Original Study

Thanks to the contribution of our participants (that’s you!), our original NIH funded study lead to several findings and the publication of several papers papers that have added to our knowledge of preschool development and preschool anxiety.


Sensory Over-Responsivity: An Early Risk Factor for Anxiety and Behavioral Challenges in Young Children.

Anxiety in children can present itself in many forms such as increased trouble controlling their sleep, increased irritability, or even increased stomach issues. These behavioral presentations of anxiety have been linked to a risk factor known as sensory over-responsivity. Sensory over-responsivity can be understood as the intense and uncommon reaction to everyday sensory stimuli like the tag on one’s shirt or loud noises. While much of the research on sensory over-responsivity has been conducted in relation to children with Autism Spectrum Disorder and their behavioral challenges, other evidence suggests there is a relationship between increased behavioral challenges and the development of psychiatric disorders such as anxiety in children without Autism Spectrum Disorder. Until this study, no longitudinal studies, that is, studies of the same group of people over multiple time frames, of the relationship between sensory over-responsivity and anxiety in children without autism were done.

In this Duke Preschool Anxiety Study, three phases of sampling were conducted to gather a sample of pre-school aged participants. From the 3.433 individuals who completed a screening phase, 917 were selected for an at home assessment of which 502 were selected for the laboratory phase. A smaller subset of 254 individuals of the 502 who had met criteria for anxiety actually participated along with a random sample of non-anxious individuals. Following this data collection, 191 individuals participated in a follow-up study, hence the longitudinal nature of the research, and 127 of these individuals had met criteria for an anxiety disorder when they were 2-5 years old. Different psychiatric assessments were administered for data collection in terms of anxiety criteria, sensory over-responsivity criteria, and associated behavioral challenges. The findings of this study support the suggestion that sensory over-responsivity during preschool is predictive of anxiety symptoms in this age group such that children with high levels of sensory over-responsivity symptoms as preschoolers have higher levels of anxiety symptoms at school-age. This means increased and unusual reactions to common stimuli can be associated with higher levels of both irritability and sleep problems at school-age.

Magnetic susceptibility of brain iron is associated with childhood spatial IQ.

A lack of iron, or iron deficiency, can have negative implications during adolescent brain development such as decreased abilities in learning, memory, verbal and non-verbal reasoning, and visual-spatial cognition. With respect to the specific regions responsible for the production, processing, and integration of iron in the brain, neuroimaging can identify which of these regions is lower in iron concentration. A magnetic resonance imaging (MRI) technique known as quantitative susceptibility mapping (QSM) that measures iron in brain tissue was used in this study to explore the relationship between lack of iron in adolescent brains and spatial IQ.

39 children aged 7 to 11 years old who met the criteria of selection participated in this study. Each of these children had participated in a longitudinal study—a study that follows the same group of individuals over an extended period of time—prior to their participation in this study, thus they had completed a cognitive assessment measuring spatial IQ and other factors at 6 years old. These scores were used in combination with a Brain MRI and QSM to gather data on brain activity for analysis. Based on the results, this study found that iron in a specific region in the brain known as the right caudate is associated with spatial IQ in school-age children. Spatial IQ is related to one’s ability to visualize and conceptualize their orientation and surroundings in space. Based on evidence as well as this study, decreased spatial IQ can be linked to iron deficiency in infants, adolescents, and adults due to multiple factors like neurotransmitters in the brain being functionally impaired.

Psychological and Psychosocial Impairment in Preschoolers With Selective Eating.

In this study, we wanted to know the importance of moderate and severe selective eating (SE) in order to help health care providers recognize when selective eating needs intervention. The participants of our study were 917 children (ages between 24 to 71 months) and their caregivers. We interviewed the caregivers with the Preschool Age Psychiatric Assessment. This assessment asked about the child’s mental conditions, home environment, and regular behaviors such as eating. A sample of 188 pairs of children and their caregivers were assessed a second time about 24.7 months after the first assessment. Our results show that both moderate and severe levels of selective eating (SE) were associated with mental health symptoms of anxiety, depression, and/or attention-deficit/hyperactivity disorder (ADHD). The more severe the selective eating was, the worse the mental health symptoms seemed to be. We believe that health care providers should intervene even at moderate levels of selective eating.

Preschool anxiety disorders predict different patterns of amygdala-prefrontal connectivity at school-age.

In this study, we wanted to know if a history of generalized anxiety, separation anxiety, and/or social phobia in preschool children was connected to amygdala-prefrontal dysregulation, which means difficulty managing emotional responses, as school-age children. We also wanted to know if distinct anxiety disorders showed any differences in patterns of this dysregulation.

The participants of our study were children who were in a 5-year study of brain development and anxiety disorders in early childhood. We used the Preschool Age Psychiatric Assessment (PAPA) to assess symptoms of generalized anxiety, separation anxiety, and social phobia among 2 to 5-year-old preschool children. We repeated the PAPA when the children turned 6 years old. When the children were 5.5 to 9.5 years old, we conducted functional MRIs (fMRIs) to assess their brains’ responses to seeing angry and fearful faces. Our results suggest that a history of preschool anxiety can have lasting effects on children’s brain development. We need more research on how different preschool anxiety disorders have different effects.

The NIMH Child Emotional Faces Picture Set (NIMH-ChEFS): a new set of children’s facial emotion stimuli.

Due to the lack of reliable and valid sets of children’s faces for emotion perception and face processing research studies, researchers attempted to test a set of images of emotional faces of children to test its validity. The sets came from the National Institute of Mental Health Child Emotional Faces Picture Set (NIMH-ChEFS) and included 482 images of fearful, angry, happy, sad and neutral faces with 2 gaze conditions: direct and averted (eyes turned away). The measure of the validity face pictures sets are determined by the extent to which raters agree between the “images a priori emotion designation and the raters’ identification of emotion type” (Egger et al., 2011). By a priori, this is the extent to which knowledge or reasoning is determined by analyzing concepts and information rather from observation or experience. Most studies in the past on child developmental psychology involve children identifying and responding to faces of adults, thus there is a need for the development of standardized stimuli sets of child faces that children can respond to in order to better study emotion perception and processing among children. The NIMH-ChEFS created an emotional stimuli set of children’s faces between the ages of 10 to 17 which were evaluated by 20 adult raters. In the picture set of over 500 pictures that were evaluated, only full front profile pictures of the children were used. The children were predominantly caucasian and female.

The evaluation of the picture set took place at Duke University Medical center by a research group that selected the best pictures for a longitudinal study on childhood anxiety disorder. The adult raters were between ages 22 to 70. The raters underwent a rating task in which a screen for each image was displayed with an evaluation component; raters completed three tasks in which they had to select the emotion the picture represented (afraid, angry, happy, sad, or neutral), they were given a slider function to rate the intensity of the emotion from “mild” to “strong”, and the rater had to rate how accurate they thought the picture represented the emotion they chose by using another slider function that ranged from “poorly” to “very well.” The tasks were labeled by the following: 1. Agreement/Disagreement rate; 2. Intensity; 3. Representativeness. Researchers examined the results by the emotion condition and gaze type, as well as the number of mismatches among the different emotions. A goodness score ranging from -100 to 100 was computed by the researchers to determine the strength of the stimulus. They compared ratings by the extent to which raters agreed or disagreed, the level of intensity and the representativeness. For about half of the stimuli (49.5%), there was an agreement among raters and the a priori classification. In analyzing the differences among the emotion types that were rated, at least 15 raters identified all the happy faces and 76% of the sad direct gaze faces were identified by the 15 raters. 6 afraid faces, 10 angry faces, 13 neutral faces and 23 sad faces had less than 75% agree among the raters. Researchers create a table to display the percent agreement and disagreement for the emotions and its accuracy for each emotion type. They found that “of the 5.7% mislabeled afraid pictures, 45% were called angry and 42% were called sad, with only 10% being labeled neutral and less than 2% labeled happy. Half of the mislabeled angry pictures were sad, with 23% choosing afraid and 27% choosing neutral. For the 0.6% mislabeled happy pictures, none labeled them as neutral. Of the mislabeled neutrals, 45% were identified as sad, with another 24% labeled as afraid and 29% labeled as angry. Lastly, for the sad faces, 47% were named angry, 23% were called afraid and 28% called neutral. Few of the non-happy faces were identified as happy.”

They found that intensity and representativeness were highly correlated; for example, happy faces were found to actually represent happiness. They found no significant differences between direct and averted gaze conditions in analyzing the ratings for intensity and representativeness. Interestingly, there were significant differences in ratings between male and female for negative emotions. For example, the afraid faces of the girls were found to be intense and represented fear much more than the faces of boys. This was also found for sad faces; however, for angry faces, the boy faces were found to be rated as more intense than angry girl faces. Between happy and neutral faces, no significance differences were found. Overall, happiness had the highest agreement rates whereas fear, anger and sadness and lower agreement rates. Sad faces were the most inaccurately identified emotion. Researchers found an acceptable set of 38 to 48 stimuli in each emotion type. Researchers identified limitations from having limited choices because for example, fear could’ve been more identifiable with surprise compared to their choices of sadness, happiness, anger and neutral. They also noted that because this study was done with adult raters, it should be tested with children raters to compare their results and level of agreement to the adults’. The researchers data support that the set used in their study is just as valid as other face stimuli sets from previous research used in psychology. From their finding that the pictures validly represent the emotion types, they can be used for further children rating but future research should also incorporate other racial groups to reduce cultural bias.

Longitudinal dimensionality of adolescent psychopathology: testing the differentiation hypothesis.

The differentiation hypothesis states that as children develop, their abilities and functioning become more differentiated in which they develop more specialized skills. Because this hypothesis has not been tested longitudinally, researchers were interested in looking at psychiatric symptoms and the developmental differentiation of psychopathology among adolescents. Previous studies looked at DSM-IV syndrome differentiation between-children, across-age and between-study samples but not within-child. This study examined the syndromes of psychopathology, syndrome differences and syndrome dimensionality of common Axis 1 Syndromes that include, major depression (MDD), generalized anxiety disorder (GAD), separation anxiety (SAD). social anxiety (SOC), attention deficient hyperactivity disorder (ADHD), conduct disorder (CD) and oppositional defiant disorder (ODD) across adolescents between age 9 to 16 in a longitudinal sample to test and better understand the differentiation hypothesis.
Children and their parents were interviewed using the Child and Adolescent Psychiatric Assessment (CAPA) which asked about the onset, intensity, frequency and duration of symptoms across 3 months. A computer algorithm then determined where the symptoms the child had met the standards for psychopathology according to the DSM-IV. In creating several models for statistical analysis, they looked at the results for within each age-group. They found that symptoms were shared across syndromes and there were some overlapping symptoms that correlated to each other. For example, fatigue symptoms for GAD and MDD were correlated. Researchers performed a dimensionality testing for syndromes from creating model comparisons. Models included within-age blocks for age 9-10, 11-13, and 14-16 for MDD/GAD, SAD, ODD, CD, SOC, Hyperactivity, Impulsivity, Inattention. An Across-block was also created to test the differentiation hypothesis. Overall for across-age, they found that an average proportion of variance in the symptoms were mostly constant across-ages for most of the syndromes but specifically, there was an increase for MDD/GAD at age 14-16 for symptoms.
From analyzing their results, they found stable dimensionality from pre-adolescence to later adolescence for the syndromes but an increase for MDD and GAD symptoms at late adolescence. In comparing emotional (i.e., Social Anxiety, MDD, GAD) and disruptive syndromes (i.e., CDD, ODD), they found that some emotional syndromes such as SOC, differentiated from disruptive syndromes but SAD were significantly less differentiated, thus the correlation between emotional and syndromes slightly differentiated. With MDD and GAD showing relatively more variance in symptoms for 14-16, researchers concluded that more studies needed to compare the dimensionality between the two for adolescence and adulthood. Additionally, results support that mild differentiation may occur before adolescence for disruptive syndromes but during adolescence, it is more so for GAD and MDD, as well as for some emotional and disruptive syndromes.