Participants
The present study is part of a larger ongoing longitudinal study (the TRIXY Early Childhood Study - Leiden the Netherlands), which includes children with SCT and typically developing children aged 1-7,5 years. The TRIXY Early Childhood Study aims to identify neurodevelopmental risk in young children with an extra X or Y chromosome. A group of 100 children with SCT (range 1-7 years old, Mage= 3.69, SD = 1.91) was included in this study, as well as a population-based group of 98 children without SCT (42 boys, Mage= 3.66, SD = 1.62). Mean age did not significantly differ between groups (t (196) = 0.11, p = .913). The SCT group consisted of 34 girls with 47,XXX (34%), 45 boys with 47,XXY (45%) and 21 boys with 47, XYY (21%). In order to investigate social attention and affect recognition outcomes in different developmental stages in early childhood, the participants were divided in three age groups: children one and two years old (n = 61, Mage = 1.47 years, SDage = 0.33, 32 SCT (6 47,XXX, 18 47,XXY, 8 47,XYY), 29 without SCT), children three and four years old (n = 83, Mage = 3.88, SDage = 0.58, 40 SCT (13 47,XXX, 19 47,XXY, 8 47,XYY), 43 without SCT), and children five, six and seven years old (n = 54, Mage = 5.86, SDage = 0.67, 28 SCT (15 47,XXX, 8 47,XXY, 5 47,XYY), 26 without SCT). To test if the frequencies of SCT types differed across age groups, a χ2 test was conducted, an no differences were observed (χ2 (4) = 8.40, p = .078).
Recruitment and assessment took place at two sites: the Trisomy of the X and Y chromosomes (TRIXY) Expert Center the Netherlands, and the eXtraordinary Kids Clinic in Developmental Pediatrics at Children’s Hospital Colorado/University of Colorado in the USA. Children in the SCT group were recruited with the help of clinical genetics departments (from the Netherlands, Colorado, and Belgium), as well as through patient-advocacy groups and social media postings. For the SCT group, recruitment bias was assessed, three subgroups were identified: (1) ‘Active prospective follow-up’, which included families who were actively followed after prenatal diagnosis (51% of the SCT group), (2) ‘Information seeking parents’, which included families who were actively looking for more information about SCT without having specific concerns about the behavior of their child (29% of the SCT group), and (3) ‘Clinically referred cases’, which included families seeking professional help based on specific concerns about their child’s development (20% of the SCT group).
The diagnosis of SCT was defined by trisomy in at least 80% of the cells, which was confirmed in the study by standard karyotyping. Sixty-seven children were diagnosed prenatally (65.3%, 20 girls with XXX, 32 boys with XXY, 15 boys with XYY), and 33 children postnatally (34.7%, 14 girls with XXX, 13 boys with XXY, 6 boys with XYY). 24 out of 45 boys with 47,XXY received testosterone treatment (53.3%).
Children without SCT were recruited from the western part of the Netherlands, and approached with information brochures about the study. All participants were Dutch (The Netherlands) or English (USA) speaking, had normal or corrected-to-normal vision, and did not have an history of traumatic brain injury. For ethical reasons, children without SCT were not subjected to genetic screening, as these children were meant to be a representation of the general population. As the prevalence of SCT is ~1 in 1000, the risk of having one or more children with SCT in group children without SCT was considered minimal and acceptable.
Eye tracking paradigms
Social attention to eyes. The Static Facial Emotions paradigm consisted of 16 static photographs of cross-cultural actors with an equal distribution of two facial emotions (happy and angry), and of male and female actors (see Figure 1). The photographs were taken from the Karolinska Directed Emotional Faces (KDEF, (Lundqvist et al., 1998). These KDEF pictures have no background, and actors have no visible beards, mustaches, earrings, eyeglasses, or make-up. The photographs were presented to the child, displayed at the center of the screen, in a counterbalanced order. The child was exposed to each picture for 3 s., with a 2 s. inter-item interval during which a attention grabber (i.e. a picture of a toy or animal, together with a sound to grab the child’s attention), was presented in one of the four corners of the screen, to prevent for the automatic response to fixate at the center of the screen.
Social attention and social load. The Dynamic Social Information eyetracking paradigm consisted of two natural and dynamic conditions: single face (SF) and multiple faces (MF). Six blocks were included (3 single face, 3 multiple faces) of 15 s. each. The total time of the stimulus set was 90 s. The blocks were presented in an alternate order (i.e. single, multiple, single, multiple, single, multiple). In each block, a video clip was presented to the child. In the single face condition, one face of a child was on the screen, in the multiple faces condition, two or more faces were on the screen (child-child or child-adult). The video clips consisted of subjects with different cultural background, and were extracted from the TV broadcasted series ‘Baby Einstein’ (Kids2, 2015, see Figure 1). The videos were accompanied by unsynchronized classical instrumental music, and no speech was involved. As this task did not involve language and used age-appropriate stimuli, it was considered to be appropriate for participants in all countries. In a group of non-clinical young children aged 3-7 years, this eye tracking paradigm was found to be significantly predictive of real-life social behaviors, and independent of age, IQ, or gender (Van Rijn et al., 2019).
Eye tracking equipment and procedures
Gaze data within specific areas of interest (AOIs) was collected using the Tobii X2-60 eye tracker (Tobii Technology AB, Danderyd, Sweden), which records the X and Y coordinates of the child’s eye position at 60 Hz by using corneal reflection techniques. The computer with eye tracker was placed on a table adapted to the height of the seat, and the child was seated in a car seat at 65 cm viewing distance. A 5-point calibration procedure was used, with successful calibration defined as a maximum calibration error of 1 degree for individual calibration points (i.e. < 1 cm at a distance of 65 cm from the eyetracker). After the calibration procedure, the child was instructed to watch the movie clips and pictures on the computer. The two eye tracking paradigms started with an attention grabber (e.g. a moving picture of an animal, shown on a black background and accompanied by a sound) to direct the attention of the child to the screen.
Only gaze data points with validity code ‘0’ (indicating that high quality data of both eyes was collected) were included in the analysis. Gaze data was processed using Tobii Studio (version 3.2.1), using the Tobii Identification by Velocity Threshold (I-VT) fixation filter. This filter controls for validity of the raw eyetracking data making sure only valid data were used (Olsen, 2012). The ‘Dynamic AOI’ tool was used to draw AOIs, drawn with a one centimeter margin, to ensure that the AOIs were sufficiently large outside the defining contours to reliably capture the gaze fixation (Hessels et al., 2016). In the Static Facial Emotions paradigm, AOIs were grouped into the category eyes, and for the whole screen, first fixations within the eye AOI, and total fixation duration within the eye AOI were measured, in order to study social attention to eyes. In the Dynamic Social Information paradigm, dynamic AOIs were grouped into the following categories: face and eyes, and for the whole screen, total fixation duration within AOIs were measured in two conditions: Single Face condition (low social load) and Multiple Face condition (high social load). In order to evaluate the amount of nonvalid eye tracking data, the total visit duration toward the whole screen was calculated, divided by the duration of the clip, multiplied by 100, reflecting the percentage of valid data collected during each of the eye tracking tests. For both paradigms, proportions fixation duration were calculated by taking the total fixation duration within the AOI, divided by the total visit duration toward the whole screen of the individual child, multiplied by 100, reflecting the percentage of time children were attending to an AOI. In the facial emotion paradigm, proportions first fixations within the AOI eyes were calculated by taking the number of photographs where participants fixated first on the eyes, divided by the total number of photographs (max = 16).
NEPSY Affect recognition
The Affect Recognition subtest of the Developmental NEuroPSYchological Assessment, second edition (NEPSY-II neuropsychological test battery, Korkman et al., 2007) was designed to assess children’s ability to discriminate among common facial emotions from photographs of children, and used in this study to measure task performance of affect recognition skills. The task has been normed with typically developing children aged 3-16 years old, and was administrated in a subsample of the study sample with the age of 3 years and older (n = 138). During the task, participants are required to match faces of different children with different cultural backgrounds who show the same emotional expressions (happy, sad, angry, disgust, fear and neutral). The participant indicates if two expressions are the same or different, determines which two faces have similar expressions, or identifies two children with expressions that match a third child’s face. The total raw score range is between 1 and 25, with higher scores reflecting a better ability to recognize facial expressions. Besides raw scores, percentile scores as compared to norms from the general population can be calculated. Dependent upon the spoken language of the child, the Dutch or English norms were used. Percentile scores were labeled as being in the average range (percentile score > 25), the borderline range (11 < percentile score > 25), the below expected level (3 < percentile score > 10), and the well below expected level (percentile score ≤ 2).
Cognitive assessment
To measure global level of intelligence and language three tests were administrated. The Bayley III (subscale cognitive scale (Bayley-III: Bayley Scales of infant and toddler development., 2009) was administered to children with the age of 1-2 years old. In the older children four subtests of the Wechsler Preschool and Primary Scales of Intelligence, 3rd edition (WPPSI-III) were used to estimate global level of intelligence (children aged 3 years: Block Design, Receptive Vocabulary, Information, Object Assembly, children aged 4 years and older: Block Design, Matrix Reasoning, Vocabulary, and Similarities, (Wechsler, 2002). The Peabody Picture Vocabulary Test (PPVT, Dunn & Dunn, 1997) was used to measure receptive language level in children aged 3 years and older.
Study procedures
Assessment took place at various sites (Colorado USA and the Netherlands) either in a quiet room at the University or at home. To standardize the testing environment, the testing set-up and research protocols were identical at all sites. Researchers from Leiden University were responsible for project and data-management (i.e., training and supervision of researchers processing and scoring of data). Administration of cognitive assessment and the NEPSY was performed on a table by trained child psychologists or psychometrists in Dutch or English (dependent on the first language of the child). The eye tracking procedure took place during a separate appointment, within one week after the NEPSY administration. The laptop with the eye tracker was placed in a small tent to standardize the testing environment, and to control for lighting conditions. The child was seated in a car seat in front of the eye tracker. The examiner was seated beside the child (directing Tobii Studio with a remote keyboard) and started the calibration procedure. Eye tracking paradigms were shown in a fixed order (single/multiple faces, facial emotions). Parents were allowed to stay in the room (out of sight) and were asked not to communicate with their child during the procedure.
Data analyses
Statistical Package for the Social Science (SPSS) version 25 was used for statistical analyses. A χ2 test was used to compare the distribution of karyotypes within the three age groups. Pearson’s correlation analyses were used to measure the association between main outcome variables (i.e. social attention and affect recognition) and global cognitive functioning and receptive language abilities. For group wise (SCT vs. children without SCT) comparisons of proportions first fixations, proportions duration fixation within the AOIs in the three age groups, and affect recognition skills in two age groups (M)ANOVAs were used. Pillai's trace was used to assess the multivariate effect. Significant multivariate effects were post-hoc analyzed with univariate ANOVAs to determine the locus of the multivariate effect. The moderating effect of age on social attention and affect recognition outcomes between the SCT and typically developing group was assessed using PROCESS, a bootstrapping, nonparametric resampling procedure (Hayes, 2009). Bootstrapping analysis with 5000 resamples was done to test for a significant moderating effect using the SPSS macro developed by Hayes (2017). Outcome variables and moderator variable (i.e. child’s age) were centered. In this analysis, the moderation effect is significant if the 95% bias corrected confidence interval for the moderator effect does not include zero. Influence of karyotype accounting for the effect of age was tested by an MANCOVA. (M)ANOVAs were used to investigate differences between recruitment groups, and influence of research sites was analyzed with independent t-tests. Statistical analyses were performed one-tailed (SCT vs. children without SCT) or two-tailed (influence of karyotype/recruitment bias/research site), and level of significance was set at p < .05. In case of significant differences, Cohen’s d or partial η2 were used to calculate effect sizes.