Search

Previous studies have shown that machine-learning (ML) algorithms can “predict” sex based on brain anatomical/ functional features. The high classification accuracy achieved by ML algorithms is often interpreted as revealing large differences between the brains of males and females and as confirming the existence of “male/female brains”. However, classification and estimation are different concepts, and using classification metrics as surrogate estimates of between-group differences may result in major statistical and interpretative distortions. The present study avoids these distortions and provides a novel and detailed assessment of multivariate sex differences in gray matter volume (GMVOL) that does not rely on classification metrics. Moreover, appropriate regression methods were used to identify the brain areas that contribute the most to these multivariate differences, and clustering techniques and analyses of similarities (ANOSIM) were employed to empirically assess whether they assemble into two sex-typical profiles. Results revealed that multivariate sex differences in GMVOL: (1) are “large” if not adjusted for total intracranial volume (TIV) variation, but “small” when controlling for this variable; (2) differ in size between individuals and also depends on the ML algorithm used for their calculation (3) do not stem from two sex-typical profiles, and so describing them in terms of “male/female brains” is misleading.

Abstract Background Studies assessing volumetric sex differences have provided contradictory results. Total intracranial volume (TIV) is a major confounding factor when estimating local volumes of interest (VOIs). We investigated how the number, size, and direction of sex differences in gray matter volume (GMv) vary depending on how TIV variation is statistically handled. Methods Sex differences in the GMv of 116 VOIs were assessed in 356 participants (171 females) without correcting for TIV variation or after adjusting the data with 5 different methods (VBM8 non-linear-only modulation, proportions, power-corrected-proportions, covariation, and the residuals method). The outcomes obtained with these procedures were compared to each other and to those obtained in three criterial subsamples, one comparing female-male pairs matched on their TIV and two others comparing groups of either females or males with large/small TIVs. Linear regression was used to quantify TIV effects on raw GMv and the efficacy of each method in controlling for them. Results Males had larger raw GMv than females in all brain areas, but these differences were driven by direct TIV-VOIs relationships and more closely resembled the differences observed between individuals with large/small TIVs of sex-specific subsamples than the sex differences observed in the TIV-matched subsample. All TIV-adjustment methods reduced the number of sex differences but their results were very different. The VBM8- and the proportions-adjustment methods inverted TIV-VOIs relationships and resulted in larger adjusted volumes in females, promoting sex differences largely attributable to TIV variation and very distinct from those observed in the TIV-matched subsample. The other three methods provided results unrelated to TIV and very similar to those of the TIV-matched subsample. In these datasets, sex differences were bidirectional and achieved satisfactory replication rates in 19 VOIs, but they were “small” (d

It is a common belief that males have superior visuospatial abilities and that differences in this and other cognitive domains (e.g., math) contribute to the reduced interest and low representation of girls and women in STEM education and professions. However, previous studies show that gender-related implicit associations and explicit beliefs, as well as situational variables, might affect cognitive performance in those gender-stereotyped domains and produce between-gender spurious differences. Therefore, the present study aimed to provide information on when, how and who might be affected by the situational reactivation of stereotypic gender-science beliefs/associations while performing a 3D mental rotation chronometric task (3DMRT). More specifically, we assessed the explicit beliefs and implicit associations (by the Implicit Association Test) held by female and male students of humanities and STEM majors and compared their performance in a 3DMRT after receiving stereotype- congruent, incongruent and nullifying experimental instructions. Our results show that implicit stereotypic gender-science associations correlate with 3DMRT performance in both females and males, but that inter-gender differences emerge only under stereotype-reactivating conditions. We also found that changes in self-confidence mediate these instructions’ effects and that academic specialization moderates them, hence promoting 3DMRT performance differences between male and female humanities, but not STEM, students. Taken together, these observations suggest that the common statement “males have superior mental rotation abilities” simplifies a much more complex reality and might promote stereotypes which, in turn, might induce artefactual performance differences between females and males in such tasks.