By guest blogger Gina Rippon
In case you hadn’t noticed, there is an ongoing debate about the existence of differences between women’s and men’s brains, and the extent to which these might be linked to biological or to cultural factors. In this debate, a real game-changer of a study would involve the identification of clear-cut sex differences in foetal brains: that is, in brains that have not yet been exposed to all the different expectations and experiences that the world might offer. A recent open-access study published in Developmental Cognitive Neuroscience by Muriah Wheelock at the University of Washington and her colleagues, including senior researcher Moriah Thomason at New York University School of Medicine, claims to have done just that, hailed by the researchers themselves as “confirmation that sexual dimorphism in functional brain systems emerges during human gestation” and in various ways by the popular press as, for example, The Times of London’s headline: “Proof at last: women and men are born to be different”.
Does this study live up to the claims made by its authors and, more excitedly, those passing the message on? I think not.
The study is certainly heroic in methodological terms, comprising a challenging recruitment and scanning strategy to capture “resting state” data (brain activity during rest rather than during an active task) from over 100 prenatal participants. As well as not having direct access to their participants, the experimenters had to devise extraordinarily complex ways of controlling for their movement – a challenge for any brain scan study, but particularly those hoping to measure connectivity. This involved the researchers identifying the “stillest” periods in each 12-24 minute foetal resting state scan, and then performing subsequent re-alignment and readjustments (or, effectively, cutting and pasting of each of lowest movement epochs together for analysis). We are told that 40 per cent of the acquired data had to be discarded in this process (something that should be borne in mind in assessing the validity and reliability of the data).
Having identified 197 regions of interest (ROIs) in these foetal brain images, the researchers looked at how the activity in each region was correlated with the activity in all the others (a stronger correlation is taken as a sign of greater connectivity), resulting in an analysis of 19,306 connections or pairings between regions. The researchers then used an innovative method to identify significant clusterings in the data, to see if there were any networks or “hubs” of connectivity. It should be noted that the correlations were calculated separately for the males and females and that an uncorrected p-threshold of 0.05 was applied to the resulting correlations, both rather questionable statistical practices when working with data sets such as these.
Early versions of the overall methodological approach used here have previously been reported by the same research team , providing fascinating insights into the formation of neural connectivity systems in utero and how these change with gestational age. The basis for the current paper is a much bigger data set than they used previously, which could have continued to develop these insights. Sadly, and rather inexplicably, their focus here is on differences in connectivity between female and male foetuses. This is unfortunate, not because identifying prenatal sex differences in the brain is not a crucial process in this debate, but because this data set just does not lend itself to this kind of “hunt-the-difference” agenda.
One problem is that, of the 118 participants whose data survived all of the necessary exclusion criteria, 48 were female and 70 were male. This imbalance, together with the skewed distribution of several other demographic factors, such as gestational age when the scans took place, meant that the researchers had to use what are known as “nonparametric statistics” for many of their comparisons. This form of analysis, as well as lacking power, precludes the kind of multivariate analysis which would have allowed the researchers to consider multiple factors at once. For example, there was no way of simultaneously testing the relative contributions of differences in gestational age and differences in foetal sex to different types of network pairs (connections between different brain regions).
A related problem is the wide age range of the participants: from 25 weeks to 39 gestational weeks. That 14 week period covers a time of dramatic cortical development, especially the formation of cortical networks – a 25 week-old brain will be very different from a 39 week-old one. Although the researchers treated gestational age as a continuous variable, there is a query as to the comparability of the male and female groups with respect to this measure. In the age window 36-40 weeks (very near full-term), there were 35 males compared with only 17 females. Even with appropriate statistical controls, this could well have skewed the findings.
Related to this, there are concerns about the “normalisation” process carried out on each scan. This allows group analysis of images from different participants by “warping” them to a standard template so that locations on each image “line up” for comparison. What to select for a standard template is a tricky question for relatively comparable adult scans; here the researchers selected a 32-week foetal brain template, which could well have been inappropriate for any of the individual scans they were processing.
False impressions:
It is also important to note what was NOT reported in this study. The researchers’ focus was almost entirely on the sex differences they found, with little comment on evidence of the sex similarities. In one of their figures, for example, the researchers illustrate 136 possible sex by connectivity comparisons, of which only 3 showed a statistically significant difference between the sexes. And remember that, overall, there were connections between over 19,000 possible regions to consider.
The researchers use the term “sexual dimorphism” or “dimorphism” more than half a dozen times, including in that powerful statement from their abstract that I quoted at the start of my blog post. But their data do NOT demonstrate a dimorphic state – as well as the much greater incidence of no differences between the groups, we have no way of assessing just how different (or not) the connectivity pathways were. As so many of the sex-linked comparisons were based merely on contrasting the number of significant correlations between brain regions for males and females – for instance, one of the figures in the paper shows significant connectivity between 5 pairs of brain networks in female foetuses, as opposed to significant connectivity between 2 different brain networks in males – we have no idea of the effect size (the practical meaning) of any differences, nor the extent of overlap between the two groups.
Any studies focussing on sex differences in the brain have the potential to help explain gender gaps, not only in behaviour, personality and preferences, but also in differences in the occurrence of physical and mental health problems. Studies such as this could be invaluable for understanding the typical and atypical developmental trajectories which may lead to such gaps. But, equally, studies such as this can contribute to stereotypical essentialist beliefs that there are innate, fixed and inevitable sex differences in the brain (and therefore in behaviour and preferences). This can have significant downstream consequences for how society comes to understand, explain and even sustain these very gender gaps.
The data in this study offer an exciting and innovative look at the development of functional brain networks in the human brain, using complex and revelatory analyses. They do not offer confirmation of sexual dimorphism in human brain networks in utero and the authors of the study should not, I believe, be making this claim.
—Sex differences in functional connectivity during fetal brain development
Post written by Gina Rippon (@ginarippon1) for the BPS Research Digest. Gina is Emeritus Professor of Cognitive Neuroimaging at the Aston Brain Centre, Aston University, Birmingham. She is a past-President of the British Association of Cognitive Neuroscience and, in 2015, was awarded an Honorary Fellowship of the British Science Association. Her research involves state-of-the-art brain imaging techniques to investigate developmental disorders such as dyslexia and autism. She also investigates the use of neuroscience techniques to explore social processes such as gender stereotyping and stereotype threat. In her new book The Gendered Brain (Bodley Head), she challenges the idea that there are two sorts of “hardwired brains”, male and female, and offers a 21st century model for better understanding of how brains get to be different.
View more here.
Credit- BPS Research Digest. Published by- Dr. Sabiha : www.drsabiha.blogspot.com
