Neuroimaging of Babies During Natural Sleep to Assess Typical Development and Cerebral Palsy

Diagnostic magnetic resonance imaging (MRI) of newborns and infants with suspected brain damage is today based on conventional structural MR images with the focus on identifying major structural brain pathology. However, advanced MRI sequences, such as advanced diffusion-weighted imaging, may be more sensitive to detecting brain damage. The NeuroImaging of Babies during natural Sleep to assess typical development and CP (NIBS-CP) study is a longitudinal cohort study of infants and toddlers aged 3-24 months, aimed at studying early brain development in infants at risk of cerebral palsy (CP) and typically developing infants using advanced MRI sequences. The NIBS-CP cohort consists of:

1. Infants at risk for CP, recruited from the Cerebral Palsy: Early Diagnosis and Intervention Trial (CP-EDIT). The ongoing Danish CP-EDIT by principal investigator Professor Christina Høi-Hansen is registered with ClinicalTrials.gov ID NCT05835674. CP-EDIT will enroll 160 infants aged 3-11 months with CP or high risk of CP. Infants will be followed longitudinally with a large battery of motor, neurological, and cognitive assessments during the first two years of life. In addition, CP-EDIT contains information from patient journals on diagnostic MRI, but CP-EDIT does not include the collection of MRI scans as part of the study. For infants enrolled in CP-EDIT, participation in NIBS-CP only includes advanced MRI scans and parent-reported questionnaires. All neurological, motor, and cognitive assessments will be undertaken in CP-EDIT.
2. Typically developing infants, recruited specifically for NIBS-CP. For these infants, participation in NIBS-CP includes advanced MRI scans, neurological, motor, and cognitive assessments (similar to the ones included in CP-EDIT), and parent-reported questionnaires. NIBS-CP provides a control cohort of typically developing infants that will give rise to a normative Danish sample of early brain development to the infants at high risk of CP, as CP-EDIT does not include a control cohort of typically developing infants.

AIMS

The NIBS-CP project aims to:

1. Establish procedures for infant and toddler MRI during natural sleep without the use of sedation or GA at Hvidovre Hospital.
2. Employ advanced MRI sequences for scanning infants and toddlers, e.g., advanced diffusion-weighted imaging, high-resolution structural MRI, and MR spectroscopy, which are likely to be more sensitive in detecting brain injury and damage than conventional diagnostic MRI.
3. Initiate a cohort of typically developing infants and toddlers, and perform longitudinal advanced MRI of infants at high risk for CP. The NIBS-CP cohort will be used to:

• Establish a dataset of normative material of early brain development of Danish children.
• Conduct normative modeling of typical and atypical early brain development, i.e., analogous to growth charts, to inform about differences in brain development at the level of a single child.
• Map relationships between early brain development, and motor function and development. Identifying predictive features of brain structure with MRI is key to the future use of early MRI in the clinical work-up, as these features may enable a better prediction of the prospects of motor function and motor development of the child.

HYPOTHESES

The primary hypotheses are that:

• Infants with CP will show reduced hand functioning which relates to decreased myelination and microstructural integrity primarily in the corticospinal tract.
• Infants with unilateral brain injuries will display asymmetrical hand functioning, which will be linked to asymmetrical corticospinal tract microstructure and myelination.

The secondary hypotheses are that:

• Reduced hand functioning and hand functioning asymmetry will also be related to:
• decreased myelination and microstructural integrity in other key motor regions, such as the basal ganglia, motor cortex, cerebellum, and white matter tracts
• the metabolic profile of the basal ganglia/internal capsule.
• Infants with CP will show reduced gross motor functioning, which relates to decreased myelination and microstructural integrity in key motor regions, such as the basal ganglia, motor cortex, cerebellum, and white matter tracts, e.g., the corticospinal tract.

Furthermore, the normative modeling framework, similar to pediatric growth curves, will be used to elucidate how individual infants with high risk for CP deviate from the norm on different brain outcome measures. The relationships between developing motor functioning and brain structural outcome measures will also be mapped, as little is known about how individual differences in hand and motor functions of the clinically used assessments with the brain outcome measures. Elucidating brain outcome measures that relate to developing motor functions is crucial for using early MRI scans in medical evaluations. Such patterns could help predict a child's motor skills and how these might develop, which could be important for directing future interventions.

METHODS

The following methods will be used for the NIBS-CP study:

Magnetic Resonance Imaging (MRI) The MRI protocol consists of several different structural MRI sequences, diffusion-weighted imaging, and MR-spectroscopy. Children will be scanned using a 3 Tesla MR scanner. The MRI protocol takes approximately 45 minutes. If the infant/toddler moves during a specific sequence, the specific sequence will be redone, if possible. In such cases, the protocol may take up to 1 hour. Earplugs as well as headphones will help minimize the scanner-related noise experienced by the infant/toddler.

Structural magnetic resonance imaging (sMRI) gives high-resolution images of the brain anatomy with different contrasts (e.g., T1-weighted and T2-weighted images). This kind of scanning allows for the quantification of different brain measures, such as volumes of specific brain regions and cortical thickness estimates. Moreover, sMRI provides the images that the neuroradiologist will read and use diagnostically.

Diffusion-weighted imaging (DWI) provides information about the microstructure of both gray and white matter tissue as well as structural connectivity.

Proton MR spectroscopy (MRS) will yield a metabolic profile of the basal ganglia/internal capsule region, including markers of neuronal integrity (e.g., N-acetyl-aspartate and glutamate), and glial markers (e.g., myoinositol). Voxels will include the left or right basal ganglia and internal capsule, to capture the metabolic profile in this region.

Sequences will be ordered so that the most important sequences (sMRI > DWI > MRS) will be acquired first, to optimize the chances of getting the sequences needed for diagnostic purposes for the infants at high risk of CP needing a diagnostic scan.

Neurological assessments Hammersmith Infant Neurological Examination (HINE) will be conducted by a pediatrician at Hvidovre Hospital for the typically developing infants and within the CP-EDIT study for the infants at risk for CP.

Motor and cognitive assessments

Motor and cognitive assessments of the typically developing infants will be conducted by physio- and occupational therapists at Hvidovre Hospital. Infants at risk for CP will undergo the same tests within the CP-EDIT protocol. The following motor and cognitive tests are included:

• Alberta Infant Motor Scale (AIMS)
• Hand Assessments for Infants (HAI)
• Bayley Scales of Infant and Toddler Development, 4th Edition (BSID-4)
• Peabody Developmental Motor Scales, 2nd Edition (PDMS-2)

Questionnaires

Neuropsychological data will be collected using the following parent-reported questionnaires for all participants (typically developing and at-risk for CP):

• Ages \& Stages (ASQ). There are two questionnaires, one assessing motor and cognitive development and one assessing socio-emotional development.
• Demographics \& Background

STUDY OUTLINE

Study outline for the typically developing infants in NIBS-CP:

The NIBS-CP study consists of three assessment rounds:

• Round 1 (inclusion) at age 3-6 months (infants may range from age 2 months or to 11 months), including MRI, HINE, AIMS, HAI, and Demographics \& Background.
• Round 2 at age 12 months, including MRI, AIMS, HAI, ASQ, and Demographics \& Background.
• Round 3 at age 24 months, including MRI, PDMS-2, ASQ, BSID-4, and Demographics \& Background.

Study outline for CP-EDIT participants in NIBS-CP:

Families enrolled in CP-EDIT will be asked to participate in three MRI scans at age 2-11 months (Round 1, inclusion), age 12 months (Round 2, first follow-up), and age 24 months (Round 3, second follow-up) in parallel to CP-EDIT. Participation in NIBS-CP only includes MRI scans and parent-reported questionnaires (ASQ and Demographics \& Background).

STATISTICAL CONSIDARATIONS Longitudinal data will be collected from 200 infants. Infant cohorts are typically smaller than adult cohorts, because of the challenges of MR-scanning this age group (Korom et al., 2022). The success rate of the scans is expected to be around 75% for sMRI, 60% for DWI, and 50% for MRS. The decline in success rate is due to the order of the sequences, given the higher risk of infants waking up the longer the scan time.

Statistical analyses will be conducted in statistical software tools, such as R and SPSS, and in image analysis specific tools, such as FreeSurfer and FSL. We will use e.g., analysis of covariance (ANCOVA) and multiple linear regression for cross-sectional data, and e.g., linear mixed models, generalized additive mixed models (GAMM), and repeated measurements analysis of variance (ANOVA) for longitudinal data

CP patient cohorts are biologically and clinically heterogeneous. Thus, in addition to examining differences in group averages, the normative modeling framework will be used (Rutherford et al., 2023, 2022) to quantify heterogeneity in structural brain measures by mapping structural brain changes at the level of the individual. Normative modeling is a leading tool in precision medicine, as it allows for elucidating differences at the individual level, mapped in relation to a reference model of normative data, similar to height growth charting in pediatric medicine. Here, normative modeling involves charting percentiles of variation across a population in terms of mappings of brain measures, or e.g., between brain measures and motor functioning. Publicly available datasets, such as the Healthy Brain and Child Development (HBCD) study, will be used to enhance our training dataset. In all models with MRI, age, sex, and variables estimating subject motion during MRI will be utilized as covariates.