Learning about the Developing Brain (LADB)

This study is closed to enrollment.

To view Final Progress Report, Click Here.

Purpose of the study

This application extends to early childhood recent work with adults and adolescents, as well as with rodents and primates, suggesting that social phobia (SoPh), separation anxiety disorder (SAD), and generalized anxiety disorder (GAD), while often comorbid, represent biologically distinct psychiatric phenotypes linked to distinct patterns of neural activity. We propose the first longitudinal fMRI and eye-tracking study of pediatric SoPh, SAD, and GAD. This study builds on our community-based study, “Preschool Anxiety Disorders in Primary Care” (PADPC; R01-MH075766-01A1; PI: Angold; co-PI: Egger, IRB #8713). Capitalizing on the intensive battery of already collected information on diagnosis, temperament, and risk factors, it extends the assessments to the realm of neuroscience. We will recruit 180-230 children from the Preschool Anxiety Disorders in Primary Care Study (PADPC), two thirds with SAD, GAD, and/or SoPh at baseline; and one third typically developing children. The children’s ages will range between 4 and 7 years old at the time of selection (120 recruited in YR 1 and 60-110 in YR 2). Children will complete eye scans at each lab visit. Children will also be invited to participate in the fMRI phase of the study as well. Our broad aims are: (1) Identification of dysfunctions of the neural circuitry involved in emotion processing in children with preschool SAD, SoPh, and GAD (2) Examination of the mechanisms underlying the persistence and development of functional abnormalities in the neural circuitry involved in emotion processing in children with preschool SoPh, GAD, and SAD.

Our hypotheses are:

Hypothesis 1. Preschool social phobia, specifically, results from amygdala and vPFC hypersensitivity to social threat.

Hypothesis 2. Preschool SAD, specifically, results from hypersensitivity in the amygdala, vPFC and the nucleus accumbens to separation threat cues.

Hypothesis 3. Preschool GAD is more similar to later childhood depression than to an anxiety disorder.  Both GAD and depression result from dysfunctions in neural circuits underlying emotion regulation and evaluation of stimulus salience including (1) hyposensitivity to positive social and non-social stimuli and (2) hypersensitivity to negative social and non-social stimuli.

Hypothesis 4. Children who meet criteria for more than one of these three anxiety disorders are expected to manifest the atypical fMRI and scan patterns characteristic of all the disorders they have.

Background and significance

SoPh, SAD and GAD are among the most common childhood psychiatric disorders, carrying substantial risks for continuing morbidity and impaired quality of life throughout the lifespan. SoPh, SAD, and GAD represent meaningfully discriminable anxiety phenotypes. These phenotypes represent functional disturbances in responses to emotionally evocative, potentially threatening stimuli discriminated along three dimensions – social vs. nonsocial, familiar vs. unfamiliar, and emotional valence. These disturbances have been linked to dysfunctions in attention, appraisal, and dysfunctions in the fear circuit, as well as associated circuits mediating face-emotion recognition and emotion regulation. This application offers the following advances: (1) No study, at any age, has examined the specificity of effects in these regions across stimulus configurations designed to address these three dimensions in the same individuals. (2) Despite the fact that comorbidity is widely recognized as one of the central problems for etiological research in psychiatry, it has typically been ignored in imaging studies. (3) There have been few imaging studies of pre-adult anxiety disorders, and none of those few has considered children under the age of 8. (4) There have been no longitudinal neurofunctional studies of anxiety disorders in childhood or adolescence whatsoever.

Defining the similarities and differences in the neural substrates and circuitry of preschool SoPh, SAD, and GAD will have enormous implications for (1) validation of these early-onset disorders, (2) the question of where to draw the line between anxious temperamental variation and clinically significant anxiety phenotypes in young children, (3) the development of treatments and preventive interventions for young children with anxiety disorders, and (4) our understanding of the etiology and developmental course of specific anxiety disorders.

Design and procedures

This is a longitudinal imaging and eye tracking follow-up study of a portion of a representative sample of preschool age children and their families who were recruited to participate in a current NIMH funded, cross-sectional study of preschool anxiety disorders (“Preschool Anxiety Disorders in Primary Care (PADPC)”; R01-MH075766-01A1; PI: Adrian Angold; co-PI: Helen Egger MD). In the initial design, we planned to   recruit 180 children who were 2-5 years old at the baseline PADPC assessment (120 with SAD, GAD, and/or SoPh at baseline; 60 typically developing children) to participate in annual assessments. In this IRB amendment, we are requesting permission to recruit an additional 50 subjects because (1) we have the resources to do so and (2) need to balance our age and gender sub-groups. There are two phases of this proposed study: the eye scan phase and the fMRI phase. In the eye scan phase, the child will complete an eye scan paradigm, as well as psychiatric, physiological, and psychosocial assessments. Parents will also complete assessments about their child’s behaviors and emotions. Children will also be recruited to participate in the fMRI phase of the study. We expect 50% of the children who participate in the eye scan phase to take part in the fMRI phase. Children who do not participate in the fMRI phase of the study will still participate in the annual eye scan phase of the study. The visual stimuli for the eye tracking and fMRI will include faces with different emotional expressions, pictures of the child and mother and toys, and scenes, with and without people, with different emotional valence.

Assessment procedures

The eye tracking will be conducted at the Duke Young Child Laboratory which is located at our Center for Developmental Epidemiology (CDE) offices in Brightleaf Square, Durham, North Carolina. After the CDE eyetracking lab we will invite the children to participate in the fMRI phase of the study. For the children participating in the fMRI phase of the study, a mock scan will be conducted at the CDE and eye scanning and the fMRI will be conducted at the Duke University Brain Imaging and Analysis Center (BIAC) in the Duke Children’s Hospital. PAPA interviews will be conducted in the parents’ homes, unless the parent prefers to be interviewed at the CDE lab.

At the end of the CDE session, we will provide the parent with information about the fMRI phase of the study and ask for consent for the child to participate. If the parent and the child are willing to try the fMRI phase of the study, we will conduct a mock scan at our CDE lab.  We will teach the child how to respond to “yes/no” questions without nodding or shaking his or her head. The child will practice staying still with the interviewer. We will collect physiological measures of sympathetic nervous system (heart rate variability, galvanic skin response, and alpha-amylase) and hypothalamic pituitary adrenal axis activity (cortisol) from our subjects. These measures and assays will enable us to quantify physiological anxiety before and after mock scanner desensitization and from the mock scan to the actual scan. Heart rate and GSR will be recorded continuously during the mock scan protocol using the electrodermal activity and electrocardiogram modules of the Biopac MP150 system. Heart rate and GSR are already being collected in the actual MRI scan using a Biopac system. Children will provide four saliva samples for the cortisol and alpha-amylase assays at the mock scan session and at the actual MRI session (total 8 samples per child): (1) upon arrival at the lab (or MRI suite), (2) immediately before the simulation scan (3) immediately after the mock scan or actual MRI and (4) and 20 minutes after the scan. At each point of saliva collection, the child’s current level of anxiety/fear will assessed using a fear thermometer presented as a visual analog scale ranging from 0 (not afraid) to 10 (most afraid I have ever been). Autonomic symptoms of anxiety will also be assessed at each point of saliva collection, using a subset of questions extracted from the Spence Children’s Anxiety Scale (SCAS), a 44-item child self-report measure. The child will color in the thermometer to show his/her level of fear/anxiety at that moment. We will send the child home with a CD of the MRI sounds to listen to prior to the MRI session and a coloring book about the MRI procedures and crayons.

CDE session. At the CDE assessment, the parent will complete psychiatric and psychosocial measures, all self-report. Next, the parent will complete an eye scan and then the child will complete an eye scan. The eye tracking system consists of a computer and a computer screen with a camera that tracks eye movements using infrared beam tracking. The screen that the child and parent watch the stimuli on has a camera built into the monitor that uses infrared beam tracking to record the parent and child’s eye movements. When the children are six years old, they will also complete the Differential Abilities Scale (DAS), a comprehensive cognitive assessment that takes about 40 minutes to complete.

We will also conduct a brief 20 minute assessment of emotion regulation. We will show the children three 2-minute movie clips (neutral, happy, and sad) while we measure eye scan paths and take physiological measurements of heart rate, stomach motility, and skin perspiration using a physiological monitoring system (BIOPAC Systems, Santa Barbara, CA). All BIOPAC Systems data acquisition and analysis systems are FDA approved and compliant with all appropriate and relevant European and worldwide standards for medical safety (Including: IEC601-1 Medical Safety Test Standards, 73/23/EEC Low Voltage Directive, 89/336/EEC Electromagnetic Compatibility Directive). Galvanic skin response (GSR), a measure of arousal as indexed by skin perspiration will be monitored from disposable silver-silver chloride electrode disks attached to the middle phalanges of the 1st and 2nd digits of the non-dominant hand. Heart rate and heart rate variability will be measuring using electrocardiogram via disposable electrodes placed above the left and right ankles and left and right wrists. Stomach motility will be measured via electrogastrogram with electrodes placed above the umbilicus and below the left rib.

All Biopac data acquisition and analysis systems are fully tested and compliant with all appropriate and relevant European and worldwide standards for medical safety (Including: IEC601-1 Medical Safety Test Standards, 73/23/EEC Low Voltage Directive, 89/336/EEC Electromagnetic Compatibility Directive). The BIOPAC monitoring system is currently being implemented in psychophysiological laboratories throughout Duke University including the DIISP Lab in the Duke Social Science Research Institute and has been used previously in IRB approved protocols including those from our laboratory. This system is used clinically for both assessment and treatment such as biofeedback. We will be conducting concurrent eye tracking while they are watching the movies. After watching each movie clip, the children will be asked about the emotions they felt during the clip.

In-home session. An interviewer will visit the parent at home or in our CDE offices to administer the Preschool Age Psychiatric Assessment (PAPA) to the parent when the child is 6 years old. A computerized version of the PAPA (ePAPA) will be administered using a tablet computer. The interview will take about 1-3 hours to complete.

BIAC session. At the BIAC children will complete the same eye tracking study that was conducted at the CDE, as well as a mock scan and the actual fMRI scan. The visit will last about an hour and half. An fMRI assessment will be conducted annually with the children.

Desensitization to MRI Procedures. Prior to the first fMRI scan, parents will be sent audiovisual materials that provide prior exposure/desensitization to visual and auditory stimuli that are encountered during a scan. As noted above, the child will participate in a play protocol to familiarize him or herself to the MRI procedures and experience. On the day of the scan, the child will participate in a mock scan in our simulation scanner which reproduces the sights and sounds of the scanning environment. The MRI simulator looks, sounds, and feels just like a real MRI scanner. During the entire actual MRI scan, an experienced interviewer will be in the scanner room to comfort the child, answer questions, and/or remove the child quickly from the scanner if he or she becomes too upset to proceed.

If the subject is a female, 8 years old or older, we will ask the parent if the child is menstruating.  If so the BIAC will administer a urine pregnancy test. The subject will pee into a small container that will be given to the technician to test. The pregnancy kits will be provided by the BIAC and are approved by the Chair of Obstetrics and Gynecology.  BIAC will keep records of the results of each pregnancy test in writing, as will we.

Selection of subjects including inclusion/exclusion criteria and how subjects will be identified

We will recruit two cohorts of children (120 children Year 1; 60 Year 2) for a total of 180 subjects. The children’s ages will range between 4 and 8 years old at the time of selection. We are planning to recruit additional children (up to 50). The main selection criterion for this study will be whether or not the child met criteria for SoPh, SAD, or GAD when they participated in the baseline assessment of the PADPC. Each cohort will be stratified by age (equal numbers of 4, 5, 6 year olds) and gender. The total sample (N=180) will included 120 anxious children (ANX) and 60 controls (CTRL) children. We anticipate that 58% of the sample will be African American (N=104) so our study cohort will somewhat over-represent the African-American community in our area.

Inclusion/exclusion criteria for the eye scan phase of the proposed study

Inclusion criteria are (1) children participating in the nested case-control study of the PADPC who (2) were between 24 and 71 months old at the time of recruitment in to the PADPC and are, at the time of recruitment, 48-107 months old and (2) either (a) meet symptom criteria for SAD, GAD, or SoPh based on the PAPA or (b) do not meet criteria for any psychiatric disorder based on the PAPA.

Exclusion criteria are those listed above for the PADPC: (1) the lack of a parent with adequate fluency in English to complete the interview, and (2) the index child is known to have mental retardation (IQ < 70), autism or other pervasive developmental disorders, or has sibling already participating in the study.

Inclusion/exclusion criteria for the fMRI phase of the proposed study

Inclusion criteria is the child must be more than 48 months old (i.e. aged 4 or older) when recruited. Exclusion criteria are a child with (1) a cardiac pacemaker, dental braces, aneurysm clip, cochlear implants, shrapnel, neurostimulators, defibrillator, artificial heart valve, or history of metal fragments in eyes, (2) a body weight of 250 lbs. or more, and/or (3) claustrophobia (4) a pregnant female subject. Subjects who are too large to fit comfortably in the scanner (highly unlikely in young children) or who have a history of claustrophobia will still be considered for the behavioral laboratory studies. Sedation will not be used for the fMRI scans.

Subject recruitment and compensation


After sending an IRB approved letter describing the study, our project coordinator will contact eligible PADPC study participants by telephone, as close as possible to the one year anniversary of their baseline assessment, and invite them to take part in this follow-up study. After scheduling an appointment over the phone, we will send printed material that explains the study and procedures, confirms the appointment, and provides directions and study contact information to the parent.


The parents and children will be compensated for participating in the study. The child’s parent will receive $40.00 for completion of the CDE lab assessment and $75 for the in-home PAPA interview ($40.00 per lab assessment session). They will receive $75 for completion of the assessment at the BIAC. Children will receive gift bags with developmentally appropriate toys to thank them for their participation at the end of each session. We will provide travel by cab to and from the lab sites if the parent cannot arrange transportation. We pay for parking at both sites.

Consent process

Parental/guardian consent will be obtained for all participants annually. In addition, each participant 6 or older will give informed assent to participate. The child’s parent/guardian must be present while the staff member is attempting to obtain informed assent from the child. The staff member who has obtained consent/assent must certify that he has carefully explained the purpose and nature of the research to the participants in appropriate language and answered any questions. If a parent wishes to allow his/her child to participate, and the child provides assent, the parent will sign IRB approved consent forms. They will be informed that they can refuse to answer any questions they choose. All consent forms will be signed in duplicate with the parent(s) receiving one original and one being kept in a locked office at the CDE. All retained copies will be kept in a locked cabinet in a locked room. Participants will be informed that participation in the study is voluntary, that they may decline to participate in the study and that this will not jeopardize care, and that they may withdraw from the study at any time.

Subjects capacity to give legally effective consent

We will not include parents who are not competent to provide consent. The parent giving consent must have legal guardianship of the child. Most of the children are too young to provide assent to participate. Nonetheless, if the child objects to participating in any part of the protocol or becomes distressed, we will stop the assessment (see above).

Study interventions

This is an observational study, not an intervention study. The study does not include clinical assessment, treatment recommendations, or treatment.

Risk benefit assessment

Risk/benefit ratio

As described below, risks of the project studies are minimal. Therefore, the benefits of participation in the proposed studies for children with and without anxiety disorders outweigh the potential risks.


The anticipated risks of participation are minimal to none but are described below.

Psychiatric, psychosocial, and cognitive assessments. The risks of paper and pencil or computer-based assessments are minimal and are limited to performance anxiety or fatigue. The parents may experience some psychological discomfort talking to the interviewer about some aspects of her life or her child’s life. The child might find the assessments frustrating, disappointing, scary or boring. At any time, the parent and/or the child may discontinue the interview/assessment or choose not to talk about a certain topic. The interviewers are also trained in procedures for cases where a participant indicates suicidal or homicidal thoughts or plans, or when there is evidence of or concern about child abuse or neglect. 

Eye tracking. There are no known risks to the use of infrared oculography. The eye tracker uses an infrared beam to track the position of the pupil and cornea over time in order to estimate position of gaze. The exposure to near-infrared radiation for subjects in these experiments is about 0.1 to 0.3 mW/cm2 (ASL product specifications).

fMRI studies. There are no known risks to the use of MRI per se; however, there are four areas of concern. The first is the potential risk of the main magnetic field attracting ferromagnetic objects toward the magnet. The second is the discomfort some subjects encounter by the confinement within the bore of the MRI system. The third is the loud noise made by the gradients during imaging. The fourth is finding atypical structural findings during the fMRI exam. These risks occur for all clinical MRI exams and are not increased by the proposed research.

Physiological assessments. There are no known risks associated with physiological assessments. Electrodes (disposable adhesive electrodes) are positioned on the fingers, stomach, and ankle for the purpose of recording the electrical activity of the heart, stomach, and skin. Thus electrodes are only mechanisms for collection.

Potential social risks. There is minimal risk of the revelation of confidential information including diagnostic information, historical information, and intellectual functioning levels. Interviewers are trained not to talk to anyone about their interviews and sign an undertaking to that effect. All data that would lead to subject identification will be labeled using a research code number only so that no names will be used in any of the data. Confidentiality of all subject information will be preserved by housing all records in locked files in locked offices. Data will be maintained on a computer accessible only to research project staff. Further, all computerized data will be maintained in a limited access, password-protected hard drive. 

Potential legal risks. There is a risk that study data might be subpoenaed for legal purposes. If funded, we will obtain a Certificate of Confidentiality from NIH. In addition, except under circumstances covered under the mandated child abuse reporting laws, and/or situations in which the child and/or a caregiver is judged clinically to be a danger to themselves or others, no information about the child or family will be shared with any individual or agency without prior written consent.

Protection against risk

We make every effort to make the visit to the CDE lab as fun and supportive as possible. The setting is cheerful and child-friendly. Between tests and procedures, children will have a great deal of freedom, attention, and a wide range of appealing activities from which they can select (e.g. to play/build with Legos, play games, or draw). Both the parent and the child will be given a healthy snack. The length of the test sessions will be adjusted to the needs of each child.

Anxiety associated with the fMRI procedures will also be reduced by allowing the child to practice the procedures in the mock scanner before the actual procedure. The mock scanner includes a recording of the noise associated with imaging and the capacity to practice versions of the study procedures before the actual fMRI. Sedation will not be used and the subject or his/her parent may abort the examination at any time.

Neuroimaging safety

There are no known biological risks to exposure to magnetic fields from MRI exams using the techniques in this study. All MRI studies will follow guidelines set by the FDA with regard to specific absorption ratio (SAR), limits on gradient slew rate (dB/dt), and noise. The fMRI studies involve a 3.0 Tesla MR scanner that conforms to FDA safety guidelines. All scanner adaptations such as echo planar imaging head coils are FDA approved. The presence of a metal foreign object implanted in the subject will be determined by the screening procedure, orienting session, interview with each potential subject and parent(s), and completion of a detailed screening questionnaire. Subjects and staff will be instructed to remove all metal objects, including clothing with metal clasps, before entering the magnet room. Some subjects may feel uncomfortable or confined once positioned within the bore of the MRI system. If subjects feel uncomfortable, the imaging procedure will be terminated and the subject will be removed from the magnet.


The anticipated benefit of this research is a better understanding of the characteristics and development of the neural substrates and genetics of specific types of early onset anxiety symptoms and disorders in preschool children. Many subjects in our other studies have also told us that they enjoy being part of a study and are proud of their contribution. There is no direct benefit from the research for the child or the parent beyond the contribution to a general understanding of mental health symptoms and disorders experienced by young children and their families. Compensation for the parents and children who participate in the study is described in section 5 above.

Costs to the subjects and compensation

Subjects will incur no costs at as result of participating in this study. This is not a treatment study. The parents and children will be compensated for participating in the study. For compensation, see section 5.

Data analysis and statistical considerations

Our overall approach to data analysis relies on the framework provided by the class of mixed (random and fixed) effects models referred to as hierarchical linear models (HLM). HLM models were developed to deal with clustering of observations within groups of subjects (for instance, of children within schools), and within-subject clustering of observations (for instance, in longitudinal studies, or studies with repeated stimulus presentations). HLM provide the means to deal uniformly with methods typically used in ROI fMRI analysis, voxel-based whole-brain fMRI analysis, and longitudinal data analysis. In each of these cases the differences in approach to analysis concern the ways in which non-independence of observations is dealt with.

In the proposed study, we take the ROI approach to testing our hypotheses, but we have the added complication of having multiple scans from the majority of subjects. In other words, we will have within-subjects clustering of scan data. This is easily dealt with in HLM by treating subject identity as a random effect, while continuing to average across all lower nested levels (resulting in analysis at the level of repeated scans (cyan in figure D.3). The simplest form of analysis here will treat the within-subject correlations between scan occasions as a nuisance factor, providing what is essentially an ANOVA (or multiple regression) with correction for the repeated measures. However, within-subject slopes (i.e., change over time) can also be modeled to determine whether different groups are changing in different ways (growth curve analysis). Such analyses can be performed in many packages, including SAS PROC MIXED (or with the SAS GLIMMIX macro if the outcomes are non-normal), Stata, and HLM. Latent growth curve analysis is a further extension that seeks to detect classes of individuals with different patterns of change over time empirically (such analyses can be performed using PROC TRAJ in SAS, or GLAMM in Stata).

Data and safety monitoring

The sources of research material collected from subjects who participate in this project will be the interviews with parents, questionnaires filled out by parents, diagnostic measures, cognitive assessment, eye tracking, electrophysiological data, and fMRI scans. The results from these measures will all be translated into written records, numerical data, or imaging data that will be stored electronically.

We intend to share de-identified data with New York University, specifically Dr. Barbara Coffey and Dr. Michael Milham. This is part of a collaborative effort with other institutions to conduct secondary analyses of previously collected data.  The New York University group will gather data on resting state fMRI data sets with accompanying MPRAGE images to spatially normalize functional data. They will be using this data to measure levels of functional connectivity between brain regions during rest by applying commonly used resting state metrics and looking at EEG amplitude of low frequency oscillations.  All the data will be fully de-identified and anonymized. Researchers will not be given or have access to any subject identifiers and will not be able to readily identify the subjects.

The minimal risks to the subjects are described in the previous section. Dr. Egger (PI) and Dr. Angold (CI) are both licensed child psychiatrists who provide medical oversight for the study and can respond to clinical needs as they arise.

Data storage and confidentiality

Each subject will have been assigned a unique numeric identifier as a PADPC participant (PAPAID #). When data are transferred to data summary sheets, only this subject number will appear on the sheet and be entered into the database. All electronic data will be double password protected and complies with Duke University IRB regulations and all HIPPA regulations. Care is taken to conceal other possible sources of identification, including addresses. Lists of participant numbers, names, and other demographic information, as well as consent forms and MRI screening forms, are kept in a locked cabinet located in a locked room at the CDE. Cross-referenced names and code numbers are included in a master list that is stored in a locked file cabinet that will be accessible only to the principal investigators and research staff. The identity of participants or any family members will not be revealed in any description or publications resulting from this research. Research records are only accessible to research investigators and research staff. All staff sign confidentiality agreements to work on our studies. MRI data will be destroyed after six years or at the conclusion of the study. Written reports and papers will not identify subjects or make identification possible without the written consent of the person concerned. Our data storage plans have been developed to be fully compliant with HIPPA safeguards. All research participants will receive a copy of Duke’s privacy policy, which they will sign. They will receive a copy for their records and the original will be kept on file for six years.

Procedures to preserve confidentiality

All information gathered for the purposes of the research projects will be treated confidentially, in compliance with the Duke University Institutional Review Board policies and consistent with the rights of all participants to privacy. All faculty and staff have taken and passed Duke’s required training modules on HIPAA regulations for protecting participants’ privacy. The Center for Developmental Epidemiology (CDE), where the study records are held, observes all approved security rules to protect the data. The CDE also has a full-time staff member, our data and information manager, Jurgen Henn, who oversees our group’s HIPAA compliance. If any new information becomes available to the investigators during the course of the research that may affect the parent’s willingness to continue to allow participation, each subject’s parent(s) will be notified. Additionally, if any new adverse effects of the procedures are demonstrated in the future, every attempt will be made to notify parents about this.

Secure research files (locked cabinets in locked rooms) will be used to store assessment and data forms derived from the research data collection. Confidentiality on computer disks used for analysis will be carefully safeguarded by participant coding, with each participant assigned a numeric code to conceal identity.