Timing of peripubertal steroid exposure predicts visuospatial cognition in men: Evidence from three samples

Abstract

Experiments in male rodents demonstrate that sensitivity to the organizational effects of steroid hormones decreases across the pubertal window, with earlier androgen exposure leading to greater masculinization of the brain and behavior. Similarly, some research suggests the timing of peripubertal exposure to sex steroids influences aspects of human psychology, including visuospatial cognition. However, prior studies have been limited by small samples and/or imprecise measures of pubertal timing. We conducted 4 studies to clarify whether the timing of peripubertal hormone exposure predicts performance on male-typed tests of spatial cognition in adulthood. In Studies 1 (n = 1095) and 2 (n = 173), we investigated associations between recalled pubertal age and spatial cognition in typically developing men, controlling for current testosterone levels in Study 2. In Study 3 (n = 51), we examined the relationship between spatial performance and the age at which peripubertal hormone replacement therapy was initiated in a sample of men with Isolated GnRH Deficiency. Across Studies 1-3, effect size estimates for the relationship between spatial performance and pubertal timing ranged from. -0.04 and -0.27, and spatial performance was unrelated to salivary testosterone in Study 2. In Study 4, we conducted two meta-analyses of Studies 1-3 and four previously published studies. The first meta-analysis was conducted on correlations between spatial performance and measures of the absolute age of pubertal timing, and the second replaced those correlations with correlations between spatial performance and measures of relative pubertal timing where available. Point estimates for correlations between pubertal timing and spatial cognition were -0.15 and -0.12 (both p < 0.001) in the first and second meta-analyses, respectively. These associations were robust to the exclusion of any individual study. Our results suggest that, for some aspects of neural development, sensitivity to gonadal hormones declines across puberty, with earlier pubertal hormone exposure predicting greater sex-typicality in psychological phenotypes in adulthood. These results shed light on the processes of behavioral and brain organization and have implications for the treatment of IGD and other conditions wherein pubertal timing is pharmacologically manipulated.

Keywords: Hormones; Isolated GnRH Deficiency; Organizational effects; Pubertal timing; Visuospatial abilities.

Fig. 1.

Dotted lines reflect endogenous hormone production across development, while the shaded box represents sensitivity to hormone-driven organizational effects. It is possible that sensitivity to hormone-driven brain organization remains constant across perinatal and peripubertal development, the constant sensitivity hypothesis (CSH; panel A); alternatively, sensitivity to hormone-driven brain organization may decrease across this window, the decreasing sensitivity hypothesis (DSH; panel B). If so, earlier puberty should be associated with greater sensitivity to hormone-driven brain organization, and greater sex typicality in adulthood. Figure adapted from Schulz and Sisk (2016).

Fig. 2.

Relationships between age of exposure to peripubertal androgens and performance on spatial cognitive tasks showing a male advantage. (A) Study 1: WAIS-R PIQ regressed on self-reported age of puberty in typically developing men (residuals after conditioning on age and gender preferences, current age, and race). (B) Study 2: MRT regressed on self-reported absolute pubertal timing in typically developing men (residuals after conditioning on sexual orientation, current age, race, and testosterone). (C) Study 3: MRT regressed on age of HRT initiation in men with IGD (residuals after conditioning on sexual orientation, current age, and race).

Fig. 3.

Results of meta-analyses on the correlation between absolute (overall r = −0.15; panel A) and relative (overall r = −0.12; panel C) pubertal timing and spatial rotation, and of omit-one-study analyses (−0.17 ≤ r ≤ −0.11; panels B and D).