Spets and Slotnick present a meta-analysis on long-term memory and sex differences. Overall findings indicate greater brain activation in men than women during memory performance. Merits of the activation likelihood estimation meta-analysis include considering study parameters and equating performance to enhance interpretability of activation differences. Variables and concepts relevant to memory and sex differences research also are discussed. As memory is essential for survival, characterizing neurobiological profiles, parsing sex and gender, will help broaden the field of long-term memory and sex differences research.
Keywords: sex, gender, long-term memory, learning, fMRISpets and Slotnick (Spets & Slotnick, 2020) present a timely meta-analysis on sex differences in brain activity with a focus on long-term memory. The long-standing question of whether there are sex differences in brain function has long existed, on cellular, molecular and behavioral levels. Here we highlight the main findings and significance of this meta-analysis, and suggest relevant variables and concepts that should be considered to even further broaden the field of long-term memory and sex differences research.
The authors present results from eight studies examining differences in brain activation between women and men, during performance of long-term memory tasks:, memory for words, faces and shapes, autobiographical memory (visual/verbal cues) and memory for abstract shapes. Using functional magnetic resonance imaging (fMRI), findings indicate greater brain activation in men than women during memory performance, in left lateral prefrontal cortex and on the right, middle occipital gyrus, fusiform gyrus, parahippocampal cortex and cerebellum. A strength of this meta-analysis was inclusion of methodological details: sample size, age, stimulus type, matched-performance, statistical testing, and fMRI contrasts of interest (selection of comparisons that isolate the cognitive process from a control condition). The authors conclude that across studies, the wide range of experimental parameters employed did not have a significant impact on the reported sex differences.
Spets and Slotnick (Spets & Slotnick, 2020) also highlight limitations of collapsing across sex due in part to generally small sample sizes in fMRI studies (typically ranging from n=10–26 ). Thus, there is a widespread practice of collapsing across sex, likely due to an even smaller number of women included in imaging investigations. Another important aspect highlighted in the meta-analysis is the merit of selecting tasks in which performance can be equated across women and men. Performance matching is relatively rare in memory studies, despite strong evidence that men typically outperform women on spatial memory tasks, whereas women typically outperform men on verbal memory tasks. Nonetheless, equating performance permits a unique opportunity to parse brain activation from performance, which is particularly important given that memory is a broad construct that involves contributions from multiple brain regions.
Challenges common to fMRI research include interpretation of observed brain activation differences. As raised in the meta-analysis, greater brain activation in men reported in the absence of task performance differences is suggested to reflect less neural efficiency in men, thereby eliciting increased activation to perform at levels equivalent to women. As acknowledged in the meta-analysis, studies of long-term memory should consider unique contributions of encoding to long-term memory (retrieval) when interpreting study findings. With the rapid advance of sophisticated statistical imaging methods, the neuroscience field has generally moved away from a systems biology approach that probes brain function using a region of interest approach. Such innovative approaches now permit investigation of integrated brain function on a network level, supporting that memory no longer solely lives in the hippocampus, but that hippocampus has synchronous (and asynchronous) connections to a constellation of other brain regions, that together produce the phenomenon of memory.
This meta-analysis also impressively describes what is known about neurobiological origins of memory in men versus women, and that existing data are limited. This highlights the critical need to collect data from equal numbers of women and men (Mazure & Jones, 2015; Woitowich et al., 2020). Over two decades ago, inclusion of women (and minority groups) was mandated in NIH-funded clinical research, to ensure that research findings are generalizable to the entire population. Yet, current studies continue to underrecruit and enroll women, which is necessary to more accurately characterize sex differences in the neurobiology underlying long-term memory. The authors also acknowledge that within existing data, there has been a lack of focus on critical factors such as sex-hormones (e.g., estrogen, progesterone, testosterone) and effects of menstrual cycle phase on long-term memory. These factors not only influence brain activity, likely mediating memory performance (Fernández et al., 2003). Using a multivariate machine learning approach, it was concluded that there are complex interactions between hormonal, genetic, and epigenetic factors that influence sex differences in brain (Kurth et al., 2020). Thus, future studies are warranted that integrate memory performance relative to circulating hormone levels, with attention paid to menstrual cycle phase in naturally cycling women, as well as regulated hormone levels in women taking oral contraceptives.
Beyond the results of this meta-analysis and the limitations raised, it is important to consider the impact of the binary categorization of sex, as sex assigned at birth (i.e., natal sex, female/male) does not always align with gender identity/felt gender. Complex differences between women and men can therefore be subject to oversimplification by classifying brains as either “female” or “male” (Kurth et al., 2020). While the differential contributions of biological factors (natal sex) and cultural factors (gender identity) are relatively unclear, stratification into female or male groups likely obscures interpretation of detected differences. Individual differences may therefore be better differentiated by incorporating an approach that examines brain differences on a continuum of femaleness to maleness. Thus “sex” should be reserved for studies reporting biological differences, whereas “gender” should be used when reporting differences that are culturally based, such as gender identity. In most studies reporting sex differences, participants are assumed to be cisgender, identity is aligned with sex assigned at birth. New directions in clinical research on long-term memory should be sensitive to these important distinctions of sex versus gender. Trajectories of change in brain, hormones and behavior across developmental stages, and use of pubertal blockers and/or cross-sex hormones that help affirm gender identity, also should be taken into account.
Memory is essential for survival, and serves to, among many roles, promote adaptation to environmental conditions. Further, learning and memory are vulnerable to elements such as aging and psychiatric illness, which can have differential impacts in women compared to men. Characterizing neurobiological profiles with consideration of parsing sex from gender, therefore, is crucial. Neurobiology underlying memory differs between women and men, as eloquently reviewed by Spets and Slotnick (Spets & Slotnick, 2020). This meta-analysis brings to the forefront, the critical need for this line of research, which may in turn, inform the development of practices that serve to promote maintenance of a healthy memory, specifically tailored to women versus men.