Skin-to-skin Contact in Neonatal Intensive Care Unit (NICU), Caregiving Touch and Neural Correlates of Slow Stroking Touch in Preterm Infants

Background and rationale One way of operationalizing social touch has been provided by the correspondence between properties of much naturalistic affiliative interpersonal touch and the unique tuning characteristics of low-threshold unmyelinated peripheral afferent fibers. The mechanosensitive C-tactile fibers are activated by gentle skin stroking of the hairy (1-10 cm/s), but not glabrous, skin in adults. In adult, C-tactile fibers activation is correlated with perceived pleasantness. C-tactile fibers are known to mature early and it has been hypothesized that they play a role in encoding socio-affective dimensions of touch. The primary cortical target for C-tactile afferents is thought to be the posterior insular cortex. Importantly, sense of touch develops prior to the 8th week of gestation and insula cortex matures at 27th week of gestation. There is reason to believe that affective touch is already processed during the early stage of development. Recent research on full-term (FT) infants highlighted that the neural correlates of affective touch processing, activated by gentle tactile stimulation are the somatosensory cortex and the insular cortex. However, to our knowledge, no study has examined whether these two regions-of-interest (ROIs), known to process gentle skin stroking, responds to gentle skin stroking in preterm infants (PT).

Moreover, despite, several change were done in the NICU to provide more natural conditions for the PT, these infants often receive very little in terms of early physical contact. Importantly, maternal skin-to-skin contact, has the potential to mitigate some of the adverse consequences of prematurity. A recent Cochrane review reported an average 60% reduction in mortality with this intervention compared to usual care in hospitalized low birth weight infants. The review further showed a 35% relative reduction in nosocomial infections or sepsis, a shorter hospital stay and lower incidence of respiratory infections. Prolonged SSC also provides more effective thermal control, promotes breast feeding, weight gain, shorter hospital stays and stronger neural responses. However, despite the importance of SSC for optimal child development, little is known whether the duration of SSC is associated with the above mentioned ROIs in PT infants.

Finally, In humans, 65% of face-to-face interactions between mothers and infants involve touch communication. Also the quality of the touch matters: gentle stroking touch generates more smiling in infants than static touch. During mother-infant face-to face interactions mothers use touch to engage with their infants, to demonstrate affection. Caregiving touch (CT) reduces infant stress by increasing positive affect19, calms infants in pain and discomfort. Also, CT is associated with sensitive caregiving behavior21, which in turn is associated with better cognitive and behavioral outcomes both in FT and PT children. Importantly, maternal sensitivity predicts infant brain volumes and functional connectivity in 3-6 month-old FT infants. Although CT/sensitive caregiving behavior might contribute to the brain responses in early infancy, the potential link between CT/sensitive caregiving behavior and the above mentioned ROIs response to gentle skin stroking remains to be tested both in PT and FT infants.

Aim 1. To examine differences between PT and FT infants in brain response in somatosensory cortex and the insular cortex (regions-of-interest: ROIs) to skin stroking by gentle brush applied by a trained experimenter during an MRI acquisition.

Experimental Design: FT and PT infants will be scanned around of 2 months of age, corrected age (CA) for prematurity. All infants will undergo an MRI exam with a 3 Tesla Philips Achieva scanner and a 32-channel head coil at the Neuroimaging Lab, Scientific Institute, IRCCS Eugenio Medea. In agreement with Tuulari procedure (Tuulari et al., 2019), during MRI acquisition, a trained experimenter manually will apply gentle brush strokes to infants' right anterior shin region in a proximal to distal direction. The length of the stimulated area will be measured to cover 15 cm, and brush strokes will be applied at a velocity of 5 cm/s for 15s, with randomized inter-stimulus intervals of 10-15 s. The experimenter will be guided by auditory cues delivered through the scanner head phones. At the scanning site, the families will receive by a trained and experienced radiographer and the researchers. Before the scan, the scanning protocol will be revised with the parents and the absence of safety risks will be confirmed by the personnel. All MRI will be performed during natural sleep. All infants will be provided with customized hearing protection. Standard ear muffs will be given to parents, as they will stay in the scanning room throughout the whole scanning session. The scanning will be observed by the personnel from the control room.

Aim 2. To investigate the association between the duration of SSC during the NICU stay and the selected ROIs, activated by gentle skin stroking in PT infants at 2 months.

Experimental Design: the SSC is a standardized care option in NICU and it is not considered an experimental technique. Thus, the possibility to compare infants whose mothers will chose SSC and those who will not, is precluded. An alternative solution will be to recruit dyads via convenience sampling, in according to inclusion criteria. The effects of SSC will be analyzed based on the duration of SSC during the infant hospitalization at the NICU. The amount of SSC will be obtained through parent-report diaries. Infants (and their mothers) will be enrolled when their medical situation will be stabilized and if the mothers will agree to perform SSC. Infants will take out of the incubator, undressed (wearing only a diaper and some-times a cap), and placed between the mother's breasts. The procedure to the MRI exam has already been reported in the aim1.

Aim 3. To assess the association between both CT/sensitive caregiving behaviors in the home environment and during mothers' displays of touching behaviors and maternal sensitivity during face-to-face interaction and brain responses in the selected ROIs to gentle skin stroking in PT and FT infants.

Experimental Design: when infants reached 2, before the MRI acquisition, mother-infant interaction will be videotaped. The infant will be placed in a supportive seat when alert and contented; the mother will ask to sit opposite her infant, and play with him/her for 5 min. Two digital cameras, one focused on the infant and the other on the caregiver, will used to produce a single image with simultaneous frontal view of the face, hands and torso of infant and mother. Infant and maternal behavior and their interactions will be coded using the coding system Global Rating Scales of mother-infant interaction (GRS), which have been extensively validated in normative and clinical populations. Maternal touch during interaction will be coded using The Caregiver-Infant Touch Scale (CITS), a reliable and systematic coding system. The CITS is a measure of the qualitative (and quantitative) changes in tactile stimulation produced by caregivers during face-to-face interaction with their infants. Maternal sensitivity and maternal touch will be coded by two trained researchers who will be blind to PT or FT conditions. Moreover, in order to assess the duration and the quality of mother's tactile stimulation in the infant's home environment two self-report measure will be used: the Baby Care Questionnaire and the Parent-Infant Caregiving Touch Scale. Both measure have been used in normative and clinical populations.

Methodologies and statistical analyses:

General methodology The present study has a short-term longitudinal and prospective study design and involves two groups of infants (PT and FT). The study is made up of three phases, corresponding to the three study aims and experimental designs reported above.

• Phase 1 (study aim 2; experimental design aim 2): time-frame from birth to NICU discharge in PT infants only.
• Phase 2 (study aim 3; experimental design aim 3): time-frame from NICU discharge to 2 months CA for PT and from birth to 2 months for FT, including the videotaping of the face-to-face interaction at home.
• Phase 3 (study aim 1; experimental design aim 1): experimental MRI scanning visit occurring around 2 months CA.

Sample size estimation To estimate sample size for minimum sample to detect a clinically significant difference we referenced Thirion et al. (2007) that indicated that N=20 is a minimum if one wants to have acceptable reliability in a MRI study. Starting from this, a final sample size N=20 per group was estimated (N=40 subjects); in agreement with the exclusion rate found in Tuulari et al. (2019) due to the excessive motion of the infants during the MRI scan visit, the investigators assumed a more conservative 10% rate, estimating a final sample size N=25 per group was estimated (N=50).

MRI data processing and analysis Preprocessing and statistical analyses will be conducted using Statistical Parametric Mapping 12 (SPM12). Functional data preprocessing will include slice time correction, realignment to the first volume of the first run, and normalized. Motion artifacts correction will be performed.

Well-established procedures for co-registration and normalization will be computed using Advanced Normalization Tools. Individual percent signal change in response to skin stroking will be then extracted from each ROI. In order to examine brain responses outside the a priori defined ROIs, an exploratory whole-brain analysis will be conducted.

Plan of analyses Primary hypotheses will be tested using regression models, including group, stratification variables and any identified baseline confounders. The alpha for primary outcomes will be set at p<.05. Hypothesis testing will be supplemented with effect size estimates and confidence intervals. Analyses will be conducted on an intent-to-treat basis. Multiple imputation methods will be used to adjust effect estimates for missing data on the assumption that data is missing at random.

Sensitivity analyses will be conducted to determine the robustness of findings to differing assumptions about missing data. Longitudinal analyses will be conducted using linear mixed modelling.

Significance and Innovation Till now few studies evaluated how affective touch may be associated to brain responses in neonates, and, to our knowledge, no study has included PT infants. Thus, to investigate the associations between SSC and later neural correlates of early tactile experience may potentially have relevant implications for early life interventions in NICU. For example, a recent study documented that unlike SSC, a mattress that attempts to mimic breathing motion and heartbeat sounds does not affect physiological responses of PT infants. This suggests that parental early physical contact have a unique and specific impact on the infant. Moreover, the study will integrate this finding with observational data from mother-infant ace to face interaction to better clarify the role of caregiving touch and how it could be related to the brain responses.