Early Human Development xxx (xxxx) xxxx

Contents lists available at ScienceDirect

Early Human Development

journal homepage: www.elsevier.com/locate/earlhumdev

Potential explanations of behavioural and other differences and similarities between males and
females with autism spectrum disorder

A B S T R A C T

Several potential explanations may be dependent on the dynamics of prenatal and postnatal testosterone in males and females, and to be consistent with Baron-
Cohen's concept of extreme male brain. This paper explores the evidence that male and female autistic subjects differ on the average in that they have had different
exposures to the causes of autism, females bearing higher genetic burdens for ASD (autistic spectrum disorder), and males having a greater exposure to high
intrauterine levels of testosterone (T). The high levels of intrauterine (and possibly postnatal) testosterone to which ASD cases have been exposed, cause a less
masculinized physical habitus (including facial features) in exposed males, and a more masculinized physical habitus in exposed females. ASD genes (as opposed to
intrauterine testosterone) are mainly responsible for a low mean IQ in ASD (especially female cases). Exposure to high intrauterine T increases the probability that
foetuses will be male, thus potentially explaining the high sex ratio (proportion male) of cases of ASD. The Gender Incoherence Model seems to be based on facts
unrelated directly to autism. The shifts towards the other sex are argued to be consequent on sex-different reactions to prenatal exposure to high T, not on the
pathology itself. The suspected underdiagnosis of female cases is partially dependent on the different proportions of environmental and genetic causes to which male
and female cases are hypothesized to have been exposed, and the consequent ‘more normal’ behaviour of female cases.

1. Introduction
Recently, strong evidence was published supporting both Baron-
Cohen's Empathizing-Systemizing theory of sex differences, and his
Extreme Male Brain (EMB) theory of autism [1]. These authors noted
that typical sex differences between healthy males and females were
attenuated in autistic subjects, and speculated that this ‘likely reflects
an interaction of biological and cultural factors’. However, some other
workers have expressed perplexity about the similarities and differ-
ences between males and females with autism spectrum disorder (ASD).
Indeed, some workers have suggested that they put into question the
concept of the Extreme Male Brain. Such doubt has been prompted by
similarities and differences in regard to the following:
1. Performances on mental rotation tests,
2. Androgynous facial features,
3. Attention to gender-related images,
4. Findings on the Cambridge Sympathy Test,
5. Findings on the Reading-the-Mind-in-the-Eyes Test,
6. Finger length ratios,
7. Behavioural differences, and.
8. Functional Magnetic Resonance Imaging, fMRI.
Our purpose is to offer potential explanations for each of these
observations to supplement the explanation cited above of Greenberg
et al. [1] (viz. an interaction of biological and cultural features). Some
of our potential explanations seem well grounded: others may be ac-
knowledged to be speculative. But, at any rate, one may conclude that
the explanation proposed by Greenberg et al. [1] is not the sole possi-
bility.
From a taxonomic standpoint, autism spectrum disorder, ASD, is an
interesting diagnosis. It is usually made on the basis of three symptoms
viz. 1. Repetitive movements, 2. Increased systematizing, and 3.
Decreased empathizing, involving impaired ability to appreciate the
minds of others and to detect its presence via their facial expressions,
thus leading to diminished eye contact. Baron-Cohen [2] characterised
this set of symptoms by the term ‘the Extreme Male Brain’. However,
these three symptoms do not intercorrelate highly, so one may suspect
that they have more than one cause. At any rate, the diagnosis of ASD
has been applied to subjects with varying behaviours, abilities and
physical habitus. This is so within ASD subjects of the same sex and as
between ASD subjects of opposite sex. This variability has puzzled some
research workers, who have interpreted such observations as putting
into question Baron-Cohen's concept of the Extreme Male Brain. The
present note has two purposes viz. 1. to summarize ASD sex similarities
and differences which have led to such puzzlement, and 2. to try to
reconcile these findings with the notion of the Extreme Male Brain by
considering them in the context of two well-supported hypotheses.
These hypotheses are now described. They have explanatory power in
this context, but, to our knowledge, have not previously been invoked
for this purpose.
2. Hypotheses that may reconcile these findings with Baron-
Cohen's concept of the extreme male brain
Hypothesis A. The multifactorial threshold hypothesis.
It is established that autism has both genetic and environmental
causes. Moreover, autism is suspected of being subject to the ‘Carter
phenomenon’ viz. the multifactorial threshold. In this, genetic and en-
vironmental causal factors are notionally additive, and, when they
reach a threshold, a case occurs. It is established that ASD females are
more likely to have affected relatives than are ASD males [3–5]. Palmer
et al. [6] presented data from 1.5 million families each with 2 children,
including 37,507 ASD cases. The overall prevalence of ASD was 1.96%
in males and 0.5% in females. Following a male proband, ASD was

https://doi.org/10.1016/j.earlhumdev.2019.104863

0378-3782/ © 2019 Published by Elsevier B.V.


diagnosed in 4.2% of female sibs and 12.9% of male sibs. Following a
female proband, ASD was diagnosed in 7.6% of female sibs and 16.7%
of male sibs. These authors suggested that these rates are consistent
with a model in which females require a higher number of genetic or
environmental risk factors to manifest ASD than males. A slight mod-
ification to that idea will be presented here. It will be argued instead
that the data favour a multifactorial threshold model in which female
cases require a higher genetic load, and male cases require a higher
environmental exposure. One such environmental cause of ASD is in-
trauterine exposure to high levels of testosterone, T ([7,8]).
Hypothesis B. The relation between prenatal and postnatal
testosterone levels in healthy males and females.
There is evidence that normal males exposed to high intrauterine
testosterone (T) levels suffer adverse postnatal hormonal consequences.
We know of no direct human experimentation on this point. However,
there is evidence in sheep (a standard animal model for human re-
production). When pregnant ewes are experimentally dosed with tes-
tosterone, their male offspring reportedly have reduced body weight,
scrotal circumference and sperm count than controls [9–11]. In con-
trast, their female offspring are phenotypically masculinized [12].
Further evidence that exposure to high levels of prenatal testosterone
masculinizes female foetuses is provided by the evidence that it causes
polycystic ovary syndrome when they are adult [13,14]. In short, pre-
natal exposure to high levels of T causes postnatal de-masculinization in
males, and masculinization in females. These phenomena occur in
otherwise healthy animals: so one may suspect that comparable phe-
nomena would happen to autistic subjects (who also ex hypothesi, have
been exposed to high intrauterine T).
Next, we note some data relating prenatal and postnatal testos-
terone (T) levels to ASD.
Brosnan [15] noted that prenatal T slows the growth of the left brain
and enhances the growth of the right brain. The right brain is associated
with visual-spatial and mathematical abilities. So it may be proposed
that some forms of autistic behaviour are directly due to prenatal
programming of the brain by exposure to high intrauterine T levels.
Strong confirmation of this hypothesis has been published e.g. by
Baron-Cohen et al. [7]. It has also been suggested [16,17] that high
adult circulating T levels contribute to autistic behaviour. The latter
authors reported that in a placebo-controlled trial of administering T to
neurotypical subjects, there was an increase of autistic traits in the
experimental group. Moreover, these authors reported that in a con-
trolled trial of human transsexuals, there was a significant increase in
autistic traits in female-to-male cases (who were treated with andro-
gens), and a decrease in autistic traits in male-to-female cases (who
were treated with anti-androgens). So the latter authors suggested that
autism might be treated by anti-androgens. This idea has not received
wide acceptance. Indeed, it has been argued that there is no appreciable
relation between adult T levels and autistic diagnoses [18]. But it
should be emphasised that single dose administrations of testosterone
have reportedly had significant behavioural effects on healthy women
[19] and men [20]. Moreover, there is a consensus that circulating
androgens are high in autistic females, and also, but less so, in autistic
males [21–24]. So one may propose, as a generalization, that mean
circulating T is high, at least at times, in cases of autism. In short, both
prenatal and postnatal T are thought to be (to varying degrees), high in
ASD.
3. Apparently anomalous findings in male and female ASD
subjects
3.1. Mental Rotation Performances
Rohde et al. [25] found no significant difference between the mental
rotation performances of males and females diagnosed with autism
spectrum syndrome. These authors wrote that ‘this does not support
2
Early Human Development xxx (xxxx) xxxx
[Baron-Cohen's] extreme male brain hypothesis of autism’. Tovainen
et al. [26] presented data on spatial behaviour in a sample of more than
13,000 twin pairs aged 19–21 years. Their data suggested that males
outperformed females on all spatial measures, including mental rota-
tion. However, they found no suggestion that prenatally transferred
testosterone (T) (as in females born in opposite-sex dizygotic twin pairs)
has any effect on mental rotation. Constantinescu et al. [27] suggested
that the early postnatal T surge may influence mental rotation in boys,
but not girls. Moreover, Pintzka et al. [28] reported a significantly
improved mental rotation performance following a single testosterone
dose to healthy women. Thus, mental rotation seems to correlate,
perhaps weakly, but positively, with postnatal circulating T levels ra-
ther than prenatal T levels. To explain the observations of Rohde et al.
[25], we invoke the two hypotheses proposed above. So we suggest that
male ASD subjects are no better than female ASD subjects on mental
rotation performance for the reason that the male ASD subjects (having
been exposed to higher intrauterine T) have moderately low postnatal T
levels that are comparable to those of ASD females.
3.2. Androgynous facial features
Women with autism spectrum disorder (ASD) reportedly have less
feminine features and larger heads than normal controls, whereas ASD
men have been judged to have comparatively feminine faces [21,29].
Bearing in mind the two hypotheses (A and B) proposed above, we
suggest that in ASD, facial morphology is partially moulded by post-
natal T levels. Ahmed et al. [30] wrote: “Abundant evidence docu-
mented in the biological and psychological literature indicates that
facial masculinity … is linked to high testosterone levels among males
…”. There is also evidence that testosterone virilizes adult female faces
[31], and that this is negatively associated with attractiveness in
women [32]. So the androgynous faces of ASD subjects could be ex-
plained by the high postnatal T in ASD women and relatively low
postnatal T in ASD men. Moreover, it should be noted that evolutionary
benefits would accrue if such a correlation (between adult T levels and
masculinity of facial features) were to exist. For instance, one evolu-
tionary benefit of such a phenomenon is that women would recognise
reproductive promise (high T levels) in the face of a masculine-looking
potential mate. Another benefit is that men would recognise the re-
productive costs of high T levels signalled by masculine faces in women.
A framework embodying this notion (of a positive correlation between
facial masculinity and circulating T in subjects of both sexes) is now
presented.
In their influential evolutionary paper, Trivers and Willard [33]
noted that in many species, males in ‘good condition’ outreproduce
females in good condition; and that females in poor condition out-
reproduce males in poor condition. These authors further noted that if
females in good condition were to produce an excess of sons, and fe-
males in poor condition an excess of daughters, then women, in general,
would produce more grandchildren. Accordingly, these authors hy-
pothesized that this actually happens. ‘Condition’ is nowhere defined by
these authors, but one strong piece of evidence supporting this pre-
diction is the finding that parents (of both sexes) with high T levels
produce an excess of sons [34]. And as was pointed out in that paper,
testosterone, at least in men, correlates positively with a number of
reproductive parameters. A bonus of this reasoning is that it would
incidentally potentially explain the (otherwise unexplained) high sex
ratio (proportion male) of autistic subjects (about 3–4 males per fe-
male) [35] because, as noted above, high T (in either parent around the
time of conception) has been hypothesized to be causally associated
with male offspring [36–38]. This hypothesis has been described by
Spiegelhalter [39] to be the ‘most coherent’ of those available to explain
the variation of sex ratio at birth.

3.3. Attention to gender-related images
Sex differences between autistic subjects have been reported in re-
gard to gender-related images. For instance, Harrop et al. [40] reported
that in contrast with boys with ASD, girls with ASD show more atten-
tion to images of toys or objects more typical of one sex than the other.
These authors also showed that girls with autism attend more to images
of faces [41]. As is suggested in the section on Hypothesis A above,
there is strong evidence that ASD females, on the average, have not
been so much exposed to high intrauterine T as ASD males. Hines et al.
(2015) noted that in human beings, T is elevated from about week 8 to
week 24 of gestation and then again during early postnatal develop-
ment. These latter authors noted that increased intrauterine exposure to
T leads to male-typical play, and to alterations to sexual orientation and
gender identity. Accordingly, we suggest that the comparatively modest
exposure of ASD females to high intrauterine T (and, as a consequence,
to more normal postnatal T) may be responsible for their more typical
attention pattern to gender-related images. In short, we suggest that the
Carter Phenomenon (Hypothesis A) may explain the sex differences in
attention to gender-related images in ASD subjects.
3.4. Findings on the Cambridge Sympathy Test
Sympathy and distress ratings, as measured by this test, reportedly
did not differ significantly as between male and female autistic subjects
[42]. These authors write: “Sympathy is a specific form of empathy and
involves both cognitive and affective empathy”. Provisionally we sug-
gest that the result of Holt et al. [42] is explained by the comparable
circulating T levels in male and female autistics as a consequence of the
difference between male and female postnatal T responses to high in-
trauterine T. We acknowledge that the test results of Holt et al. [42] are
the consequence of complex mental phenomena and so may have ad-
ditional explanations.
3.5. Findings on the Reading-the-Mind-in-the-Eyes Test
Autistic men and women reportedly show no significant difference
in performance on this test. But the performance of females with autism
differed from that of normal female controls more than the corre-
sponding difference in men ([43]). Testosterone reduces functional
connectivity during the ‘Reading the Mind in the Eyes’ test [44]. So we
suggest provisionally that comparative lack of sex difference in ASD
subjects on this test is due to the postnatal sex-different T responses to
sex differences in the levels of intrauterine T to which male and female
autistic subjects have been subjected (Hypothesis B).
3.6. Sex-differences in the 2D:4D finger length ratios of autistic subjects
It is generally agreed that the finger length ratio, 2D:4D, may be
used as a rough proxy for a subject's exposure to intrauterine testos-
terone [45], healthy men in general having lower values of 2D:4D than
healthy women. Masuya et al. [46] reported that men with ASD had
lower (more masculine) 2D:4D ratios than healthy control men. In
contrast, women with ASD had higher 2D:4D ratios than controls.
Schieve et al. [47] reported similar findings for ASD children. So one
may infer that male ASD subjects have been exposed to higher in-
trauterine T levels than female ASD subjects. Since (as noted above),
there is good evidence that intrauterine T is an environmental cause of
ASD, these data constitute powerful confirmation of the hypothesis that
ASD is subject to the Carter Phenomenon viz. that male cases have been
more exposed to environmental causes, and that female cases carry a
heavier genetic load (as argued above in Hypothesis A). Comparable
data on large samples of children were presented by Barona et al. [48]
and Schieve et al. [47]. These authors showed that (in accordance with
EMB Theory), males had lower (more masculinized) 2D:4D ratios than
females. However, these authors failed to find that ASD symptoms were
3
Early Human Development xxx (xxxx) xxxx
related to 2D:4D in females. Barona et al. [48] suggested that this latter
finding is in conflict with EMB Theory. We suggest alternatively that if
it were accepted that ASD is subject to the multifactorial threshold ef-
fect, then it would follow that female, as contrasted with male, ASD
cases, on the average, have been less exposed to high intrauterine T and
thus, less likely to yield a significant correlation between their symp-
toms and 2D:4D ratios. This line of argument may reconcile EMB
Theory with the findings and comments of Barona et al. [48] on female
ASD cases.
3.7. Behavioural differences between male and female autistic subjects
It has been reported that there are behavioural differences as be-
tween male and female autistic subjects. For example, Werling and
Geschwind [49] suggest that there are fewer restricted and repetitive
behaviours and externalizing problems in females. Rubenstein et al.
[50] concur and also noted that intellectual disability is more common
in female cases. Moreover, Antezana et al. [51] suggest that quality of
the restricted and repetitive behaviours differ as between the sexes. It
seems that sex differences in levels of pre-and post-natal testosterone
may account for most of these differences. In contrast, the established
lower mean intellectual level of ASD females seems likely to be a spe-
cific consequence of ASD genes, a higher genetic burden of which is
carried by ASD females. A relevant fact relates to cases of Asperger's
syndrome (viz. high-performing autistic people). In these, the sex ratio
(males to females) is of the order of 8–10 to 1. These behavioural dif-
ferences between male and female ASD subjects have probably led to
some underdiagnosis of females. This possibility arises as follows. Allely
[52] writes: “Females with ASD may display superficial social skills
which may mask their ASD symptomatology, impacting on the identi-
fication of the disorder – known as the ‘camouflage hypothesis’. It is
increasingly recognised that ASD females are more able than ASD males
to engage in behaviour which is socially acceptable, particularly those
females with high cognitive abilities. We suggest that this sex difference
in behaviour stems from the differential exposure to high intrauterine T
of male and female ASD subjects.
3.8. Functional magnetic resonance imaging, fMRI
Functional magnetic resonance imaging, fMRI, measures brain ac-
tivity by detecting changes associated with blood flow. This technique
relies on the fact that cerebral blood flow and neuronal activity are
coupled. When an area of the brain is in use, blood flow to that region
also increases. Functional imaging studies have shown that certain
brain regions show greater activity during resting states than during
cognitive tasks. This has led to the finding that these regions constitute
a network supporting a default mode of brain function. It has been
shown that in normal subjects there are sex differences in the default
mode network with regard to autism spectrum traits [53]. A study in-
volving both normal and ASD subjects showed differential neural ex-
pressions: female cases showed a shift towards the normal male con-
nectivity pattern and male cases showed a shift towards the normal
female pattern [54]. These authors suggested that this impugned Baron-
Cohen's EMB Theory, writing: “Our data support the notion that ASD is
a disorder of sexual differentiation rather than a disorder characterised
by masculinization of both genders”. Floris et al. (2018) presented data
on a large sample of normal and ASD males and females. They offered a
similar view: they suggested that there are two competing models for
ASD, the EMB Theory and the Gender Incoherence (GI) model. The
latter predicts a shift towards maleness in females and towards fema-
leness in male ASD cases. However, this seems equivocal evidence
against the EMB Theory because they also reported co-existing but
network-specific shift-towards-maleness and shift-towards-femaleness
in ASD males. A shift-towards-maleness reportedly involved the default
network, while a shift-towards-femaleness mostly occurred in the so-
matomotor (of, or pertaining to, movements of the body) network.

Moreover, Kozhemiako et al. [55] reported that males and females with
ASD tend to follow the male pattern of developmental changes in in-
terhemispheric connectivity, thus supporting Baron-Cohen's EMB
Theory. Lastly, Zeestraten et al. [56] reported frontal connectivity ab-
normality in males with autism, but not in females with autism. We
suggest that the different brain connectivities in male and female ASD
brains are programmed by their different exposures to intrauterine T
(consequent on the multifactorial threshold phenomenon as described
in Hypothesis A).
4. Summary and conclusions
The foregoing eight potential explanations may be seen to be de-
pendent on the dynamics of prenatal and postnatal T in males and fe-
males, and to be consistent with Baron-Cohen's concept of EMB. We
suggest that there is good evidence that:
1. Male and female autistic subjects differ on the average in that
they have had different exposures to the causes of autism, females
bearing higher genetic burdens for ASD, and males having a greater
exposure to an environmental cause viz. high intrauterine levels of
testosterone.
2. The high levels of intrauterine (and possibly postnatal) testos-
terone to which ASD cases have been exposed, cause.
a) a less masculinized physical habitus (including facial features) in
exposed males, and.
b) a more masculinized physical habitus in exposed females.
3. ASD genes (as opposed to intrauterine testosterone) are mainly
responsible for a low mean IQ in ASD (especially female cases).
4. Exposure to high intrauterine T increases the probability that
foetuses will be male, thus potentially explaining the high sex ratio
(proportion male) of cases of ASD.
5. The Gender Incoherence Model seems to be based on facts un-
related directly to autism. The shifts towards the other sex are argued to
be consequent on sex-different reactions to prenatal exposure to high T,
not on the pathology itself.
6. The suspected underdiagnosis of female cases is partially de-
pendent on the different proportions of environmental and genetic
causes to which male and female cases are hypothesized to have been
exposed, and the consequent ‘more normal’ behaviour of female cases.
7. The fact that so many potential explanations may plausibly be
based on the two hypotheses (A and B here) must give added credibility
to those hypotheses.
Note
For illustrative purposes, it would be useful if we could provide
normative data on pre- and post-natal testosterone levels in random
samples of male and female autistic subjects and controls. We were
unable to locate such data.
Acknowledgments
WHJ is indebted to Professor Sir David Spiegelhalter FRS
(Laboratory of Statistics, University of Cambridge) and to Professor
Robert Trivers for their generous assessments of his hypothesizing.
Declaration of competing interest
There are no known conflicts of interest associated with this pub-
lication and there has been no significant financial support for this work
that could have influenced its outcome.
References
[1] ’
[2] S. Baron-Cohen, The extreme male brain theory of autism, Trends Cogn. Sci. 6
4
Early Human Development xxx (xxxx) xxxx
(2002) 248–254.
[3] J. Martin, R.K. Walters, D. Demontis, M. Mattheisen, S.H. Lee, E. Robinson, et al., A
genetic investigation of sex bias in the prevalence of attention-deficit hyperactivity
disorder, Biol. Psychiatry 83 (2018) 1044–1053.
[4] E.B. Robinson, P. Lichtenstein, H. Anckarsater, F. Happe, A. Ronald, Examining and
interpreting the female protective effect against autistic behaviour, Proceedings of
the National Academy of Sciences USA 110 (2013) 5258–5262.
[5] S. Sumi, H. Taniai, T. Miyachi, M. Tanemura, Sibling risk of pervasive develop-
mental disorder estimated by means of an epidemiologic survey in Nagoya, Japan,
J. Hum. Genet. 51 (2006) 518–522.
[6] N. Palmer, A. Beam, D. Agniel, A. Eran, A. Manrai, C. Spettell, et al., Association of
sex with recurrence of autism spectrum disorder among siblings. J.A.M.A, Pediatrics
171 (2017) 1107–1112.
[7] S. Baron-Cohen, B. Auyeung, B. Norgaard-Pedersen, D.M. Hougaard, M. Abdallah,
L. Melgaard, et al., Elevated fetal steroidogenic activity in autism, Mol. Psychiatry
20 (2015) 369–376.
[8] Cherskov, A., Pohl, A., Allison, C., Zhang, H., Payne, R.A., Baron-Cohen, S. 2018.
Polycystic ovary syndrome and autism: a test of the prenatal sex steroid theory.
Translational Psychiatry Vol 8, Article 136
[9] M.P. Recabarren, P.P. Rojas-Garcia, R. Einspanier, V. Padmanabhan, T. Sir-
Peterman, S.E. Recabarren, Pituitary and testis responsiveness of young male sheep
exposed to testosterone excess during fetal development, Reproduction 145 (2013)
567–576.
[10] S.E. Recabarren, P.P. Rojas-Garcia, M.P. Recabarren, V.H. Alfaro, R. Smith,
V. Padmanabhan, et al., Prenatal testosterone excess reduces sperm count and
motility, Endocrinology 149 (2008) 6444–6448.
[11] S.E. Recabarren, M. Recabarren, D. Sandoval, A. Carrasco, V. Padmanabhan,
Puberty arises with testicular alteration and defective AMH expression in rams
prenatally exposed to testosterone, Domest. Anim. Endocrinol. 61 (2017) 100–107.
[12] V. Padmanabhan, M. Manikkam, S. Recabarren, D. Foster, Prenatal testosterone
excess programs reproductive and metabolic dysfunction in the female, Mol. Cell.
Endocrinol. 246 (2006) 165–174.
[13] S.S. Jonker, S. Louey, C.E. Roselli, Cardiac myosite proliferation and maturation
near term is inhibited by early gestation maternal testosterone exposure, Am. J.
Physiol. Heart Circ. Physiol. 315 (2018) H1393–H1401.
[14] Tata, B., Mimouni, N.E.H., Barbotin, A.-L., Malone, S.A., Loyens, A., Pigny, D.D.
2018. Elevated prenatal anti-Mullerian hormone reprograms the fetus and induces
polycystic ovary syndrome in adulthood. Nature Medicine 24, issue 6, pp 834-846.
[15] M.J. Brosnan, Digit ratio and faculty membership: implications for the relationship
between prenatal testosterone and academia, Br. J. Psychol. 97 (2006) 455–466.
[16] D.A. Geier, M.R. Geier, A prospective assessment of androgen levels in patients with
autistic spectrum disorders: biochemical underpinnings and suggested therapies,
Neuroendocrinol. Lett. 28 (2007) 565–573.
[17] D.A. Geier, J.K. Kern, P.G. King, L.K. Sykes, M.R. Geier, An evaluation of the role
and treatment of elevated male hormones in autism spectrum disorders, Acta
Neurobiol. Exp. 72 (2012) 1–17.
[18] H. Takagishi, T. Takahashi, T. Yamagishi, M. Shinada, K. Inukai, S. Tanida, et al.,
Salivary testosterone levels and autism spectrum quotient in adults,
Neuroendocrinol. Lett. 31 (2010) 837–841.
[19] D. Enter, P. Spinhoven, K. Roelofs, Dare to approach: single dose testosterone ad-
ministration promotes threat approach in patients with social anxiety disorder, Clin.
Psychol. Sci. 4 (2016) 1073–1078.
[20] Wu, Y., Clark, L., Zilioli, S., Eisenegger, C., Gillan, C.M., Deng, H. et al. 2018. Single
dose testosterone administration modulates emotional reactivity and counterfactual
choice in healthy males. Psychoneuroendocrinology 90, 127=133.
[21] S. Bejerot, J.M. Erikkson, S. Bonde, et al., The extreme male brain revisited: gender
coherence in adults with autism spectrum disorder, Br. J. Psychiatry 201 (2012)
116–123.
[22] L. Ruta, E. Ingudomnukul, K. Taylor, B. Chakrabarti, S. Baron-Cohen, Increased
serum androstenedione in adults with autism spectrum disorders,
Psychoneuroendocrinology 36 (2011) 1154–1163.
[23] E. Schwarz, P.C. Guest, H. Rahmoune, L. Wang, Y. Levin, E. Ingudomnukul, et al.,
Sex specific serum biomarker patterns in adults with Asperger’s syndrome, Mol.
Psychiatry 16 (2011) 1213–1220.
[24] Tan, D.W., Maybery, M.T., Clarke, M.W., Di Lorenzo, R., Evans, M.O., Mancinone,
M. et al. 2018. No relationship between autistic traits and salivary testosterone
concentrations in men from the general population. PLOS ONE Volume 13, Article
e0198779.
[25] M.S. Rohde, A.L. Georgescu, K. Vogeley, R. Fimmers, C.M. Falter-Wagner, Absence
of sex differences in mental rotation performance in autism spectrum disorder,
Autism 22 (2018) 855–865.
[26] Tovainen, T., Pannini, G., Papageorgiou, K.A., Malanchini, M., Rimfeld, K.,
Shakeshaft, N. et al. 2018. Prenatal testosterone does not explain sex differences in
spatial ability. Scientific Reports Vol 8, Article 13653.
[27] Constantinescu, M., Moore, D.S., Johnson, S.P., Hines, M. 2018. Early contribution
to infants' mental rotation abilities. Developmental Science Issue 4, Article e12613.
[28] Pintzka, C.W.S., Evensmoen, H.R., Lehn, H., Haberg, A.K. 2016. Change in spatial
cognition and brain activity after a single dose of testosterone in healthy women.
Behavioural Brain Research 298, 78-90, Part B.
[29] Gilani, S.Z., Tan, D.W., Russell-Smith, S.N. et al. 2015. Sexually dimorphic facial
features vary according to level of autistic-like traits in the general population.
Journal of Neurodevelopmental Disorders 7, Article 14.
[30] S. Ahmed, J. Sihvonen, S. Vahamaa, CEO facial masculinity and bank risk-taking,
Personal. Individ. Differ. 138 (2019) 133–139.
[31] S. Mackenzie, C. Wilkinson, Morphological and morphometric changes in the faces
of female-to-male (FtM) transsexual people, International Journal of

Transgenderism 18 (2017) 172–181.
[32] F. Probst, C. Bobst, J.S. Lobmaier, Testosterone-to-oestradiol ratio is associated with
female facial attractiveness, Q. J. Exp. Psychol. 69 (2016) 89–99.
[33] R.L. Trivers, D.E. Willard, Natural selection of parental ability to vary the sex ratio
of offspring, Science 179 (1973) 90–92.
[34] W.H. James, Evolution and the variation of mammalian sex ratios at birth: reflec-
tions on Trivers and Willard (1973), J. Theor. Biol. 334 (2013) 141–148.
[35] S. Baron-Cohen, M.V. Lombardo, B. Auyeung, E. Ashwin, B. Chakrabarti,
R. Knickmeyer, Why Are Autism Spectrum Conditions more Prevalent in Males?
PLOS BIOLOGY 9, issue 6; article e1001081, (2011).
[36] W.H. James, The evidence that mammalian sex ratios at birth are partially con-
trolled by parental hormone levels at the time of conception, J. Theor. Biol. 180
(1996) 271–286.
[37] W.H. James, Further evidence that mammalian sex ratios at birth are partially
controlled by parental hormone levels around the time of conception, Hum. Reprod.
19 (2004) 1250–1256.
[38] W.H. James, Evidence that mammalian sex ratios at birth are partially controlled by
parental hormone levels around the time of conception, J. Endocrinol. 198 (2008)
3–15.
[39] D. Spiegelhalter, Sex by Numbers, Profile Books in association with the Wellcome
Collection, London, 2015.
[40] C. Harrop, D. Jones, S. Zheng, S. Nowell, B.A. Boyd, N. Sasson, Circumscribed in-
terests and attention in autism: the role of biological sex, J. Autism Dev. Disord. 48
(2018) 3449–3459.
[41] C. Harrop, D. Jones, S. Zheng, S.W. Nowell, B.A. Boyd, N. Sasson, Sex differences in
social attention in autism spectrum disorder, Autism Res. 11 (2018) 1264–1275.
[42] Holt, R., Upadhyay, J., Smith, P., Allison, C., Baron-Cohen, S. 2018. The Cambridge
Sympathy Test: self-reported sympathy and distress in autism. PLOS ONE Volume
13 issue 7: Article e0198273.
[43] Baron-Cohen, S., Bowen, D.C., Holt, R.J., Allison, C., Auyeung, B., Lombardo, M.V.
et al. 2015b The ‘Reading the Mind in the Eyes’ test: complete absence of typical sex
differences in ~ 400 men and women with autism. PLOS ONE Vol 10, Article
e01365al 21.
[44] P.A. Bos, D. Hofman, E.J. Hermans, E.R. Montoya, S. Baron-Cohen, J. van Honk,
Testosterone reduces functional connectivity during the ‘Reading the mind in the
eyes’ test, Psychoneuroendocrinology 68 (2016) 194–201.
[45] J.T. Manning, Digit Ratio: A Pointer to Fertility, Behavior and Health, Rutgers
University Press, London, 2002.
[46] Masuya, Y., Okamoto, Y., Inohara, K., Matsumura, Y., Fujioka, Y., Wada, Y. et al.
2015. Sex-different abnormalities in the right second to fourth digit ratio in
⁎
Corresponding author.
Early Human Development xxx (xxxx) xxxx
[47] L.A. Schieve, L. Tian, N. Dowling, L. Croen, J. Hoover-Fong, Associations between
the 2nd to 4th digit ratio and autism spectrum disorder in population-based samples
of boys and girls: findings from the study to explore early development, J. Autism
Dev. Disord. 48 (2018) 2379–2395.
[48] Barona, M., Kothari, R., Skuse, D., Micali, N. 2015. Social communication and
emotional difficulties and second-to-fourth digit ratio in a large community-based
sample. Molecular autism 6, article 68.
[49] D.M. Werling, D.H. Geschwind, Sex differences in autism spectrum disorders, Curr.
Opin. Neurol. 26 (2013) 146–153.
[50] E. Rubenstein, L.D. Wiggins, L.-L. Lee, A review of the differences in developmental,
psychiatric and medical endophenotypes between males and females with autism
spectrum disorder, J. Dev. Phys. Disabil. 27 (2015) 119–139.
[51] L. Antezana, R.S. Factor, E.E. Condy, M.V. Strege, A. Scarpa, Gender differences in
restricted and repetitive behaviors and interests in youth with autism, Autism Res.
12 (2) (2019) 274–283.
[52] C.S. Allely, Understanding and recognising the female phenotype of autism spec-
trum disorder and the ‘camouflage’ hypothesis: a systematic PRISMA review.
Advances in autism 5, Special Issue S1 (2019) 14–37.
[53] Jung, M., Mody, M., Saito, D.N., Tomoda, A., Okazawa, H. 2015. Sex differences in
the default mode network with regard to autism spectrum traits: a resting-state
fMRI study. PLOS ONE Vol 10, Issue 11, Article e0143126.
[54] K. Alaerts, S.P. Swinnen, N. Wenderoth, Sex differences in autism: a resting-state
fMRI investigation by functional brain connectivity in males and females, Cognitive
and Affective Neuroscience 11 (2016) 1002–1016.
[55] N. Kozhemiako, V. Vakorin, A. Nunes, G. Iarocci, U. Ribary, Extreme male devel-
opmental trajectories of homotopic brain connectivity in autism, Hum. Brain Mapp.
40 (2019) 987–1000.
[56] E.A. Zeestraten, M.C. Gudbrandsen, E. Daly, M. Thiebaut de Schotten, M. Catani,
Sex differences in frontal lobe connectivity in adults with autism spectrum condi-
tions, Translational Psychiatry 7 (2017) e1090.
⁎
William H. James, Victor Grech ,
a
Galton Laboratory, Department of Genetics, Evolution and Environment,
University College London, United Kingdom of Great Britain and Northern
Ireland
b
Department of Paediatrics, Medical School, Mater Dei Hospital, Malta
E-mail addresses: james@ucl.ac.uk (W.H. James),
victor.e.grech@gov.mt (V. Grech).