Study Design
Fifteen children born extremely preterm (25-27 gestation weeks, GW; 53% female and 47% male) were selected from a wide sample of infants diagnosed with prenatal and/or perinatal risk factors for brain damage and followed up into childhood. These children were part of the longitudinal transdisciplinary research protocol of the Neurodevelopmental Research Unit of the Institute of Neurobiology, National Autonomous University of Mexico. The protocol includes newborn infants aged three months or less (corrected age) with prenatal and perinatal risk factors for brain damage, excluding the presence of genetic factors associated with brain damage, cardiovascular pathologies, brain malformations, and/or chromosomal aberrations. The infants are evaluated by a neuropediatrician and screened with electroencephalographic, auditory and visual event-related potential, brain magnetic resonance imaging (MRI), nutrition, attention, neurohabilitation, rehabilitation, neurodevelopmental, socioeconomic, and psychological evaluations and followed up until eight years of age (Harmony, 2017, 2021).
The present work analyzes data regarding prenatal and perinatal risk factors, clinical MRI brain damage markers, MRI subcortical volumes (acquired at the beginning and after neurohabilitation), treatment adherence, and neurodevelopment outcomes.
Prenatal and Perinatal Risk Factors
A neuropediatrician collected risk factors and diagnostic label data from the report generated by the hospital where each infant was born. The participants’ data is summarized in Table 1.
Katona Neurohabilitation Treatment
Katona Neurohabilitation is a diagnostic and therapeutic tool based on the vestibular and proprioceptive stimulation produced by the intensive practice of genetically determined, human-specific elementary movements. These complex movements are assessed, and their practice is indicated by a specialized physiotherapist who, according to the diagnostic evaluation results of muscular tone and neurological alterations in the infant, decides the infant-specific treatment plan and trains the primary care provider(s) on the administration of the appropriate positions and movements to practice with the infant at least three times a day, every day. The physiotherapist supervises and adjusts the treatment during assessment/feedback follow-up sessions (Barrera Reséndiz, 2015; Harmony, 2021). Treatment adherence was evaluated as the number of assessment/feedback follow-up sessions attended by the primary care provider with the infant (0-12 assessment/feedback sessions).
Brain MRI
MRI Acquisition
Infant's brains were scanned during natural sleep and using ear protection. MRI acquisition was performed at the beginning and after treatment either with an Intera Phillips 1.0T (Philips Medical Systems, Best, Netherlands) or a Discovery MR750 3.0T GE (General Electric Healthcare, Milwaukee, Wisconsin, US) scanner. Ages at scan are shown in Figure S1.
Anatomical images acquired with the Intera scanner included axial and sagittal T1-weighted conventional spin echo (SE), with a repetition time (TR)/echo time (TE) = 405/15 ms, flip angle = 62°, 15 slices, slice thickness = 5 mm, matrix = 256 × 166, field of view (FoV) = 220 × 220 mm, and voxel size = 6.0 × 0.8 × 0.8 mm3; and axial and coronal T2-weighted SE, with TR/TE = 2600/150 ms, flip angle = 90°, 30 slices, slice thickness = 6 mm, matrix = 256 × 153, FoV = 200 × 200 mm2, and voxel size = 0.8 × 6.6 × 0.8 mm3.
The images acquired with the Discovery MR750 scanner used a 16-channel neurovascular head coil (HDNV). These images included: coronal 3D T1-weighted SPGR, with TR/TE = 6/2 ms, flip angle = 12°, 392 slices, slice thickness = 1 mm, matrix = 224 × 224, FoV = 220 × 220 mm2, and voxel size = 0.8 × 0.5 × 0.8 mm3; and coronal 3D T2-weighted SE, with TR/TE = 2500/68 ms, flip angle = 90°, 196 slices, slice thickness = 1 mm, matrix = 224 × 224, FoV = 220 × 220 mm2, and voxel size = 0.8 × 1.0 × 0.8 mm3.
MRI Individual Analyses
After formatting the images with the dcm2bids and dcm2niix tools (Bedetti et al., 2021, Li et al., 2016), the MRI Denoising Package (Coupé et al., 2012) for Matlab (version R2022a) was applied to improve the signal-to-noise ratio. Then, the images were resliced to a 1 x 1 x 1 mm3 voxel size using the mri_convert tool of the FreeSurfer software (version 7.0.2). The FreeSurfer pipeline was used to segment brains through the Aseg tool (Fischl et al., 2002) in images of children older than two years. Images of two-year-old or younger brains were segmented using the Infant FreeSurfer pipeline (Zöllei et al., 2020). Figure S2 shows an example of brain segmentations from images acquired at the beginning and after Katona neurohabilitation. Once extracted, the volume values were divided by the estimated total intracranial volume, to control individual differences in head size. Both the absolute and the resultant relative volumes were used to perform statistical analyses.
Neuropsychological Screening
The Mental Development Index (MDI) and the Psychomotor Development Index (PDI) of the Bayley Scales of Infant Development, Second Edition (Bayley-II; Bayley, 1993), were used to evaluate the neurodevelopmental outcomes at three years of age.
Statistical Analyses
The SPSS v15.0 software was used to perform the statistical analyses. First, variable normality was assessed with Kolmogorov-Smirnov tests. Because several variables did not show normal distributions, Wilcoxon signed-rank tests were applied to evaluate differences in subcortical volumes at the beginning compared to after the treatment (p < 0.05).
Then, Spearman’s partial correlations, controlled by sex, were performed to evaluate the associations between subcortical relative volumes, age at first scan, time between scans, treatment adherence, and neurodevelopmental outcomes (p < 0.05; uncorrected).
Finally, hierarchical regressions were performed to search for the combinations of variables that best predicted MDI and PDI scores (p < 0.05). The variables were introduced in two stepwise blocks: 1) age at first scan, time between scans, sex, and treatment adherence, and 2) subcortical relative volumes.