1. A Multisite Neurobehavioral Assessment of FASD
Sarah Mattson, PI
CIFASD Clinical Group PI Meeting
January 17-18, 2008

2. Key Personnel Colleen Adnams, Co-PI, subcontract
Claire D. Coles, PI, subcontract
Julie A. Kable, Co-I, subcontract
Wendy Kalberg, Co-PI, subcontract
Philip A. May, PI, subcontract
Edward P. Riley, Co-PI
Elizabeth R. Sowell, PI, subcontract

3. Objectives
To determine whether a neurobehavioral phenotype exists in children with fetal alcohol syndrome
Whether the same phenotype exists in children with FASD who lack facial dysmorphology
Whether the phenotype can be used for differential diagnosis
Secondary aims, involving collaboration with other CIFASD projects and cores, are to determine the relationship between brain dysmorphology, facial dysmorphology, and neurobehavioral function.

4. Methods
A standard neurobehavioral protocol will be administered to four groups of children at six sites
executive function, working memory, verbal function, and psychological symptomatology.
In addition to children with FASD and non-exposed controls, children with low IQ scores or ADHD will be included as contrast samples.
Using this heterogeneous sample and multivariate statistical methods, neurobehavioral profile specific to FASD will be sought.
participants will also be assessed using methodology prescribed by the Dysmorphology Core and the facial and brain imaging projects of the CIFASD. Data from three broad domains (neurobehavior, dysmorphology, and brain morphology and function) will be analyzed both separately and together to address the main aim of the CIFASD: improving the diagnostic criteria for FASD.

5. Functions Previously Associated with the Neurobehavioral Core
Preparing all testing materials
Purchasing tests and test materials
Creating and distributing administration manuals, data collection forms, summary forms, scoring manuals
Ensuring reliability of data collection
Conducting two training sessions during the first year with follow- up sessions each year for reliability purposes
Collaborating with the Informatics Core on creating the data dictionary for the input tool for the current data and on pilot testing of the input and upload tools created for this project

6. Relationships with Other Projects/Investigators
Brain Imaging
San Diego, Los Angeles, South Africa
Facial Imaging
San Diego, Los Angeles, Atlanta, South Africa
Dysmorphology
All sites
The proposed project directly relates to the consortium goals of Enhanced understanding of the neurobehavioral phenotype of FAS and ARND and Establishing standardized diagnostic criteria and methods of assessing FASD. To this end, it demonstrates integration with other projects including those entitled “Mapping the Brain, the Face and Neurocognitive Function in FASD” (E. Sowell, PI) and “3D Facial Imaging in FASD” (T. Foroud, PI). In addition, support from the Dysmorphology (K. Jones, PI) and Informatics (C. Stewart, PI) Cores are integral to this project. The Dysmorphology Core will provide physical examinations required for diagnosis and grouping and the Informatics Core will provide data support. During the last funding period, collaborations with these Cores were very successful. The Dysmorphology Core examined 744 cases across the sites and the Informatics Core created a database consisting of more than 800 neurobehavioral variables that was used successfully across sites.
All collaborations have resulted in published papers.

7. Participating Sites
Center for Behavioral Teratology, San Diego State University, San Diego, CA
Marcus Institute, a Division of Kennedy-Krieger Institute at Emory University, Atlanta, GA
University of New Mexico, Albuquerque, NM
Seven Northern Plains communities, including six Indian reservations
University of Cape Town, South Africa
the University of California, Los Angeles, Los Angeles, CA
The combination of these sites allows for a study population that is both large in number and heterogeneous in nature. Such a sample will ensure our results are both accurate and unbiased by specific site characteristics.

8. Sample Sizes Site FAS/D CON IQ ADHD San Diego 50 25 Los Angeles 25-35
Atlanta
25-50
New Mexico 80 40
N. Plains
S. Africa
100
Total
360
130
235
Table 2. Proposed sample sizes for each site in the
four participant groups
Site
FAS/FASD
CON IQ
ADHD
San Diego 50 25
Los Angeles
25-35
Atlanta
25-50
New Mexico 80 40
Northern Plains
South Africa
100
Total
360
130
235

9. Progress of Previous Award
Two studies involving all sites are in Manuscript phase
Virtual water maze - being revised
Neurobehavioral Profile - in preparation
Three imaging studies published
Two RSA 2008 Abstracts
A. Previous Award. The progress report from the renewal application, which details our progress during the previous funding period is included in Appendix 1. Briefly, two studies are in manuscript phase, including analysis of the virtual water maze task and the latent profile analysis of the CIFASD neurobehavioral data. The manuscript for the watermaze study was not accepted by Neuropsychologia and we are reanalyzing and revising the manuscript for another journal. The latent profile analysis was presented at RSA in 2007 and we are preparing the manuscript for publication. We have also published three papers documenting results from imaging studies, and these are detailed in the appendix.

10. Center for Behavioral Teratology, San Diego State University
Neurobehavioral Profile of Children with Heavy Prenatal Alcohol Exposure
S.N. Mattson, S.C. Roesch, E.P. Riley, C. Adnams, I. Autti-Rämö, Å. Fagerlund, W. Kalberg, M. Korkman, P.A. May, and the CIFASD
Center for Behavioral Teratology, San Diego State University
Research Supported by NIAAA Grants U01AA014834, U01AA014786, U24AA014830, U24AA014811
Background and General Method
Results
Table 2. Distribution of subjects in the alcohol-exposed (ALC) and control (CON) groups in the two classes resulting from the latent profile analysis.
The Collaborative Initiative on Fetal Alcohol Spectrum Disorders (CIFASD) is an international multi-site project developed, in part, to define a behavioral phenotype for prenatal alcohol exposure. A broad, predominantly nonverbal battery was selected to assess children with heavy prenatal alcohol exposure while reducing the influence of differences in language and culture between sites. Data were collected from four sites in three countries: United States (San Diego, CA and Northern Plains Indians), South Africa (Cape Town), and Finland (Helsinki). Children with heavy prenatal alcohol exposure (ALC) and controls with minimal or no prenatal alcohol exposure (CON) were tested in their native language by a trained psychometrist or psychologist on a battery of neuropsychological tests developed by the CIFASD Neurobehavioral Core working group.
A 2-class solution had the best fit relative to other models tested according to the Lo-Mendell Rubin Test (p = .018).
Class 1: 78 subjects (46.4%)
Class 2: 90 subjects (53.6%)
Subjects in Class 2 performed better than subjects in Class 1. See Table 1.
Chi-square analysis indicated exposure group and class were related (p < .001). See Table 2.
Class membership was not related to FAS diagnosis (p > .05). See Table 3.
Results were replicated in the larger sample (N=309) estimating missing values.
Measures that discriminated groups were tests of general intellectual function (Leiter), executive function (DKEFS Fluency & Trails, PPT), interhemispheric transfer (Finger Localization), short-term memory and attention (CANTAB Spatial Span), and fine motor speed (Grooved Pegboard). See Table 1 for conditional response means and Figure 1 for graphical representation.
Measures that did not discriminate groups were tests of basic motor, rule learning, object and spatial memory (CANTAB Motor Screening, Big/Little Circle, Pattern Recognition Memory, Spatial Recognition Memory, Spatial Working Memory), reversal learning (Visual Discrimination Learning), and attention (NES-3).
Table 3. Distribution of subjects in the FAS and ARND subgroups in the two classes resulting from the latent profile analysis. Diagnosis was missing for 5 subjects.
Measures
The Neurobehavioral Core test battery consisted of 16 neurocognitive tests and 6 tests of psychological symptomatology. The tests were selected to probe specific deficits that had been established in previous human and animal studies, and included measures of arousal and attention (NES-3 Continuous Performance Test, Reversal Learning Test); executive function (Progressive Planning Test, D-KEFS Verbal Fluency, D-KEFS Trail Making); general intellectual ability (Leiter-R); memory (CANTAB Spatial Recognition Memory, CANTAB Pattern Recognition Memory); spatial learning (Virtual Morris Water Maze); motor function (Grooved Pegboard Test, CANTAB Motor Screening); working memory (CANTAB Spatial Working Memory); interhemispheric transfer (Finger Localization Test); short-term memory/attention (CANTAB Spatial Span); concentration (Leiter Cancellation Test); visual-motor integration (VMI); and psychological symptomatology (Sluggish Cognitive Tempo, Disruptive Behavior Rating Scale, Achenbach Child Behavior Checklist, Achenbach Teacher Report Form, Achenbach Youth Self-Report Form, Pictoral Depression Scale). Only the neurocognitive tests were included in the current study.
Figure 1. Standardized mean differences based on the conditional response means (Table 1) for the neurocognitive variables included in the latent profile analysis. Standardized mean differences can be interpreted as effect sizes (d).
Table 1. Conditional response means and standard deviation for neurocognitive variables included in the latent profile analysis. Only variables with medium or greater effect sizes were included in this analysis.
Statistical Analyses
Neurocognitive data were analyzed using latent profile analysis (LPA), which is a statistical method for finding subgroups of related cases. It is a person-centered approach that groups individuals into classes or profiles based on the similarity of individuals’ response patterns. Thus, LPA derives classes of individuals who are similar to individuals within their own class but differ from individuals in other classes.
Based on one-way ANOVAs on all neurocognitive variables, variables that had medium effect sizes as a function of group (ALC vs. CON) were submitted to LPA.
168 subjects (94 ALC, 74 CON) had data on all 25 variables.
Conclusions
These results reinforce our ability to conduct multi-site, international research and suggest that neurobehavioral tests can be used to define a clinical profile of children affected by heavy prenatal alcohol exposure. Further, these data point to specific neurobehavioral domains as particularly sensitive to this exposure.
Future research should continue to build on this profile by including a broader array of neurocognitive tests and diverse samples.
For correspondence:

12. Laboratory Validation of Parent Reports of Behavior
Parent reports indicate inattention and hyperactivity, however laboratory validation of these deficits are unavailable
Children in the CIFASD study were assessed using parent reports of inattention and hyperactivity (DBD). Validation of these domains were sought using Wrist-worn actigraphs and CPT (NES3).
Only SD Subjects were included (N=67)
Groups were similar on age, ses, handedness, race, ethnicity, and sex
The ALC group had lower FSIQ scores groups
All testing began between 8:50-9:45AM

13. Parent Report of Inattention and Hyperactivity

14. Laboratory Measure of Inattention

15. Laboratory Measure of Activity Levels

16. Objective Measure of Activity Levels
Wrist-worn actigraphs were used to measure activity levels of 75 children who participated in CIFASD testing in San Diego
Three groups were included: ALC, CON, ADHD
Groups were similar on age, SES, handedness, race, ethnicity, and sex
The ALC group had lower FSIQ scores than the other two groups
Activity was measured during the first 5 hours of neuropsychological testing
All testing began between 8:50-9:45AM

17. Laboratory Measure of Activity Levels

18. Phil’s slides

19. Neuropsychology Summary: Plains Indian Data
We analyzed the results of the first round of CIFASD neuropsychology tests of the battery for Plains Indians (2005 – 2007).
Interest was in examining trends that speak to the neurobehavioral profile of FASD children.
Also we were interested in whether the complex tests/tasks were more discriminating of the more severely affected FASD children (FAS and PFAS).

26. Progress of New Award
Conference Calls. Monthly conference calls have been held with PIs and key Personnel.
IRB Approvals. We submitted applications for IRB approval for the neuropsychological, dysmorphology, brain imaging, and 3D imaging portions of the study at each site. Approvals were granted at the SDSU and Atlanta sites and are pending at all other sites.
Subcontracts. All three subcontracts have been submitted to the sites and we are awaiting their return.

27. Progress of New Award
Hiring. We have in place staff necessary to conduct the study, including a psychometrist, recruiter, and research associates/assistants. The subcontract sites have personnel in place, with the exception of Atlanta and South Africa which are waiting for funds to be released to hire personnel.
Purchasing. We have purchased materials and equipment necessary to begin data collection at each site. Materials are being distributed.

28. Progress of New Award
Material Development. We have created working drafts of our test administration materials and scoring materials.
Database Development. We have been working with the informatics core to develop the input tool for the new neurobehavioral test battery (CIFASD Neurobehavioral Phase II). We have received and piloted the beta version of this input tool and are actively working with them to move it towards its final form.

29. Progress of New Award Training #1: December 3-5, 2007
In attendance were data collection staff from University of New Mexico and UCLA.
Participants were trained on data collection procedures and received materials for pilot testing and instructions for reliability procedures.
Training #2: January 14-15, 2008
All sites were represented for training on scoring, data entry, input tool, and reliability procedures.
The goal of training is to ensure that all sites are collecting data in the same manner, and are reliable with each other.

30. Next Steps...
All sites are in the process of practicing test administration procedures and will submit the reliability materials for approval prior to data collection
We are working on streamlining the battery and the materials
Data collection to begin (hopefully) in February
Neuropsych, facial imaging, brain imaging, dysmorphology