Fetal behavior during MRI changes with age and relates to network dynamics

Abstract

Fetal motor behavior is an important clinical indicator of healthy development. However, our understanding of associations between fetal behavior and fetal brain development is limited. To fill this gap, this study introduced an approach to automatically and objectively classify long durations of fetal movement from a continuous four-dimensional functional magnetic resonance imaging (fMRI) data set, and paired behavior features with brain activity indicated by the fMRI time series. Twelve-minute fMRI scans were conducted in 120 normal fetuses. Postnatal motor function was evaluated at 7 and 36 months age. Fetal motor behavior was quantified by calculating the frame-wise displacement (FD) of fetal brains extracted by a deep-learning model along the whole time series. Analyzing only low motion data, we characterized the recurring coactivation patterns (CAPs) of the supplementary motor area (SMA). Results showed reduced motor activity with advancing gestational age (GA), likely due in part to loss of space (r = -.51, p < .001). Evaluation of individual variation in motor movement revealed a negative association between movement and the occurrence of coactivations within the left parietotemporal network, controlling for age and sex (p = .003). Further, we found that the occurrence of coactivations between the SMA to posterior brain regions, including visual cortex, was prospectively associated with postnatal motor function at 7 months (r = .43, p = .03). This is the first study to pair fetal movement and fMRI, highlighting potential for comparisons of fetal behavior and neural network development to enhance our understanding of fetal brain organization.

Keywords: coactivation patterns; deep learning; fetal fMRI; fetal motor behavior; motor cortex; motor development; network dynamics.

FIGURE 1

An example (GA = 32 weeks) of fetal motor behavior captured by fMRI. Top panels depict three fetal brain volumes drawn from a period of extreme repositioning. The mask used to segment the volume at each time is shown in turquoise. The brains approximate coronal, sagittal, and axial views, left to right. Plotted, below, is the frame‐wise displacement (FD) over the time series. Circles in the lower plot indicate the time point corresponding to the brain figures shown in top panels. fMRI, functional magnetic resonance imaging; GA, gestational age.

FIGURE 2

Workflow of the full analysis. For each subject, a 12‐min scan, or two 6‐min fMRI scans, was auto‐masked by the deep learning model. The “MCFLIRT” realignment tool was then applied to the resulting 4D extracted brain series, to generate motion parameters for the full‐time series. Fetal motor behavior measures of mean frame‐wise displacement (FD), max FD, and maximum displacement were derived from the output of this realignment step. FD measures reflect the magnitude of brain location changes from volume to volume, while the maximum displacement measures the maximum movement over the entire 12 min. All the above steps are framed within yellow boxes. In parallel, the CAPs analysis was conducted only on the low‐motion data (indicated by light green boxes). fMRI volumes with significant head motion were manually identified, using FSL image viewer, and excluded. Remaining segments of relatively low‐motion fMRI data are henceforth referred to as “the low‐motion segments.” For each 4D low‐motion segment, manual masking, reorientation, motion correction, and normalization to a 32‐week fetal brain template (Serag et al., 2012) were performed. Preprocessed segments were then concatenated to form a single series and realignment was applied to correct potential misalignment between segments. This was followed by ICA‐denoising and smoothing. CAP analysis was conducted on the first 100 processed volumes, which assures consistent quantity of data across subjects. CAP, coactivation pattern; fMRI, functional magnetic resonance imaging; ICA, Independent Component Analysis.

FIGURE 3

Motor behavior changes across gestational age. Fetal mean (left) and max (middle) frame‐wise displacement (FD), and the maximum brain displacement (right) across all volumes in the fMRI scan are all negatively correlated with gestational age. fMRI, functional magnetic resonance imaging.

FIGURE 4

SMA‐CAPs. The six SMA CAPs (z values, thresholded at |z| > 0.5) are displayed on the left, with pie charts indicating the amount of time each pattern coactivated with the left SMA, right SMA, or bilateral SMA seeds. CAPS were calculated on the first 100 low‐motion fMRI volumes. More frequent CAP 1 occurrence was marginally correlated with lower fetal gestational age (top right) and more frequent CAP 5 occurrence was significantly correlated with lower fetal maximum brain displacement across the entire fMRI scan. The frequency of CAP1 and CAP5 occurrence is indicated on the top X axis of both scatterplots and the frequency of GA and max FD values is indicated on the right Y axis of each plot, respectively. There were no other significant associations between CAP occurrence, fetal gestational age, and fetal motor behavior. CAP, coactivation pattern; FD, frame‐wise displacement; fMRI, functional magnetic resonance imaging; GA, gestational age; SMA, supplementary motor area.

FIGURE 5

Prediction effect of CAP3 occurrence on postnatal motor function (Bayley Composite score) at 7 months. Fetuses who spent more time in CAP 3, representing coactivation of right occipital regions with the SMA seeds, had more mature motor abilities at 7 months after birth. The distribution of Bayley motor composite scores is displayed on the right Y axis and the distribution of CAP 3 occurrence is displayed on the top X axis. CAP, coactivation pattern; SMA, supplementary motor area.