NIH T32 in Complexity and Neurobiology


The overarching aim of this training program is to produce researchers poised to apply state-of-the-art analytic tools to understand the neurobiology of drug abuse, with a specific focus on characterizing the neural substrates of addiction and other comorbid psychopathologies, always with an eye toward clinical application.

Both pre and post-doctoral scholars are instructed in the application of complex systems and data science approaches to the neurobiology of substance abuse, with the dual strategies of:

(1) training candidates with expertise in data science, applied mathematics, computer science, and complex systems to apply their skills to the neuroscience of addiction; and
(2) training candidates pursuing addiction research from neuroscience, psychiatry, psychology and genetics in the theory and application of Big Data methods, including network analysis, machine learning algorithms, cross-validation frameworks, and other advanced statistical techniques.

The core curriculum incorporates:

(1) the established complex systems and data science graduate certificate at the University of Vermont
(2) course work in neuroscience, psychology and addiction, including classes focused on developing human subjects research skills, as well as
(3) specialized courses designed to directly and effectively bridge the gap between the core disciplines.

Armed with coherent domain knowledge and practiced with advanced methods for complex systems, trainees will develop analysis pipelines that:

(1) incorporate sophisticated aggregation of longitudinal and multi-modal datasets, including various neuroimaging modalities, genetic information, survey and clinical data;
(2) harness the power of supercomputing and modern machine learning algorithms to step beyond linear and univariate effects; and
(3) address questions of immediate clinical importance to substance abuse that can inform the determination of risk factors, treatment and intervention strategy, and policy decisions.

Open Positions

Two Post-doctoral Fellowship Positions:

Starting January 2019. This training program will be demanding, invigorating, hopefully enlightening, and career-shaping. Come join us in Vermont for an exciting adventure!

Contacts: or

The Curious Platypus


Candidates must have completed doctoral training in psychology, neuroscience, bioinformatics, or a related discipline, and have a record of research in genetics or neuroimaging.

Candidates must be a U.S. citizen, noncitizen national, or lawfully admitted for permanent residence at the time of appointment.

More Details:

currently accepting applications for 2 post-doctoral fellowships (NIH-NRSA), starting as early as January 2019, within an exciting new training program on the application of Big Data methods to large-scale, multi-modal (neuroimaging, genetic, psychometric) datasets, in the context of addiction research. These positions are renewable for up to 3 years.

Stipends will be approx. $49,000, depending on previous post-doctoral experience (0-2 years allowable).

Selected candidates will be based in the University of Vermont Department of Psychiatry, and may choose a mentor from a number of participating faculty during the interview process. In addition to the Department of Psychiatry, selected candidates will have the opportunity to interact and collaborate with faculty across a broad range of departments, within several different colleges at UVM. These include Cellular, Molecular, and Biomedical Sciences, Experimental Psychology, Clinical Psychology, as well as the Departments of Mathematics and Statistics, and Computer Science. Trainees will also participate in activities of the Vermont Complex Systems Center, and will have supercomputing capability through the Vermont Advanced Computing Core. Trainees will have direct access to, and be able to harness local expertise with large neuroimaging datasets including IMAGEN, ENIGMA-Addiction, and The Adolescent Brain Cognitive Development (ABCD) Study. Candidates will also have the opportunity to interact with collaborators at the University of California, San Diego, including Anders Dale and Terry Jernigan. This is a far-reaching and truly interdisciplinary opportunity for selected candidates to receive state-of-the-art training in the application of machine-learning methodologies to the largest existing neuroimaging datasets.

Candidates must have completed doctoral training in psychology, neuroscience, bioinformatics, or a related discipline, and have a record of research in genetics or neuroimaging. Programming experience (MATLAB, Python, R, etc.) will be highly valuable, but is not absolutely required as a prerequisite, (though interest in learning to program is an absolute must). Candidates must also demonstrate some basic mathematical and statistical competence, and show interest in expanding their knowledge in this area. Trainees will be selected on the basis of academic record, interviews, and references, and must be U.S. citizens, noncitizen nationals, or lawfully admitted for permanent residence at the time of appointment.

As a Post-Doctoral researcher, the candidate will be responsible for neuroimaging and/or genetic data analysis in the context of a project designed by the candidate themselves in collaboration with their chosen mentor. The candidate will be expected to prepare manuscript(s) on this project for publication. Additional responsibilities will include:
• Completing coursework (1 course/semester) for the Certificate of Graduate Study in Complex Systems.
• Attending lab meetings (1/week), journal clubs (1/week), and a program-specific seminar.
• Gaining experience designing experiments and collecting data.
• Preparing a grant application at the conclusion of the training program.

My Image

Data Sets



    IMAGEN is a multi-site European study of adolescent development. 2,400 fourteen year-olds underwent extensive neuroimaging (four fMRI tasks, resting state fMRI, DTI, anatomical scans), genotyping (SNPs imputed to 1000 Genomes, gene expression, methylation CNVs) and phenotyping (substance use, mental health, personality, cognition, IQ, parental substance use and personality). Subjects were re-assessed for substance use, mental health and personality at age 16, ages 18/19, and age 23, and thus far, 1,500 subjects have returned to the lab repeating the extensive neuroimaging and phenotypic assessments. As a former PI on the Dublin, Ireland site, PI Garavan has full access to these data and complete copies exist on the PI’s servers at UVM.



    ENIGMA Addiction Working Group: The Addiction working group is an international collaboration based on the ENIGMA model. To discover unknown trait-SNP associations, an unbiased search across the whole genome, known as a genome-wide association study (GWAS), requires the testing of hundreds of thousands to millions of SNPs. Since a stringent threshold, conventionally p ≤ 5x10-8, is required to correct for multiple comparisons and the sample size of a typical neuroimaging study is less than N=100, sharing data across multiple sites is necessary to achieve sufficient statistical power to perform a GWAS on neuroimaging phenotypes. The Addiction working group collectively possesses neuroimaging, behavioral, and genomic data obtained from over 15,000 subjects from labs in 14 countries. Multiple imaging modalities are available including structural MRI, DTI and resting-state fMRI. Joint projects with other working groups of the ENIGMA Consortium are also planned to investigate the brain sources of comorbidity (e.g. Depression, Bipolar Disorder, Schizophrenia, OCD and ADHD). Hugh Garavan and Scott Mackey (UVM Psychiatry) lead the working group and have full access to the shared data.

  • ABCD:


    ABCD: The Adolescent Brain Cognitive Development study is a landmark, $300 million, multi-institute NIH grant that aims to track adolescent neurodevelopment. 12,000 subjects, aged 9 and 10, will be recruited at 21 sites across the USA, will undergo extensive neuroimaging (structure and function), genotyping and phenotyping (psychometrics, biospecimens and mobile technologies), and will be followed for ten years with annual lab visits and biennial neuroimaging. PI Garavan is also PI at the UVM ABCD site and is an associate director of the project’s Coordinating Core. The ABCD study will be the largest neuroimaging study ever and presents unparalleled opportunities to exploit “Big Data” methodologies to characterize brain function and the multitude of interacting influences on its development.

  • Generation R:

    Generation R:

    Generation R: The Generation R Study is based in Rotterdam, the second largest city in the Netherlands, and conducted at one large academic medical institution, the Erasmus Medical Center. In total, 9,778 expectant mothers were enrolled in the study between 2002 and 2006. A total of 9,745 live births (51% female) were born from enrolled mothers. After exclusion of mothers who did not give consent for further studies and families living outside of the study area, a total of 7,893 children were available for post-natal follow-up. Assessments began prenatally, which included fetal ultrasounds during early (gestational age < 18 weeks), mid (18 – 25 weeks) and late (> 25 weeks) pregnancy, repeated questionnaires, anthropometry, blood and urine sampling. Postnatal assessments included questionnaires at 2, 6, 12, 18, 24, 30 months and 3, 6, and mostly recently completed 9-10 years. At 3 months, an in-home visit was performed, in total 6690 children visited the research center at age 6 years. Data collection includes cognitive testing, behavioral assessments and motor development. Multimodal neuroimaging—including diffusion and resting state data—have been collected on participants at ages 5-7 and 9-10 years of age in more than 5000 children. Epigenetic data have also been collected on 1400 study participants, of which 500 with three repeated measures.

  • Human Connectome Project (HCP):

    Human Connectome Project (HCP):

    Human Connectome Project (HCP): Successful charting of the human connectome in healthy adults will pave the way for future studies of brain circuitry during development and aging and in numerous brain disorders. The HCP aims to map the human connectome in 1200 healthy adults and is making these data freely available to the scientific community using a powerful, user-friendly informatics platform. The HCP protocol includes extensive phenotyping, structural (T1- and T2-weighted and diffusion), and functional (task-based and resting state) MRI, as well as resting state and task based EEG and MEG on a subset of participants. As of December 2015, preprocessed and quality controlled data are available for 970 participants. In addition, the HCP Lifespan project is collecting multi-modal imaging and behavioral data in participants from six age groups (4-6, 8-9, 14-15, 25-35, 45-55, 65-75) using a shortened protocol, compatible with the full HCP protocol, that is better tolerated by those younger and older participants.

  • Nathan Kline Institute (NKI) / Rockland Sample (NKI/RS):

    Nathan Kline Institute (NKI) / Rockland Sample (NKI/RS):

    Nathan Kline Institute (NKI) / Rockland Sample (NKI/RS): NKI/RS is designed to be a phenotypically rich, heterogeneous, nationally representative neuroimaging sample, consisting of data obtained from individuals between the ages of 4 and 85 years-old ( All individuals included in the sample undergo semi-structured diagnostic psychiatric interviews, and complete a battery of psychiatric, cognitive and behavioral assessments in order to provide comprehensive phenotypic information for the purpose of exploring brain/behavior relationships. A full list of phenotypic assessments can be viewed here:; the neuroimaging protocol, which includes structural (T1-weighted and diffusion), and functional (task-based and resting state) MRI, can be viewed here: The primary goal of the NKI/RS is to generate a large scale, extensively phenotyped dataset for the purpose of discovery science. As of July 2015, preprocessed and quality controlled data are available for 689 participants viewable here.

  • NIH Normal Brain Development Study:

    NIH Normal Brain Development Study:

    NIH Normal Brain Development Study: The NIH MRI Study of Normal Brain Development is a large, multi-site project that provides a normative database to study relations between healthy brain maturation and behavior (Evans, 2006). Subjects were recruited throughout the United States utilizing a population-based sampling method aimed at minimizing selection bias (Waber et al., 2007). Using available U.S. Census 2000 data, a representative, typically developing sample was recruited at 6 pediatric study centers. The 6 pediatric centers consisted of: Children's Hospital (Boston), Children's Hospital Medical Center (Cincinnati), University of Texas Houston Medical School (Houston), UCLA Neuropsychiatric Institute and Hospital (Los Angeles), Children's Hospital of Philadelphia (Philadelphia) and Washington University (St. Louis). Recruitment was monitored throughout the study, ensuring that enrollment across all pediatric centers was demographically representative with regard to age, gender, ethnicity and socioeconomic status (full demographic features of subjects are provided in Evans, 2006). Specifically, census data were used to define low-, medium-, and high-income categories for families in the overall population, as well as to determine the expected distribution of race/ethnicity within each of the income categories. Race/ethnicity by income categories were distributed across the study’s planned age distribution. Regionally specific targets were subsequently created for each pediatric study center based on postal code census data. Subjects were not recruited through convenience volunteer methods, but were obtained and then screened through targeted mailings. The Objective 1 database (release 4.0) used in this study included 431 healthy youth, and upon enrollment (i.e., first study visit), ages ranged from 4 years and 6 months to 18 years and 3 months. The study followed a longitudinal design such that participants underwent MRI brain scanning and behavioral testing on three separate visits, occurring at roughly 2-year intervals.

  • Vermont Family Study:

    Vermont Family Study:

    Vermont Family Study: The Vermont Family Study is a smaller clinical sample of probands, their siblings, and their parents, which was collected at the University of Vermont. A total of 399 children from 205 families participated in the study. Children were selected to score highly on either attention problems or aggressive behavior and were examined using measures of dimensional psychopathology, DSM diagnoses, temperament, and a large number of family variables. Genetic data was collected on parents and children using the Psych SNP chip.

  • Dysregulation Family Study:

    Dysregulation Family Study:

    Dysregulation Family Study: The Dysregulation Family Study is also a smaller clinical sample collected at the University of Vermont. It is the combination of two studies, one of which is ongoing. A total of 237 probands have participated with family members, for a total of 546 individuals. Children were selected to have either broad problems with emotional regulation or to be controls from the clinic or from the community. This sample includes data from DSM and dimensional measures of psychopathology, temperament, eye tracking for emotion recognition, heart rate variability during frustration, delay discounting, and whole epigenome data on a subsample.

Program Faculty

The Curious Platypus

Supported By

NIH/NIDA T32DA043593 Award
Training in Complex Systems and Data Science Approaches Applied to the Neurobiology of Drug Use

Stacks Image 526078

Stacks Image 526080