Data on parent-infant physical closeness and infants' auditory environment will be collected among preterm infants when they are at gestational age of 32 to 34 weeks. The follow-up includes eye-tracker test at 7 months of corrected age for face preferences of the infants and simultaneously parents' eye movements and pupil diameter responses. During the second year, the follow up includes MacArthur Communicative Development Inventories (lexical development) at 12 and 24 months of age; language development test (Reynell Developmental Language Scales III) and developmental test (Bayley Scales for Infant development Edition III) at 24 months of corrected age.

Study population: 63 infants from Turku University Hospital and 53 infants from Tallinn Children's Hospital were recruited in 1/2017-12/2020. Inclusion criteria: Born before 32+0 week of gestation without life-threatening diseases or congenital anomalies or syndromes.

Infant auditory environment was recorded for 16 hours at 32+0 to 33+6 weeks of postmenstrual age. In addition, Parental Closeness Diary was recorded for 14 days to measure the parents' presence and parent-infant closeness to get an estimate of language environment of the child during a 14-day period at around 32-34 weeks of gestation. Edinburgh Postnatal Depression Scale was collected during the recording day from both parents to identify parents' risk for perinatal depression. During the day of recording, the research nurse reported the time when a nurse was in the patient room. The auditory environment will be analyzed quantitatively and qualitatively.

Outcomes: Infant face preference for familiar and unfamiliar faces. At the corrected age of 7 months (± 7 days), an eye-tracking based test battery will be used to assess infants' attention to familiar and unfamiliar faces. During the eye-tracking testing, the infants will sit on their parents' lap at ~60 cm distance from a stimulus monitor. Infants' eye movements will be measured by using a screen-based eye-tracker and a video camera.

The test will consist of a task assessing infants' attention to faces expressing neutral, happy and fearful emotion, posed by the infant's own parents and an unfamiliar (female) adult. In this task, the first stimulus (a non-face pattern or a face) is presented at the participant's point of fixation in the center of the computer display, and the second stimulus (a high-contrast distractor) 1 second later ~13-15° to the left or right of the center. Gaze disengagement times from the central to the lateral stimulus will be extracted from the eye tracking data (or from video data, if valid eye tracking data are missing). Disengagement times from trials conforming to the following a priori inclusion criteria will be used in the analyses: i) a sufficient length of fixation on the central stimulus (i.e.,70 % of the time) during the analysis period, which start at the onset of the trial and ends at the onset of gaze disengagement or at maximum time out value (1000 ms post-target), ii) sufficient number of valid samples in the gaze data (i.e., no gaps longer than 200 ms), and, iii) valid information about the eye movement from the central to the lateral stimulus (i.e., the eye movement did not occur during a period of missing gaze data) will be retained for analysis. The duration of attention dwell time on the first stimulus (face or non-face pattern) will be determined for the period starting 150 ms from the onset of the lateral stimulus and ending 1000 ms after the lateral stimulus onset. The duration will then be converted to a normalized dwell time index score by using the formula described in Leppänen et al. (2015). The shortest acceptable saccadic eye movement latency (150 ms) results in a dwell time score of 0, and the longest possible latency (or a lack of saccade, which is equal to the last measured time point of the first stimulus at 1000 ms) in a dwell time index score of 1. The dwell time indices will be calculated separately for each trial and averaged for non-face patterns (i.e., low-salience stimulus), for faces by familiarity (across emotion categories), and for faces by emotion category (across familiarity and separately by familiarity). The primary outcome of the analysis will be the mean dwell times for familiar and unfamiliar faces, adjusted for the mean dwell time for respective non-face patterns. All infants with >2 valid trials per condition will be retained in the final statistical analyses.

Parents' attention and physiological responses to infant faces. Parents attentional and physiological (i.e., pupil dilation) to non-face patterns and faces of their own infant(s) and unknown infants will be assessed at infant age 7 months. Parents will be asked to provide pictures of their own infant(s) prior to each laboratory visit. Pictures displaying positive/content and distressed emotions will be asked.

Picture from an existing dataset will be used as unknown infant pictures. Non-faced patterns will be created by phase-scrambling the pictures so that they retain the low-level characteristics of the original faces, but are not identifiable as faces.

To assess attention, the adults will be asked to follow (by looking) a geometric shape (a checkerboard pattern, measuring 2x2 degrees) as it shifts location from the center of the screen (500-msec) to the left or right side of the screen (~13 degrees from the center). After the participant has shifted gaze from the center to the side (as determined by the eye tracker), the pattern will be replaced by a picture of a non-face pattern or a face for 300-msec. The side of the non-face pattern and face (A= non-face left, face right or B= face left, non-face right) will be alternated between blocks of 20 trials following an ABBA-BAAB design. Using this approach, estimates of saccadic reaction times (SRTs) to the location of own infant face(s) vs. non-faces and two unknown infant faces vs non-face patterns. Because saccadic reaction times will always be initiated by the shifting location of the shape (a checkerboard pattern), any difference in mean SRT must reflect an endogenous or high-level bias towards one of the stimuli (e.g. own infant face), instead of exogenous stimulus properties. To extract attention indicators from the adult data, the latency of gaze shifts between the center of the screen and lateral stimulus positions (following shifts in stimulus position) will be analyzed by using the same procedures as those used for analyzing infant attentional data. To assess parents' sympathetic activation to infant cues, we will follow the procedures of our previous study (Yrttiaho et al., 2017). The adult will see a series of trials consisting of a short 1000-ms pre-stimulus interval with the non-face pattern and a 4000-ms face stimulus (own or unknown infant). Pupil size will be measured from both eyes during the pre-stimulus interval (with random visual pattern that has the same low-level stimulus characteristics as paired face stimulus) and the presentation of the face stimuli. By matching the low-level characteristics of the two consecutive stimuli, the present approach minimizes the possibility that the recorded pupil dilation responses are driven by changes in stimulus luminance. Latencies for each stimulus category will be extracted and used as a primary dependent variable in subsequent statistical analyses. Physiological data from adults (pupil diameter) will be preprocessed using the steps described in our previous study (Yrttiaho et al., 2017), and pupil diameter in two time windows (based on prior studies, these windows will be 300-1200 and 3000-4000 ms after face onset, respectively) will be averaged to extract of pupil constriction and dilation, respectively. The dilation phase will be used as the primary index of sympathetic arousal to infant distress cues.

Information about the child's developing abilities of early language (receptive and expressive language development) will be assessed by MacArthur-Bates Communicative Development Inventory (MCDI) at 12 and 24 months of corrected age.

At two years of corrected age the language skills of the children were measured using the Reynell Development Language Scales III and MacArthur Communicative Development Inventories (the Words and Sentences form). In addition, children's cognitive development was measured using the Bayley Scales of Infant Development III.

Inclusion Criteria:

• Born before 32+0 week of gestation

Exclusion Criteria:

• Life-threatening diseases, congenital anomalies or syndromes or chromosomal abnormalities.

63 preterm infants with their parents were recruited from the NICU of Turku University Hospital and 53 preterm infants with their parents from the NICU of Tallinn Children's Hospital in 1/2017-12/2020.