[Back]

A. Taymourtash, E. Schwartz, K. Nenning, R. Licandro, D. Sobotka, M. Diogo, P. Golland, E. Grant, G. Kasprian, D. Prayer, G. Langs:
"Functional thalamocortical connectivity development revealed by in-utero resting state fMRI";
Poster: In Utero-MRI Workshop, Oxford, United Kingdom; 2020-01-07 - 2020-01-10.

Introduction: The thalamus is regarded as a key hub to receive peripheral sensory information from midbrain and send to the cerebral cortex. It has been suggested that the normal development of thalamocortical connections during the mid- and late gestational periods is critical for the functional specification of the cortex and deviation from this process might be associated with different neurodevelopmental disorders after birth such as autism, schizophrenia, and bipolar disorder. Until recently, what is known about this comes from histological and magnetic resonance imaging on post mortem foetuses or the application of resting-state fMRI to preterm infants. However, these identified patterns cannot ideally describe normal brain development since the underlying reasons for preterm delivery or environmental factors cannot be segregated from neurological process in these cases. This study presents, to our knowledge, the first in-vivo analysis of the age-dependent development of thalamocortical connectivity in the linear and non-linear frameworks. In addition, we explored the dynamic relation between structure and function of thalamus by forming functionally specialized thalamic segments and investigating their alteration during pre-, post-, and expansion periods.
Methods: The low signal-to-noise ratio of in-utero BOLD fMRI and many motion-related artifacts due to uncontrollable fetal movement and maternal respiration necessitate customized processing strategy that mitigates the effect of such issues on data. Thus, preprocessing pipeline included the following steps: correction for distortions induced by magnetic field inhomogeneity, slice timing correction, motion correction, de-spiking, temporal high-pass filtering (0.01 Hz) of the demeaned and detrended BOLD time-series, rejection of outlier volumes1, and linear regression of the nuisance regressors derived from CFS and white matter (aCompCor)2. Motion correction as a significant step in the preprocessing pipeline was performed using a recently proposed method for fetal brain rs-fMRI3,4 in which all slice stacks each acquired at the same time point are registered to a super-resolution reference volume in a two-steps slice-to-volume registration/reconstruction cycle. Individual functional connectivity analysis was performed in the native functional space. For this, cortical regions of interest (ROIs) were first obtained using an automatic atlas-based segmentation of T2 scans of the same subject acquired during the same scan session as the fMRI volumes, using a publicly available atlas of fetal brain anatomy5. The resulting parcellation consists of 98 ROIs and was mapped to the motion corrected fMRI space using a rigid transformation computed between each individual structural T2 scan and the first volume of fMRI data. Partial correlations were then computed between the time course of each parcel, calculated by averaging over its voxels´ time course, and the time course of each thalamic voxel. A winner-takes-all strategy was adopted to assign cortical labels to each thalamic voxel based on the highest partial correlation. The variation of each linear/non-linear parameters was assessed using uncertainty calculated during permutation test procedure.
Results: This study includes 24 in-utero BOLD fMRI sequences obtained from fetuses between 19 and 40 weeks of gestation (mean ± SD: 28.4 ± 5.4). None of the cases showed any neurological pathologies. Pregnant women were scanned on a 1.5T clinical scanner (Philips Medical Systems, Best, Netherlands) using single-shot echo-planar imaging (EPI), and a sensitivity encoding (SENSE) cardiac coil with five elements (three posterior, two anterior) wrapped around the mother´s abdomen. Image matrix size was 144×144, with 3×3mm2 in-plane resolution, 3mm slice thickness, a TR/TE of 1000/50 ms, and a flip angle of 90. Each scan contains 96 volumetric images obtained in an interleaved slice order to minimize cross-talk between adjacent slices. From this cohort, four subjects were excluded due to excessive image artifacts, and the average percentage of volumes kept after preprocessing and rejection of outlier volumes was 84.58%. Seven cortical labels including occipital, parietal, central, cingulate, insula, frontal, and temporal were extracted for each subject. The non-linear sigmoid function described the development of functional thalamocortical connectivity significantly better than linear function. Its results confirmed the presence of timing patterns in regional thalamocortical connections with the peak of growth rate around the transition from second to the third trimester. In addition, thalamus parcellation topography showed bilateral symmetry in their cortical labelling for all three expansion periods.
Conclusion: These results suggest that resting-state fMRI has the potential to identify maturational trajectories of the developing brain which is critical for our understanding of brain functioning, the consequences of insult, periods of vulnerability, and for early detection of clinically significant neural anomalies.