery, Ajmal Mian, Anna Hunt, Mark Walters & Andrew J. O. Whitehouse
Scientific Reports 7, Article number: 9348 (2017)
06 March 2017
31 July 2017
Published online:
24 August 2017




University of Western Australia researchers’ model of face gender published in PLoS ONE

Garland, Carys Face ‘model’ accurately weighs gender points. ScienceNetwork WA. July 6th 2014.


The mathematical model of face gender that these UWA researchers have come up with seems like a sensible enough idea to me (and who am I to criticise?) but I’m very doubtful of just about everything stated about face gender and its relation to autism that is written in the Science Network article.

Gilani SZ, Rooney K, Shafait F, Walters M, Mian A (2014) Geometric Facial Gender Scoring: Objectivity of Perception. PLoS ONE 9(6): e99483. doi:10.1371/journal.pone.0099483


Oh wow! My idea is being developed by the scientists, and they are scientists in my home town.

I’ve just noticed a story that has been run on the ABC programs 7.30 WA and State to State about researchers in Perth (scientists, doctors and an orthodontist), including some at PMH, who are creating a database of normal children’s faces to create the Perth face-space project, which appears to be a tool in development for the identification of the countless rare (genetic?) diseases which have characteristic facial appearance or facies. This knowledge can be shared globally, so this is Western Australia’s gift to the world. Wow!  We can’t make really good coffee in Perth, but we can do some interesting things here. There is already a research paper by Perth researchers and a Belgian researcher published in which this type of 3D face database method has been researched as a possible tool for monitoring and discriminating a group of rare metabolic diseases in which disease progression alters the facial appearance. I’m very excited to learn about this project because the idea of using face recognition to diagnose or identify rare diseases and rare genetic syndromes is an idea that has been obvious to me for many years, as a super-recognizer who is not only able to memorize faces very well, but is also able to compare and analyse faces with a degree of unconscious skill that is probably above average. I’m sure that most people have an awareness of the significance of facial appearance.

One important consideration needs to be factored into this kind of project – the definite possibility that the characteristic facial features that are being studied can be artificially altered before the patient ever goes near a face scanner or a medical face photographer. Many good parents spend a small fortune with othodontists getting their offspring’s teeth straightened and in doing this they are often erasing one of the signs of a genetic or developmental anomaly. Children can also be the subject of plastic surgery on the face, especially if they were born with a disfiguring facial defect.

One thing that I don’t think is mentioned in this news story is the fact that the studying of faces as an element of medical diagnosis is nothing new at all and does not require any fancy new technology or photographing of patients at all. This new project looks like it will be a great refinement of an idea, but I don’t think it will achieve anything that hasn’t already been done before using human abilities alone. Descriptions of rare diseases and genetic syndromes in medical textbooks or online info sources often feature photographs of patient’s faces that display characteristic features linked with the diseases and conditions. These features can also be described in detailed technical/medical language, much like detailed technical botanical descriptions of the parts of plant species using specialized terms (jargon). Doctors who specialize in genetics or related specialties know what this all means, and they should also have a high to super-recognizer ability to recognize facial phenotypes or at least have the ability to do visual image matching/comparison from the photographs. A specialist doctor with access to photographic resources and good eyes and sound and well-connected intra-cranial face processing hardware should be able to consider the patient’s face properly in a diagnostic process, no online database needed. But this process is labour-intensive, so I can see a use for an online face database.

3D Camera used to detect disease. http://www.abc.net.au/news/2013-04-26/3d-camera-used-to-detect-disease/4654822

Stefanie Kung, Mark Walters, Peter Claes, Jack Goldblatt, Peter Le Souef, and Gareth Baynam A Dysmorphometric Analysis to Investigate Facial Phenotypic Signatures as a Foundation for Non-invasive Monitoring of Lysosomal Storage Disorders. JIMD Reports. 2013; 8: 31–39. Published online 2012 June 10. doi: 10.1007/8904_2012_152  http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3565667/

Saw, Samantha Defining normal. InkWire. April 17, 2014. http://inkwirenews.com.au/2014/04/17/defining-normal/