Exploring moment-to-moment brain signal variability before and after pregnancy: preliminary results

2030 

Abstract Submission 

Sara Halmans¹, Damiaan Denys¹, Kristoffer Månsson², Elseline Hoekzema¹

¹Amsterdam UMC, Amsterdam, The Netherlands, ²Centre for Psychiatry Research, Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden

Sara Halmans  
Amsterdam UMC
Amsterdam, The Netherlands

Damiaan Denys  
Amsterdam UMC
Amsterdam, The Netherlands

Kristoffer Månsson  
Centre for Psychiatry Research, Department of Clinical Neuroscience, Karolinska Institutet
Stockholm, Sweden

Elseline Hoekzema  
Amsterdam UMC
Amsterdam, The Netherlands

Pregnancy represents an important endocrine event and life transition in women that is associated with pronounced structural and functional changes in the brain.^1-2 Increased prevalence of psychiatric symptoms during the perinatal period emphasize the importance of understanding the neural underpinnings of this life transition.³ Despite the focus of functional magnetic resonance imaging (fMRI) research on the interpretation of mean blood-oxygen-level-dependent (BOLD) signals, several studies have shown that the within-individual variability of brain responses represents a key component of neural processes.^4-6 Research indicates that fMRI signal variability fluctuates over the lifespan, and neural variability has also been linked to both psychiatric symptoms and cognitive performance.^6-8 In this study, we aim to examine if becoming a mother is associated with alterations in moment-to-moment variability in resting-state neural response. Here, our objective was to compare the nulliparous control group and women that want to become pregnant for the first time at baseline.

Herein, 110 women took part in a prospective cohort study, involving longitudinal data of first- (n=40) and second-time mothers (n=30), who were scanned 1) before conception, 2) at an early stage and 3) at a later postpartum stage (see Figure 1). A nulliparous control group (n=40) was scanned twice at a similar time interval as the first-time mothers (i.e., ~ 12 months). Resting-state fMRI was performed at all sessions. fMRI preprocessing included a manually denoising procedure.⁹ Briefly, independent component analysis (ICA), was used to separated brain signals into components, and manual decisions were made whether each component were neural activity or not (e.g., movement, noise). The noise components were subsequently regressed out of the signal. The standard deviation of the BOLD signal (SD_BOLD) was calculated over the entire time series of a single scan per voxel (see Figure 2), allowing us to investigate whole-brain voxel-wise signal variability. To delve into possible dynamic aspects within the resting-state, the five-minute resting-state scan was also segmented into five one-minute intervals for inter-group comparisons. Case-control comparisons on resting-state SD_BOLD data were performed using independent t-tests in Statistical Parametric Mapping (SPM12; implemented in Matlab R2022a). Alpha was set at p < 0.001 (whole-brain uncorrected).

By use of SD_BOLD resting-state data at the PRE time point, we compared the nulliparous control group (CTRL) and the group of women wanting to become pregnant (PREG) using two contrasts: 1) CTRL > PREG, 2) PREG < CTRL. No significant clusters are reported. Comparisons of the shorter one-minute segments of the resting-state data also showed no significance. Further analyses of the later time points (i.e., postpartum), to investigate longitudinal changes, in both resting-state and task-based fMRI data are still pending.

The absence of significant differences in both directions indicates that the control group and the pregnancy group do not differ in neural variability during resting-state fMRI. This was expected, given that the two groups at the PRE time point differ only in the current desire to have a child in the pregnancy group. It is known that becoming a mother renders strong changes in neural grey matter structure and neural activity.² Therefore, future analyses of this longitudinal dataset may further elucidate whether pregnancy is associated with changes in neural variability. Additionally, previous research suggests that task-based variability may outperform resting-state variability.⁶ Including task-based data may therefore also provide insights into whether differences between task-based and resting-state variability are also evident in the context of pregnancy.

Lifespan Development Other ²

Activation (eg. BOLD task-fMRI)

Motion Correction and Preprocessing

Task-Independent and Resting-State Analysis ¹

BOLD fMRI

FUNCTIONAL MRI

Other - Pregnancy; Signal variability; resting-state