SI: GENETIC NEUROIMAGING IN AGING AND AGE-RELATED DISEASES
The ENIGMA Consortium: large-scale collaborative analyses
of neuroimaging and genetic data
Paul M. Thompson & Jason L. Stein & Sarah E. Medland & Derrek P. Hibar &
Alejandro Arias Vasquez & Miguel E. Renteria & Roberto Toro & Neda Jahanshad &
Gunter Schumann & Barbara Franke & Margaret J. Wright & Nicholas G. Martin &
Ingrid Agartz & Martin Alda & Saud Alhusaini & Laura Almasy & Jorge Almeida &
Kathryn Alpert & Nancy C. Andreasen & Ole A. Andreassen & Liana G. Apostolova &
Katja Appel & Nicola J. Armstrong & Benjamin Aribisala & Mark E. Bastin & Michael Bauer &
Carrie E. Bearden & Ørjan Bergmann & Elisabeth B. Binder & John Blangero &
Henry J. Bockholt & Erlend Bøen & Catherine Bois & Dorret I. Boomsma & Tom Booth &
Ian J. Bowman & Janita Bralten & Rachel M. Brouwer & Han G. Brunner &
David G. Brohawn & Randy L. Buckner & Jan Buitelaar & Kazima Bulayeva &
Juan R. Bustillo & Vince D. Calhoun & Dara M. Cannon & Rita M. Cantor &
Melanie A. Carless & Xavier Caseras & Gianpiero L. Cavalleri & M. Mallar Chakravarty &
Kiki D. Chang & Christopher R. K. Ching & Andrea Christoforou & Sven Cichon &
Vincent P. Clark & Patricia Conrod & Giovanni Coppola & Benedicto Crespo-Facorro &
© The Author(s) 2014. This article is published with open access at Springerlink.com
Guest Editor: John D. Van Horn
This article reviews work published by the ENIGMA Consortium and its
Working Groups (http://enigma.ini.usc.edu). It was written collaboratively;
for important intellectual content, using Google Docs for parallel editing, and
approved it. Some ENIGMA investigators contributed to the design and
implementation of ENIGMA or provided data but did not participate in the
is available at http://enigma.ini.usc.edu/publications/the-enigma-consortium-
in-review/ For ADNI, some investigators contributed to the design and
implementation of ADNI or provided data but did not participate in the
analysis or writing of this report. A complete listing of ADNI investigators
is available at http://adni.loni.usc.edu/wp-content/uploads/how_to_apply/
ADNI_Acknowledgement_List.pdf The work reviewed here was funded by
(2012); the funding for listed consortia is also itemized in Stein et al. (2012).
P. M. Thompson (*) I D. P. Hibar I N. Jahanshad I
I. J. Bowman I C. R. K. Ching I H. Dong I C. D. Leonardo I
A. W. Toga
Imaging Genetics Center, Institute for Neuroimaging and
Informatics, Keck School of Medicine, University of Southern
California, 2001 N. Soto Street, Los Angeles, CA 90033, USA
G. W. Montgomery I L. Strike
Genetic Epidemiology Laboratory, Queensland Institute of Medical
Research, Brisbane, Australia
P. H. Lee I J. W. Smoller
Broad Institute of Harvard and MIT, Boston, MA, USA
A. A. Vasquez I B. Franke I J. Bralten I H. G. Brunner I
Department of Human Genetics, Radboud University Medical Centre,
Nijmegen, The Netherlands
A. A. Vasquez I B. Franke I J. Bralten I J. Buitelaar I S. E. Fisher I
C. Francks I M. Rijpkema I M. P. Zwiers
Donders Institute for Brain, Cognition and Behaviour, Department of
Cognitive Neuroscience, Radboud University Medical Centre, Nijmegen,
Human Genetics and Cognitive Functions, Institut Pasteur, Paris,
Brain Imaging and Behavior
Joanne E. Curran & Michael Czisch & Ian J. Deary & Eco J. C. de Geus & Anouk den Braber &
Giuseppe Delvecchio & Chantal Depondt & Lieuwe de Haan & Greig I. de Zubicaray &
Danai Dima & Rali Dimitrova & Srdjan Djurovic & Hongwei Dong & Gary Donohoe &
Ravindranath Duggirala & Thomas D. Dyer & Stefan Ehrlich & Carl Johan Ekman &
Torbjørn Elvsåshagen & Louise Emsell & Susanne Erk & Thomas Espeseth & Jesen Fagerness &
Scott Fears & Iryna Fedko & Guillén Fernández & Simon E. Fisher & Tatiana Foroud &
Peter T. Fox & Clyde Francks & Sophia Frangou & Eva Maria Frey & Thomas Frodl &
Vincent Frouin & Hugh Garavan & Sudheer Giddaluru & David C. Glahn & Beata Godlewska &
Rita Z. Goldstein & Randy L. Gollub & Hans J. Grabe & Oliver Grimm & Oliver Gruber &
TulioGuadalupe &RaquelE.Gur &RubenC.Gur &HaraldH.H.Göring &SaskiaHagenaars &
Tomas Hajek & Geoffrey B. Hall & Jeremy Hall & John Hardy & Catharina A. Hartman &
Johanna Hass & Sean N. Hatton & Unn K. Haukvik & Katrin Hegenscheid & Andreas Heinz &
Ian B. Hickie & Beng-Choon Ho & David Hoehn & Pieter J. Hoekstra & Marisa Hollinshead &
Avram J. Holmes & Georg Homuth & Martine Hoogman & L. Elliot Hong & Norbert Hosten &
Jouke-Jan Hottenga & Hilleke E. Hulshoff Pol & Kristy S. Hwang & Clifford R. Jack Jr &
Mark Jenkinson & Caroline Johnston & Erik G. Jönsson & René S. Kahn &
Dalia Kasperaviciute & Sinead Kelly & Sungeun Kim & Peter Kochunov & Laura Koenders &
Bernd Krämer & John B. J. Kwok & Jim Lagopoulos & Gonzalo Laje & Mikael Landen &
Bennett A. Landman & John Lauriello & Stephen M. Lawrie & Phil H. Lee &
CNRS URA 2182 ‘Genes, synapses and cognition’, Institut Pasteur,
Sorbonne Paris Cité, Human Genetics and Cognitive Functions,
Université Paris Diderot, Paris, France
D. C. Glahn
Olin Neuropsychiatry Research Center, Institute of Living, Hartford
Hospital, Hartford, CT, USA
D. C. Glahn I C.-s. Li I G. D. Pearlson I E. Sprooten I A. M. Winkler
Department of Psychiatry, Yale University School of Medicine, New
Haven, CT, USA
K. Appel I H. J. Grabe
Department of Psychiatry and Psychotherapy, University of
Greifswald, Greifswald, Germany
O. A. Andreassen I A. Christoforou I S. Djurovic I T. Espeseth I
S. Giddaluru I U. K. Haukvik I S. Le Hellard I M. Mattingsdal I I. Melle I
V. M. Steen I L. T. Westlye
NORMENT, KG Jebsen Centre for Psychosis Research,
Oslo University Hospital and Institute of Clinical Medicine,
University of Oslo, Oslo, Norway
S. Alhusaini I G. L. Cavalleri I C. D. Whelan
Department of Molecular and Cellular Therapeutics, Royal College
of Surgeons in Ireland, Dublin 2, Ireland
L. Almasy I J. Blangero I M. A. Carless I J. E. Curran I
R. Duggirala I T. D. Dyer I H. H. H. Göring I E. K. Moses I
C. P. Peterson
Department of Genetics, Texas Biomedical Research Institute,
San Antonio, TX, USA
N. C. Andreasen I B.-C. Ho
Department of Psychiatry, University of Iowa, Iowa City, IA, USA
L. G. Apostolova I G. Coppola
Department of Neurology, David Geffen School of Medicine, UCLA,
Los Angeles, CA, USA
B. Aribisala I M. E. Bastin I T. Booth I I. J. Deary I D. C. Liewald I
L. M. Lopez I M. Luciano I S. Muñoz Maniega I N. A. Royle I
J. M. Starr I M. C. Valdés Hernández I J. M. Wardlaw
Centre for Cognitive Ageing and Cognitive Epidemiology,
The University of Edinburgh, 7 George Square, Edinburgh, UK
B. Aribisala I S. Muñoz Maniega I J. M. Wardlaw
Scottish Imaging Network, A Platform for Scientific Excellence
(SINAPSE) Collaboration, Scotland, UK
B. Aribisala I M. E. Bastin I S. Muñoz Maniega I N. A. Royle I
M. C. Valdés Hernández I J. M. Wardlaw
Brain Research Imaging Centre, The University of Edinburgh,
E. B. Binder I M. Czisch I D. Hoehn I B. Müller-Myhsok I B. Pütz I
P. G. Sämann I C. Wolf
Max Planck Institute of Psychiatry, Munich, Germany
V. D. Calhoun I J. M. Shoemaker I J. Turner
The Mind Research Network, Albuquerque, NM, USA
D. I. Boomsma I E. J. C. de Geus I A. den Braber I I. Fedko I
J.-J. Hottenga I D. van ‘t Ent
Department of Biological Psychology, VU University, Neuroscience
Campus, Amsterdam, The Netherlands
J. Bralten I G. Fernández
Department of Cognitive Neuroscience, Radboud University
Medical Centre, Donders Institute for Brain, Cognition and
Behavior, Nijmegen, The Netherlands
Brain Imaging and Behavior
Stephanie Le Hellard & Herve Lemaître & Cassandra D. Leonardo & Chiang-shan Li &
Benny Liberg & David C. Liewald & Xinmin Liu & Lorna M. Lopez & Eva Loth &
Anbarasu Lourdusamy & Michelle Luciano & Fabio Macciardi & Marise W. J. Machielsen &
GlendaM.MacQueen & UlrikF.Malt & RenéMandl & DaraS.Manoach &Jean-LucMartinot &
Mar Matarin & Karen A. Mather & Manuel Mattheisen & Morten Mattingsdal &
Andreas Meyer-Lindenberg & Colm McDonald & Andrew M. McIntosh &
Francis J. McMahon & Katie L. McMahon & Eva Meisenzahl & Ingrid Melle &
Yuri Milaneschi & Sebastian Mohnke & Grant W. Montgomery & Derek W. Morris &
Eric K. Moses & Bryon A. Mueller & Susana Muñoz Maniega & Thomas W. Mühleisen &
Bertram Müller-Myhsok & Benson Mwangi & Matthias Nauck & Kwangsik Nho &
ThomasE.Nichols &Lars-GöranNilsson &AllisonC.Nugent &LarsNyberg &ReneL.Olvera &
Jaap Oosterlaan & Roel A. Ophoff & Massimo Pandolfo & Melina Papalampropoulou-Tsiridou &
Martina Papmeyer & Tomas Paus & Zdenka Pausova & Godfrey D. Pearlson &
Brenda W. Penninx & Charles P. Peterson & Andrea Pfennig & Mary Phillips & G. Bruce Pike &
Jean-Baptiste Poline & Steven G. Potkin & Benno Pütz & Adaikalavan Ramasamy &
Jerod Rasmussen & Marcella Rietschel & Mark Rijpkema & Shannon L. Risacher &
Joshua L. Roffman & Roberto Roiz-Santiañez & Nina Romanczuk-Seiferth & Emma J. Rose &
Natalie A. Royle & Dan Rujescu & Mina Ryten & Perminder S. Sachdev & Alireza Salami &
Theodore D. Satterthwaite & Jonathan Savitz & Andrew J. Saykin & Cathy Scanlon &
Lianne Schmaal & Hugo G. Schnack & Andrew J. Schork & S. Charles Schulz & Remmelt Schür &
Larry Seidman & Li Shen & Jody M. Shoemaker & Andrew Simmons & Sanjay M. Sisodiya &
Colin Smith & Jordan W. Smoller & Jair C. Soares & Scott R. Sponheim & Emma Sprooten &
R. M. Brouwer I H. E. Hulshoff Pol I R. S. Kahn I R. Mandl I
R. A. Ophoff I H. G. Schnack I R. Schür I M. Van den Heuvel I
N. E. M. Van Haren
Brain Center Rudolf Magnus, University Medical Center Utrecht,
Utrecht, The Netherlands
R. L. Buckner I J. Fagerness
Massachusetts General Hospital, Boston, MA, USA
Karakter Child and Adolescent Psychiatry University
Center, Nijmegen, The Netherlands
N. I. Vavilov Institute of General Genetics, Russian Academy of
Sciences, Gubkin str. 3, Moscow 119991, Russia
J. R. Bustillo
Department of Psychiatry, University of New Mexico, Albuquerque,
V. D. Calhoun
Department of Electrical and Computer Engineering, University of
New Mexico, Albuquerque, NM, USA
D. M. Cannon I G. Donohoe I L. Emsell I C. McDonald I C. Scanlon
Clinical Neuroimaging Laboratory, National University of Ireland
Galway, University Road, Galway, Ireland
R. M. Cantor I S. Fears I R. A. Ophoff
Center for Neurobehavioral Genetics, University of California, Los
Angeles, CA, USA
M. M. Chakravarty
The Kimel Family Translational Imaging Genetics Laboratory, The
Centre for Addiction and Mental Health, Toronto, ON, Canada
A. Christoforou I S. Le Hellard I V. M. Steen
Dr Einar Martens Research Group for Biological Psychiatry,
Department of Clinical Medicine, University of Bergen, Bergen,
Dr Einar Martens Research Group for Biological Psychiatry,
Center for Medical Genetics and Molecular Medicine, Haukeland
University Hospital, Bergen, Norway
T. W. Mühleisen
Department of Genomics, Life and Brain Center, University of Bonn,
S. Cichon I T. W. Mühleisen
Institute of Human Genetics, University of Bonn, Bonn, Germany
S. Cichon I T. W. Mühleisen
Institute for Neuroscience and Medicine (INM-1), Research Centre
Jülich, Jülich, Germany
Division of Medical Genetics, Department of Biomedicine,
University of Basel, Basel, Switzerland
B. Crespo-Facorro I I. R. Roiz-Santiañez I I. D. Tordesillas-Gutierrez
Department of Psychiatry, Marqués de Valdecilla University
Hospital, IFIMAV, School of Medicine, University of Cantabria,
Brain Imaging and Behavior
John M. Starr & Vidar M. Steen & Stephen Strakowski & Lachlan Strike & Jessika Sussmann &
Philipp G. Sämann & Alexander Teumer & Arthur W. Toga & Diana Tordesillas-Gutierrez &
Daniah Trabzuni & Sarah Trost & Jessica Turner & Martijn Van den Heuvel &
Nic J. van der Wee & Kristel van Eijk & Theo G. M. van Erp & Neeltje E. M. van Haren &
Dennis van ‘t Ent & Marie-Jose van Tol & Maria C. Valdés Hernández & Dick J. Veltman &
Amelia Versace & Henry Völzke & Robert Walker & Henrik Walter & Lei Wang &
Joanna M. Wardlaw & Michael E. Weale & Michael W. Weiner & Wei Wen & Lars T. Westlye &
Heather C. Whalley & Christopher D. Whelan & Tonya White & Anderson M. Winkler &
Katharina Wittfeld & Girma Woldehawariat & Christiane Wolf & David Zilles & Marcel P. Zwiers &
Anbupalam Thalamuthu & Peter R. Schofield & Nelson B. Freimer & Natalia S. Lawrence &
Wayne Drevets & the Alzheimer’s Disease Neuroimaging Initiative, EPIGEN Consortium,
Saguenay Youth Study (SYS) Group
Abstract The Enhancing NeuroImaging Genetics through
Meta-Analysis (ENIGMA) Consortium is a collaborative
network of researchers working together on a range of large-
scale studies that integrate data from 70 institutions
worldwide. Organized into Working Groups that tackle
questions in neuroscience, genetics, and medicine, ENIGMA
studies have analyzed neuroimaging data from over 12,826
subjects. In addition, data from 12,171 individuals were
provided by the CHARGE consortium for replication of
findings, in a total of 24,997 subjects. By meta-analyzing
B. Crespo-Facorro I R. Roiz-Santiañez I D. Tordesillas-Gutierrez
Centro Investigación Biomédica en Red SaludMental (CIBERSAM),
L. M. Lopez
Department of Psychology, The University of Edinburgh, Edinburgh,
C. Depondt I M. Pandolfo
Department of Neurology, Hopital Erasme, Universite Libre de
Bruxelles, 1070 Brussels, Belgium
G. I. de Zubicaray
School of Psychology, University of Queensland, Brisbane,
QLD 4072, Australia
Department of Medical Genetics, Oslo University Hospital, Oslo,
G. Donohoe I S. Kelly I D. W. Morris
Neuropsychiatric Genetics Research Group, Department of
Psychiatry, Institute for Molecular Medicine and Trinity College
Institute for Neuroscience, Trinity College, Dublin, Ireland
S. Ehrlich I R. L. Gollub I M. Hollinshead I A. J. Holmes I
D. S. Manoach
MGH/HMS Martinos Center for Biomedical Imaging,
Massachusetts General Hospital, Charlestown, MA, USA
S. Ehrlich I J. Hass
University Hospital C.G. Carus, Department of Child and Adolescent
Psychiatry, Dresden University of Technology, Dresden, Germany
S. Erk I A. Heinz I S. Mohnke I N. Romanczuk-Seiferth I H. Walter
Department of Psychiatry and Psychotherapy, Charité,
Universitaetsmedizin Berlin, Charité Campus Mitte, Berlin, Germany
T. Espeseth I L. T. Westlye
Department of Psychology, University of Oslo, Oslo, Norway
S. E. Fisher I C. Francks I T. Guadalupe
Max Planck Institute for Psycholinguistics, 6500 AH Nijmegen, The
P. T. Fox
Research Imaging Institute, UT Health Science Center at San
Antonio, San Antonio, TX, USA
P. T. Fox
South Texas Veterans Health Care Center, San Antonio, TX, USA
Neurospin, Commissariat à l’Energie Atomique, Paris, France
R. L. Gollub I A. J. Holmes I P. H. Lee I D. S. Manoach I
J. L. Roffman I J. W. Smoller
Department of Psychiatry, Massachusetts General Hospital, Boston,
H. J. Grabe I K. Wittfeld
German Center for Neurodegenerative Diseases (DZNE), University
of Greifswald, Greifswald, Germany
O. GrimmI A. Meyer-Lindenberg I M. Rietschel
Central Institute of Mental Health, Medical Faculty Mannheim,
University of Heidelberg, Mannheim, Germany
Brain Imaging and Behavior
results from many sites, ENIGMA has detected factors
that affect the brain that no individual site could detect
on its own, and that require larger numbers of subjects
than any individual neuroimaging study has currently
collected. ENIGMA’s first project was a genome-wide
association study identifying common variants in the
genome associated with hippocampal volume or
intracranial volume. Continuing work is exploring genetic
associations with subcortical volumes (ENIGMA2) and
white matter microstructure (ENIGMA-DTI). Working
groups also focus on understanding how schizophrenia,
bipolar illness, major depression and attention deficit/
hyperactivity disorder (ADHD) affect the brain. We
review the current progress of the ENIGMA Consortium,
along with challenges and unexpected discoveries made on
Origins of brain imaging in human populations
During the “Decade of the Brain” in the 1990s (Jones and
Mendell 1999), a number of major neuroimaging centers
began to scan hundreds of patients and healthy individuals
O. Gruber I B. Krämer I S. Trost I D. Zilles
Center for Translational Research in Systems Neuroscience and
Psychiatry, Department of Psychiatry, Georg August University,
C. Bois I R. Dimitrova I S. Hagenaars I J. Hall I S. M. Lawrie I
A. M. McIntosh I M. Papalampropoulou-Tsiridou I M. Papmeyer I
J. Sussmann I H. C. Whalley
Division of Psychiatry, Royal Edinburgh Hospital, University of
Edinburgh, Edinburgh, UK
J. Hardy I M. Ryten I D. Trabzuni
Department of Molecular Neuroscience, UCL Institute, London, UK
K. Hegenscheid I N. Hosten
Department of Diagnostic Radiology and Neuroradiology, University
of Greifswald, Greifswald, Germany
C. A. Hartman I P. J. Hoekstra
Department of Psychiatry, University Medical Center Groningen,
University of Groningen, Groningen, The Netherlands
G. Homuth I A. Teumer
Interfaculty Institute for Genetics and Functional Genomics,
University of Greifswald, Greifswald, Germany
National Institute of Health Research Biomedical Research Centre for
Mental Health, South London and Maudsley National Health Service
Foundation Trust, London, UK
King’s College London, Institute of Psychiatry, London, UK
I. Agartz I C. J. Ekman I E. G. Jönsson I B. Liberg
Department of Clinical Neuroscience, Karolinska Institutet and
Hospital, Stockholm, Sweden
D. Kasperaviciute I M. Matarin I S. M. Sisodiya
Department of Clinical and Experimental Epilepsy, UCL Institute of
Neurology, London, UK
Center for Computational Biology and Bioinformatics, Indiana
University School of Medicine, Indianapolis, IN, USA
S. Kim I K. Nho I S. L. Risacher I A. J. Saykin I L. Shen
Department of Radiology and Imaging Sciences, Center for
Neuroimaging, Indiana University School of Medicine, Indianapolis,
L. E. Hong I P. Kochunov
Maryland Psychiatric Research Center, Department of Psychiatry,
University of Maryland School of Medicine, Baltimore, MD, USA
D. G. Brohawn I J. Fagerness I P. H. Lee I J. W. Smoller
Psychiatric and Neurodevelopmental Genetics Unit, Center for
Human Genetic Research, Massachusetts General Hospital, Boston,
X. Liu I F. J. McMahon I G. Woldehawariat
Mood and Anxiety Disorders Section, Human Genetics Branch,
Intramural Research Program, National Institute of Mental Health,
National Institutes of Health, US Dept of Health and Human Services,
Bethesda, MD, USA
Taub Institute for Research on Alzheimer Disease and the Aging
Brain, Columbia University Medical Center, New York, NY, USA
G. Schumann I G. Delvecchio I D. Dima I E. Loth
MRC-SGDP Centre, Institute of Psychiatry, King’s College London,
F. Macciardi I S. G. Potkin I J. Rasmussen I T. G. M. van Erp
Department of Psychiatry and Human Behavior, University of
California, Irvine, CA, USA
K. A. Mather I P. S. Sachdev I W. Wen
Centre for Healthy Brain Ageing, School of Psychiatry, University of
New South Wales Medicine, Sydney, New South Wales, Australia
Department of Biomedicine, Aarhus University, Aarhus, Denmark
Department of Genomic Mathematics, University of Bonn, Bonn,
Research Unit, Sorlandet Hospital HF, Kristiansand, Norway
Brain Imaging and Behavior
using a variety of neuroimaging methods. The accelerating
pace of data collection was driven mainly by the wide
availability of MRI around the world. The structure and
function of the living brain was beginning to be mapped in
unprecedented detail in human populations.
In a typical neuroimaging study—both now and 20 years
ago—between ten and a few hundred subjects might have
factors that affect brain structure and function. Early
studies—such as lesion studies—correlated radiological
measures with clinical diagnosis and behavior, but the
study of large populations represented a new movement
in human brain mapping. Fundamental questions in
neuroscience could now be examined—what are the
effects of aging, degenerative disease and psychiatric
illness on the living brain? How do brain measures
relate to cognition and behavior? Do brain measures
predict our risk for disease, or prognosis in those who
There was growing confidence that questions of broad
societal and medical impact could be better understood if
enough brain scans were collected—projects were initiated
to examine effects on the brain of psychiatric medications,
drugs and alcohol abuse, dietary factors, and many other
factors including education, cardiovascular fitness, as well as
pharmacologic and behavioral interventions.
At the same time, the broad availability of brain scans
led to the development of widely adopted tools to analyze
the resulting data. Software such as Statistical Parametric
Mapping (SPM; Friston et al. 1995; Frackowiak 1997),
K. L. McMahon
Centre for Advanced Imaging, University of Queensland, Brisbane,
Ludwig-Maximilians-University (LMU), Munich, Germany
B. W. Penninx I L. Schmaal I D. J. Veltman
Department of Psychiatry and Neuroscience Campus Amsterdam,
VU University Medical Center, Amsterdam, The Netherlands
E. K. Moses
Centre for Genetic Origins of Health and Disease, The University
of Western Australia, Perth, Australia
B. A. Mueller I S. C. Schulz I S. R. Sponheim
Department of Psychiatry, University of Minnesota Medical Center,
Minneapolis, MN, USA
Department of Psychology, Stockholm University, Stockholm,
L. Nyberg I A. Salami
Umeå Center for Functional Brain Imaging (UFBI), Umeå
University, Umeå, Sweden
R. L. Olvera
Department of Psychiatry, UT Health Science Center at San Antonio,
San Antonio, TX, USA
Rotman Research Institute, University of Toronto, Toronto, ON,
The Hospital for Sick Children, University of Toronto, Toronto, ON,
N. J. van der Wee
Department of Psychiatry and Leiden Institute for Brain and
Cognition, Leiden University Medical Center, Leiden,
A. Ramasamy I M. Ryten I M. E. Weale
Department of Medical and Molecular Genetics, King’s College
London, London, UK
P. S. Sachdev
Neuropsychiatric Institute, Prince of Wales Hospital, Sydney, New
South Wales, Australia
Laureate Institute for Brain Research, Tulsa, OK, USA
T. Foroud I A. J. Saykin
Department of Medical and Molecular Genetics, Indiana University
School of Medicine, Indianapolis, IN, USA
Department of Psychiatry, Beth Israel Deaconess Medical Center,
Boston, MA, USA
Department of Neuroimaging, Institute of Psychiatry, King’s College
London, London, UK
NIHR Biomedical Research Centre for Mental Health at South
London and Maudsley NHS Trust and Institute of Psychiatry,
King’s College London, London, UK
S. R. Sponheim
Minneapolis VA Health Care System, Minneapolis, MN, USA
M.-J. van Tol
Behavioural and Cognitive Neuroscience Neuroimaging Center,
University Medical Center Groningen, Groningen, The Netherlands
Institute for Community Medicine, University of Greifswald,
Brain Imaging and Behavior
FSL (Jenkinson et al. 2012), BRAINS (Pierson et al.
2011) and FreeSurfer (Fischl et al. 2004) among many
other tools, were widely distributed over the internet. This
made it feasible to analyze neuroimaging data and
compute standardized measures from brain scans in a
consistent and agreed way, albeit with methods that are
Early neuroimaging consortia
Early consortium efforts in neuroimaging included the
International Consortium for Brain Mapping (ICBM;
Mazziotta et al. 1995), which recognized the need to establish
normative data on the brain from a wide range of human
populations scanned in different parts of the world. The
ICBM began with an effort to scan around 150 healthy
subjects in Los Angeles, Montreal, and San Antonio, Texas,
and grew to include sites in Europe and Asia that broadened
the age range and ethnic groups assessed. Later, the ICBM
also extended the depth of the neuroimaging measures to
include functional MRI and even post mortem histology and
cytoarchitecture (Amunts et al. 1999).
Given the wide variations in brain anatomy even among
healthy subjects, consortia such as the ICBM developed a
range of “average” anatomical templates based on MRI scans
of hundreds of healthy subjects. Analysis software for brain
images disseminated these average brain templates, and
provided methods to relate new data to previously compiled
M. W. Weiner
Departments of Radiology, Medicine, Psychiatry, University of
California, San Francisco, CA, USA
Department of Child and Adolescent Psychiatry, Erasmus University
Medical Centre, Rotterdam, The Netherlands
M. Rijpkema I M. P. Zwiers
Radboud University NijmegenDonders Institute for Brain, Cognition
and Behavior, Centre for Cognitive Neuroimaging, Nijmegen,
M. Jenkinson I A. M. Winkler
Oxford Centre for Functional MRI of the Brain (FMRIB), University of
Oxford, Oxford, UK
J. B. J. Kwok I P. R. Schofield
Neuroscience Research Australia, Sydney, Australia
J. B. J. Kwok I P. R. Schofield
School of Medical Sciences, University of New South Wales, Sydney,
R. L. Buckner I M. Hollinshead I A. J. Holmes
Center for Brain Science, Harvard University, Cambridge, MA, USA
S. N. Hatton I I. B. Hickie I J. Lagopoulos
The Brain and Mind Research Institute, University of Sydney,
J. M. Starr
Alzheimer Scotland Dementia Research Centre, University
of Edinburgh, Edinburgh, UK
Centre for Regenerative Medicine, University of Edinburgh,
M. E. Bastin I C. Smith
Centre for Clinical Brain Sciences, The University of Edinburgh,
L. de Haan I L. Koenders I M. W. J. Machielsen
Department Early Psychosis, Academic Psychiatric Centre, AMC,
UvA, Amsterdam, Netherlands
D. I. Boomsma I E. J. C. de Geus I Y. Milaneschi I B. W. Penninx
EMGO+Institute, VU University Medical Center, Amsterdam,
A. J. Schork
Cognitive Science Department, UC San Diego, La Jolla,
U. K. Haukvik
Department of Psychiatric Research, Diakonhjemmet Hospital, Oslo,
Department of Clinical Neuropsychology, VU University,
Amsterdam, The Netherlands
Reta Lila Weston Institute and Department of Molecular
Neuroscience, UCL Institute of Neurology, London, UK
Department of Neurology and NeuroSurgery, McGill University,
Montreal, Quebec, Canada
J. Almeida IM. Phillips I A. Versace
Department of Psychiatry, University of Pittsburgh, Pittsburgh,
M. Alda I T. Hajek
Department of Psychiatry, Dalhousie University, Halifax, Nova
Department of Psychiatry, University of Oxford, Oxford, UK
B. Mwangi I J. C. Soares
Department of Psychiatry and Behavioral Sciences, University
of Texas Medical School, Houston, TX, USA
Brain Imaging and Behavior
atlases and datacollections.This led tothe notionof statistical
representations of imaging signals in standardized coordinate
spaces—or “statistical parametric maps”. The wide adoption
of these standard spaces—such as the ICBM or MNI
(Montreal Neurological Institute) spaces—was eased by the
development of automated registration and alignment
methods (Woods et al. 1993; Collins et al. 1994; Ashburner
et al. 1999; Jenkinson et al. 2002) that allowed researchers to
rapidly align their own data to the templates. This effort led to
the rise of voxel-based morphometric approaches and
statistical mapping approaches in general. These
the world—a movement that was stimulated by the
development of the Talairach and Tournoux brain atlases,
which defined anatomical regions in stereotaxic space
(Talairachetal.1993).The Talairachatlas was among the first
to compile a coordinate-based reference system, and it
allowed researchers worldwide to relate their findings to
existing data collections. In the mid-1990s, a group in San
Antonio developed the “Talairach Daemon”, allowing
electronic pooling of findings from brain mapping studies
based on their coordinates in Talairach space. In addition to
the use of standard anatomical templates for reporting results,
this coordinate system opened the door for clinically-oriented
consortia to scan and analyze large-scale patient populations
in a consistent way. The rapid development of nonlinear
registration methods also made it possible to improve the
alignment of new datasets to digital anatomical templates,
for coordinate based reporting of results.
The Alzheimer’s Disease NeuroimagingInitiative(ADNI),
for example (Weiner et al. 2012), scanned around 800 people
University of Texas Center of Excellence on Mood Disorders,
Department of Psychiatry, UT Medical School, Houston, TX, USA
C. E. Bearden I G. Coppola I S. Fears
Department of Psychiatry and Biobehavioral Sciences and the Center
for Neurobehavioral Genetics, The Semel Institute for Neuroscience
and Human Behavior, UCLA, Los Angeles, CA, USA
Longitudinal Studies Section, Translational Gerontology Branch,
National Institute on Aging, Baltimore, MD, USA
Berlin School of Mind and Brain, Humboldt University Berlin,
M. Bauer I A. Pfennig
Department of Psychiatry and Psychotherapy, Carl Gustav Carus
University Hospital, Dresden, Germany
H. J. Grabe
Department of Psychiatry and Psychotherapy, Helios Hospital
Stralsund, Stralsund, Germany
Department of Psychiatry, Harvard Medical School, Harvard
University, Cambridge, MA, USA
Department of Psychiatry, Brown University, Providence, RI, USA
Psychosis Research Unit, Mount Sinai School of Medicine,
New York, NY, USA
Department of Psychiatry and Behavioral Neuroscience, University
of Cincinnati College of Medicine, Cincinnati, OH, USA
R. E. Gur I R. C. Gur I T. D. Satterthwaite
Department of Psychiatry, University of Pennsylvania, Philadelphia,
R. C. Gur
Philadelphia Veterans Administration Medical Center, Philadelphia,
E. M. Frey I T. Frodl
Department of Psychiatry and Psychotherapy, University
Regensburg, Regensburg, Germany
Department of Psychiatry and Psychotherapy, Trinity College,
University Dublin, Dublin, Germany
Stockholm Brain Institute, Stockholm, Sweden
Department of Psychology and Neuroscience Institute,
Georgia State University, Atlanta,
G. B. Hall
Department of Psychology, Neuroscience and Behaviour, McMaster
University, Hamilton, ON, Canada
V. P. Clark
Department of Psychology, University of New Mexico, Albuquerque,
E. Bøen I T. Elvsåshagen I U. F. Malt
Department of Psychosomatic Medicine, Oslo University Hospital,
I. Agartz I Ø. Bergmann I E. Bøen I T. Elvsåshagen I U. F. Malt
Institute of Clinical Medicine, University of Oslo, Oslo, Norway
S. L. Risacher
Indiana Alzheimer Disease Center, Indiana University School of
Medicine, Indianapolis, IN, USA
G. B. Pike
Department of Radiology, University of Calgary, Calgary, Alberta,
Brain Imaging and Behavior
in its first phase, including healthy elderly people, individuals
disease. ADNI began in 2005, after testing the feasibility and
reproducibility of a range of scanning protocols. This led to
North America (Leow et al. 2009; Jahanshad et al. 2010; Jack
2012; Zhan et al. 2012). Many other neuroimaging consortia
havebeen established, including the functional BrainImaging
Research Network (FBIRN) (Potkin and Ford 2009) which
has developed standardized methods for multi-center
functional MRI studies (Glover et al. 2012) and the Mind
Clinical Imaging Consortium (Gollub et al. 2013) focusing
on schizophrenia, as well as research networks focusing on
pediatric imaging (Evans 2006), autism (Ecker et al. 2013),
HIV/AIDS (Cohen et al. 2010) and many others. In fact, the
successes of these multi-site initiatives have led to large-scale
neuroimaging efforts being initiated and funded in other
countries (Carrillo et al. 2012; Alzheimer’s Association
2013; White et al. 2013).
Genome-wide association studies (GWAS)
At the same time, a number of genetic studies using twin or
family-based designs had shown that many brain-derived
measures were significantly heritable (Thompson et al.
2001; Baaré et al. 2001; White et al. 2002; Wright et al.
2002; van Erp et al. 2004a, b; Hulshoff Pol et al. 2006;
Winkler et al. 2010; Kochunov et al. 2010; Blokland et al.
2012; Koten etal. 2009). In other words, a substantial fraction
of the variability in brain measures—especially structural but
T. E. Nichols
Department of Statistics & Warwick Manufacturing Group,
The University of Warwick, Coventry, UK
K. Alpert I L. Wang
Departments of Psychiatry and Behavioral Sciences and Radiology,
Northwestern University, Chicago, IL, USA
B. A. Landman
Electrical Engineering, Vanderbilt University, Nashville, TN, USA
A. C. Nugent
Experimental Therapeutics and Pathophysiology Branch,
National Institute of Mental Health,
Bethesda, MD, USA
Institute of Clinical Chemistry and Laboratory Medicine,
University of Greifswald, Greifswald, Germany
Institute of Neuroscience and Physiology, University of Gothenburg,
Department of Medical Epidemiology and Biostatistics, Karolinska
Institutet, Stockholm, Sweden
M. M. Chakravarty
Institute of Biomaterials and Biomedical Engineering, University of
Toronto, Toronto, ON, Canada
Faculty of Community Medicine, University of Tulsa, Tulsa,
Maryland Institute for Neuroscience and Development (MIND),
Chevy Chase, MD, USA
G. M. MacQueen
Mathison Centre for Mental Health Research and Education,
Hotchkiss Brain Institute, University of Calgary,
Calgary, Alberta, Canada
Munich Cluster for Systems Neurology (SyNergy), Munich,
MRC Centre for Neuropsychiatric Genetics and Genomics,
Institute of Psychological Medicine and Clinical Neurosciences,
Cardiff University, Cardiff, UK
Neuroscience and Mental Health Research Institute, Cardiff
University, Cardiff, UK
J. B. J. Kwok
School of Medical Sciences, University of New South Wales,
Kensington, NSW, Australia
K. S. Hwang
Oakland University William Beaumont School of Medicine,
Rochester Hills, MI, USA
CHU Sainte Justine University Hospital Research Center, Montreal,
Addictions Department, King’s Health Partners, King’s College
London, London, UK
P. T. Fox
South Texas Veterans Health Care System, San Antonio, TX, USA
H. Lemaître I J.-L. Martinot
de Médecine, Paris Sud University-Paris Descartes University, Maison de
Solenn Paris, SHFJ Orsay, Paris, France
Department of Genetics, King Faisal Specialist Hospital and
Research Centre, Riyadh, Saudi Arabia
K. van Eijk
Department of Psychiatry, Rudolf Magnus Institute, University
Medical Center Utrecht, Utrecht, The Netherlands
Brain Imaging and Behavior
also some functional measures, and even brain metabolites
(Batouli et al. 2012)—could be explained by genetic
relationships among individuals. The total amount of gray
and white matter in the brain, the overall volume of the
brain—and even activation patterns on fMRI or connections
tracked with diffusion MRI—were more similar among
family members than unrelated individuals (Peper et al.
2007; Koten et al. 2009; Glahn et al. 2010; Brouwer et al.
2010; Fornito et al. 2011; Blokland et al. 2012; Jahanshad
et al. 2013a; Thompson et al. 2013; Van den Heuvel et al.
Arguably, it is equally important to identify regions or
measures with low heritability as well. The reliability of
imaging measures varies considerably by region or network,
and so does the ability to detect heritability, even if present.
Such information is immensely useful in constraining the
potential phenotypes worth pursuing and interpreting results;
we consider this further below.
N. J. Armstrong
School of Mathematics and Statistics, University of Sydney,
School of Medicine, University of Nottingham,
K. D. Chang
Department of Psychiatry, Stanford University School of Medicine,
Stanford, CA, USA
Aging Research Center, Karolinska Institutet and Stockholm
University, Stockholm, Sweden
Hellen Wills Neuroscience Institute, Brain Imaging Center,
University of California at Berkeley, Berkeley, CA, USA
Department of Psychiatry, University of Missouri, Columbia,
G. D. Pearlson
Departments of Psychiatry and Neurobiology,
Yale University School of Medicine,
New Haven, CT, USA
C. R. Jack Jr
Mayo Clinic, Rochester, MN, USA
E. J. Rose
Transdisciplinary and Translational Prevention Program,
RTI International, Baltimore, MD, USA
R. Z. Goldstein
Departments of Psychiatry and Neuroscience, Icahn School of
Medicine at Mount Sinai,
New York, NY, USA
Department of Psychiatry, University of Halle,
H. J. Bockholt
Advanced Biomedical Informatics Group, llc.,
Iowa City, IA, USA
J. L. Stein
of Medicine, University of California Los Angeles, Los Angeles, CA
S. E. Medland
QIMR Berghofer Medical Research Institute, Quantitative Genetics,
M. E. Renteria I N. G. Martin
QIMR Berghofer Medical Research Institute, Genetic Epidemiology,
M. J. Wright
QIMR Berghofer Medical Research Institute, Neuroimaging Genetics,
C. E. Bearden
Department of Psychology, UCLA, Los Angeles, CA, USA
Department of Pharmacological and Biomolecular Sciences, University of
V. M. Steen
Dr. E. Martens Research Group for Biological Psychiatry, Center for
Department of Psychiatry, UHC University of Vermont, Burlington,
Centre for Healthy Brain Ageing, Psychiatry, University of New South
Wales (UNSW), Sydney, Australia
N. B. Freimer
Center for Neurobehavioral Genetics, Dept. of Psychiatry and Biobehavioral
Sciences, UCLA School of Medicine, Los Angeles, CA, USA
N. S. Lawrence
School of Psychology, University of Exeter, Exeter, UK
Janssen Research & Development, of Johnson & Johnson, Inc.,
1125 Trenton-Harbourton Road; Titusville, NJ, 08560, USA
Brain Imaging and Behavior
Despite the highheritabilityof many brain measures(h2up
to 0.89; Kremen et al. 2009; or even up to 0.96: van Soelen et
al. 2012), the specific genetic variants that contribute to this
variability remain largely unknown. A possible exception is
the Alzheimer’s disease (AD) risk gene, APOE: carriers of
one or more risk-conferring alleles (APOE4) demonstrate
accelerated gray matter loss with age (Lu et al. 2011). They
also have a roughly three-fold increased risk for late-onset
AD, for each risk allele they carry (Corder et al. 1993). In a
recent meta-analysis of 35 prospective cohort studies with an
average follow-up of 2.9 years, the odds ratio for conversion
from mild cognitive impairment to Alzheimer’s dementia in
APOE4 carriers was determined to be 2.29, relative to non-
carriers (Elias-Sonnenschein et al. 2011). Other prior papers
reported a higher odds ratio, around 4 for heterozygotes and
>7 for homozygotes, with some differences depending on the
ancestry of the cohort. According to another more recent
review, one copy of ApoE4 increases risk by ~2.6–3.4, and
homozygotes for ApoE4 have an odds ratio of 14.9 compared
to the reference genotype of E3/3 (Liu et al. 2013).
A number of groups around the world began to perform
GWAS on measures derived from brain images, with the goal
offinding new genetic variants that might account for more of
the variation in brain structure and function, and also for
disease risk. The genetic variants of interest in a GWAS are
variants in the genetic code. SNPs are DNA sequence
variations that occur when a single nucleotide (A, G, C, or
T)isaltered;SNPs are thought tobepoint mutations thatwere
not so damaging that evolution allowed them to be retained in
a significant proportion of the population of a species. Within
a population, SNPs can be assigned a minor allele
frequency—the proportion of chromosomes in the population
behind this approach. Before we discuss GWAS, it is worth
noting a distinction between narrow and broad-sense
heritability: broad-sense heritability is the proportion of
variation in a phenotype (here, individual variations in brain
measures) that can be explained by genetic effects. These
effects may include dominance and epistasis—interactions
between SNPs or genes in different parts of the genome. The
narrow-sense heritability is the proportion of variance in a
brain measure that is accounted for by additive genetic factors
(andthisistypicallya smaller proportionofthe traitvariance).
that GWAS aims to detect.
In a typical GWAS analysis, one might test ~2.5 million
common SNPs in the genome, to see if any of these genetic
variants are associated with a trait, such as a brain-derived
measure, or a specific disease such as AD. Although not the
only important type of genetic variation, SNPs can be
measured using readily available genotyping arrays, and
individually provide adequate statistical power as the variants
are common enough to test their effects statistically. Because
genotyped, many authors have argued that this genotyping
technology is much less expensive than whole-genome
sequencing. However, new technologies using low coverage
sequencing with imputation may in some cases yield several
times the effective sample size of GWAS based on SNP array
data, and a commensurate increase in statistical power as
described in Pasaniuc et al. (2012).
GWAS has had many successes. Many common poly-
morphisms have now been found that increase genetic risk
for AD (Harold et al. 2009; Lambert et al. 2009; Naj et al.
2011), age-related cognitive decline (Davies et al. 2012),
schizophrenia (Almasy et al. 2008; Stefansson et al. 2009;
Ripke et al. 2011; Rietschel et al. 2012), bipolar disorder
(Sklar et al. 2011; Cichon et al. 2011) as well as obesity
(Yang et al. 2012), alcohol drinking (Schumann et al. 2011),
tobacco smoking (Thorgeirsson et al. 2008), cardiovascular
disease (CARDIoGRAMplusC4D Consortium et al. 2013),
osteoporosis (Estrada et al. 2012), prevalent psychiatric
disorders (Cross-Disorder Group of the Psychiatric
Genomics Consortium et al. 2013) and for many other traits
Imaging may play a role in finding out how these genes
create risk for illness through their impact on the brain, by
comparing brain scans of carriers versus non-carriers. One
such example is the ZNF804A story. A variant within
ZNF804A was the first genome-wide significant SNP
associated with risk for schizophrenia (O’Donovan et al.
2008). The function of this variant was initially not clear.
Prominent papers later appeared (e.g., Esslinger et al. 2009)
using imaging to establish disturbed connectivity as a
some variant in ZNF804A (or some variations in linkage
disequilibrium with them) must be functional in the human
brain. This was one of many early studies to validate the
intermediate phenotype strategy in psychiatry.
Ongoing work comparing genome-wide data from patients
with AD and healthy elderly people had begun to unearth a
growing set of new AD risk genes (Bertram 2009). By 2009,
meta-analyses of GWAS from multiple elderly cohorts had
implicated a trio of new AD risk genes—CLU, CR1 and
AD risk by around 10–20 %, consistently, in cohorts around
the world (Logue et al. 2011). Additional AD risk variants
were rapidly discovered as GWAS expanded to more
populations with dementia and healthy controls (Hollingworth
et al. 2011).
A flurry of such studies occurred—some showed brain
differences in Alzheimer’s disease risk gene carriers a full
50 years before AD typically strikes (Braskie et al. 2011;
Bralten et al. 2011). Others showed a pattern of brain changes
in unaffected carriers that resembled the “footprint” of
Brain Imaging and Behavior
al. 2011, Rajagopalan et al. 2013). These findings will require
follow-up but illustrate the potential of using neuroimaging
measures to explore the effects of genetic variation.
But a much more adventurous goal provided the driving
force behind the new and emerging fields of imaging
genomics. This goal was to use neuroimaging data directly,
to screen the genome for common variants that might affect
the brain. In other words, rather than using the images in
secondary studies of what disease risk genes do, images could
be screened to discover important genetic associations.
(Instead of imaging genetics, the somewhat interchangeable
to any method that directly assesses variation in the genome,
as opposed to studies that may assess a single locus only, or
simpler family studies that may not even analyze DNA). The
growing computational power to screen very large
neuroimaging datasets—for the purpose of extracting
meaningful features from them—made this an interesting
and achievable objective. Advocates of “imaging genetics”—
the genetic analysis of brain images (Glahn et al. 2007; Turner
et al. 2006)—suggested that it might even be more efficient to
screen traits derived from brain images to provide
endophenotypes for brain disorders.
The main motivation to screen brain images was to find
some heritable measure of disease burden that might be closer
to the underlying genetic effect than clinical diagnosis based
on cognitive and clinical tests. The endophenotype
hypothesis, long advocated by psychiatric geneticists such as
Irving Gottesman (Gottesman and Gould 2003; Blangero
2004; Goldman 2012; White and Gottesman 2012; Kendler
and Neale 2010) suggested that one might fruitfully apply
genetic screening to any reliable and heritable biomarkers of
a disease—measures from the blood or cerebrospinal fluid
(CSF), or even from brain scans, which by now had become
an illness or disorder (see Gottesman and Gould 2003)
the illness/disorder of interest, (ii) be heritable, (iii) be state-
independent, i.e., seen in people even when they do not show
symptoms of the illness/disorder, (iv) co-segregate with
illness/disorder within families, and (v) be observed in
relatives of affected family members at a higher rate than in
the general population.
The search for endophenotypes of disease, for genetic
analysis, is related to the goal of finding biomarkers for AD
or any psychiatric illness, although the quest for biomarkers
pre-dated efforts to find endophenotypes of disease. In
addition, biomarkers may not be stable, as they may change
during the disease course. The term “biomarkers” has been
used with many different meanings, but in general biomarkers
are measures of disease burden that can be objectively
quantified, ideally allowing more objective or earlier
diagnosis, and making it easier to test the effects of treatment
Advocates of using imaging for genetic analysis pointed to
several advantages that imaging provides now, as well as
several potential advantages that it could provide in the
foreseeable future. First, neuroimaging can yield reproducible
function. Structural measures of the brain, from MRI, tend to
have relatively high reproducibility across measurement
methods, and are generally consistent with expert tracings of
the same structures (see Supplement of Stein et al. 2012; many
Fig. 1 Steps involved in a genome-wide association study. A heritable
state, or continuous, such as the intracranial volume (ICV) - is extracted
from brain imaging scans from a large group of people. To determine if
there is any statistical association between this brain measure and the
inter-subject variations at a single SNP, the genetic variations among
individuals can be assessed at a single location along the genome, and
correlated with differences in the trait of interest (here, ICV). Genome-
wide association scans involve an unbiased search across the whole
genome to discover novel genetic loci associated with the trait. Testing
a million or more SNPs requires a strict multiple comparisons correction
threshold, to avoid reporting spurious results; normally, credible findings
have to achieve a significance value more extreme than p<10−8. The so-
called “Manhattan plot” on the right (by analogy with the Manhattan
skyline in New York) displays the −log10of the p-value for associations
between the brain measure and genetic variation at each position along
the genome; the higher the point on the plot, the more likely it is that an
association exists. Of course, it is important not to see genome-wide
significance as a “binary state”, whose conditions are either fulfilled or
not—but rather a measure of the level of evidence for a genetic
association. Findings in these plots must typically be replicated in several
independent cohorts before they are considered credible or generalizable
Brain Imaging and Behavior
studies have investigated the reliability of measures from brain
MRI, e.g., Pengas et al. 2009). In a recent GWAS analysis,
Holmes et al. (2012) showed high reliability for automated
brain volumes of hippocampus (r=.98), amygdala (r=.91),
and intracranial volume (r=.99) for a cohort of data collected
also studied the inter-scanner reliability of the FIRST software
for segmentation (Patenaude et al. 2011), and found it to be
high. However, it is overly optimistic to always expect high
reproducibility from automated segmentations of brain MRI,
and the reproducibility is region specific. For example, both
FSL and FreeSurfer tend to do less well in segmenting small
in accuracy and reliability among different methods for
automated segmentation of the brain (Shokouhi et al. 2011).
Cortical thickness and other local gray matter density measures
can show reduced reliability owing to sensitivity to image
contrast variability, which becomes particularly challenging in
structures, such as the caudate, may even show systematic
biases in certain populations because of tissue class ambiguity
that arises as a consequence of white matter degradation near
delineate accurately due to the large intersubject variability. As
noted below, one goal of ENIGMA has been to screen brain
measures for reproducibility, heritability, and association with
analysis (we return to this topic below; see also Table 1).
Second, measures of brain volume, integrity, receptor
distribution, or chemical composition, might be more directly
unknown, and known candidate genes—such as growth
factors, transcription factors, guidance molecules, or
neurotransmitters and their transporters. Many of these had
already been implicated in the risk for psychiatric illness, and
imaging offered the opportunity to study differences in brain
connectivity or function, in carriers of genetic variants
associated with disease risk. In the future, advanced MRI
methods such as [1H]MR spectroscopy could be used for
population genetic studies of the brain’s chemical
composition, or PETstudies to measure receptor distribution,
common and widelyavailable types of brain scans—MRI and
DTI (Diffusion Tensor Imaging).
At the sametime, someseasonedgeneticists wereskeptical
about imaging genetics. Concerns were raised about the costs
of image acquisition, relative to a standard psychiatric
diagnostic test. For a study to be feasible, the cost of data
collection must be borne in mind, regardless of which method
is ultimately more efficient in discovering genes and
mechanisms contributing to brain disease.
By 2009, a number of GWAS studies had been performed
on neuroimaging data. Among the first studies to report a
positive finding—a so-called “genome-wide hit”—was the
report by (Potkin et al. 2009a) that identified a key genetic
variant in TOMM40, in linkage disequilibrium with APOE,
the known risk gene for AD. Using hippocampal atrophy as a
quantitative phenotype in a genome-wide scan, they assessed
381 participants in the ADNI (Alzheimer’s Disease
Neuroimaging Initiative) study, to identify SNPs for which
there was an interaction between the genotype and diagnosis
on the quantitative trait. Variants in TOMM40 appeared to
affect hippocampal volume differently in AD patients versus
controls. Working with genetic-founder populations, some
GWAS-based studies revealed genome-wide hits in relatively
small samples; in the Saguenay Youth Study, genetic
variations in the KCTD8 region were associated with brain
size in a community-based sample of adolescent girls
(rs716890, P=5.40×10−9); (Paus et al. 2012). Furthermore,
genotype in the top hit (rs716890) interacted with prenatal
exposure to maternal cigarette smoking vis-à-vis cortical area
and cortical folding; in exposed girls only, this genotype
explained ~21 % of variance in the cortical area (Paus et al.
The ADNI work analyzed a publicly available dataset
(adni.loni.usc.edu) for which MRI scans and GWAS data
could be freely downloaded from approximately 800 people.
The ease of access of the ADNI data, and the principle of
broad dissemination of the genomic data with fairly few
restrictions led to over 100 genetic studies of the ADNI data,
many of them using GWAS designs (Saykin et al. 2010).
The ADNI dataset stimulated work in imaging genetics, as
evidenced by the large number of published papers using the
data (all ADNI genetics studies from the period 2009–2012,
are reviewed by Shen et al., 2013; this volume). Even so, data
from other elderly cohorts is needed to determine how well
ADNI’s genetic findings generalize to other non-selected
cohorts. ADNI deliberately selected participants who are
predominantly Caucasian, and relatively free from drug or
alcohol abuse and vascular disease. Genetic associations
detected in ADNI may be stronger or weaker, or even not
supported at all, in cohorts with different ethnic composition,
or with other co-morbid health conditions. Because of this, it
is important to recognize that some genetic associations may
depend on the cohort studied, diagnostic or demographic
criteria, and we should not expect all true genetic associations
to be detectable in all cohorts.
Sample sizes and power for GWAS
As with other phenotypes, the first GWAS findings for
imaging phenotypes were difficult to replicate, probably due
to the limited availability of replication samples. Prior work
had cautioned that, using the most standard kind of univariate
analysis of SNP effects, very large sample sizes would be
required—far larger than a typical neuroimaging study—to
Brain Imaging and Behavior
discover influential genetic variants, unless their effect sizes
on the brain phenotypes being analyzed were unusually large
the number of null genetic variants assessed, the expected
effect size of the genetic variant (typically less than 1 % of
the trait variance), and the population frequency of the variant
(Potkin et al. 2009b; Flint et al. 2010; Paus et al. 2012). The
power obviously also depends on the genetic architecture of
the trait—key factors are the large number of effectively
independent LD regions in the genome (regions of linkage
disequilibrium, with correlated SNPs) and the number of
detectable causal loci per phenotype (which is typically
small). The first series of studies identifying genetic markers
for MRI phenotypes that replicated across independent
samples, used a Norwegian discovery sample and replicated
the findings in the ADNI or PING cohorts (Joyner et al. 2009;
Rimol et al. 2010; Bakken et al. 2012).
Even for high frequency variants, neuroimaging databases
observed effect sizes. We must bear in mind that, in the
standard experimental design (see below for others), the
significance of a genome-wide hit has to be at least 20 million
to one, to account for the very large number of variants tested.
Some have argued that neuroimaging studies reporting effects
of candidate genes—such as COMT or BDNF—are also at
risk for false positive effects, in the sense that any number of
genes could have been assessed, with no way to verify
whether the paper was selectively reporting the successes
(Flint and Munafo 2013; Ioannidis 2005). While this problem
is shared by selective reporting of results in many fields,
imaging genetics is particularly at risk because of the ease
of retesting the same data. Some have raised the concern
that a considerable proportion of neuroimaging genetics
associations, especially those found in small samples, may
Table 1 Selection of brain measures for genetic analysis. In ENIGMA,
the brain measures chosen for analysis had to be feasible to measure
consistentlyand efficiently at a largenumber ofsites, accordingtoagreed
protocols (available at enigma.ini.usc.edu). As power is limited in
GWAS, various tactics may be useful in the future to boost the power
to find genetic associations. Some of these are categorized here
Power enhancement approachPrinciple Pros and cons
1. Enhance the datasetIncrease the sample size (some GWAS studies now
assess 100,000+ subjects (Lango Allen et al. 2010;
Speliotes et al. 2010))
Increase genomic coverage/sequencing
Identifies variants with smaller effect sizes, but is more
power to detect effects of low frequency variants. Also
is more costly than genotyping common SNPs through
genotyping arrays, but the cost is rapidly decreasing
need to correct for the number of measures assessed,
which may be large (e.g., in “voxel-based” GWAS;
Stein et al. 2010; Ge et al. 2012); if too many are
assessed, power is low
Avoids heavy statistical correction, but may miss
unexpected variants or phenotypes
This may empower genomic screens of complex
phenotypes (e.g., genome-wide connectome-wide
screens; Jahanshad et al. 2013b), see ENIGMA-DTI
(Jahanshad et al. 2013a).
GWAS on the resulting “genetic clusters” appears to
have higher power than standard voxel-based
approaches (Chiang et al. 2011, 2012; Chen et al.
Balances the strength of the genetic signal for the
endophenotype and the strength of its relation to the
disorder of interest
Difficulttoapplyto distributedremote datasets formeta-
relevance of the signal
Increase the range of phenotypes studied
2. Data reduction Focus on candidate SNPs/genes, candidate pathways,
Heritability screening—remove or de-emphasise
measures with low heritability
2.1. Based on classical genetics
Genetic Clustering—find parts of an image or 3D
cortical surface with common underlying genetic
2.2. Based on relevance to disease Endophenotype ranking value (ERV; Glahn et al. 2012),
aims to rank biomarkers in terms of their promise as
endophenotypes for any heritable illness
2.3. Based on using multivariate
Meda et al. 2012);
Sparse regression (Vounou et al. 2010, 2012; Ge et al.
2012; Silver et al. 2012), compressive sensing,
parallel ICA, machine learning methods
3. Multimodality approaches
imaging modalities or other
biomarkers at the same time
Use multiple predictors in the genome or image or both
Multimodal data fusion using ICA (Calhoun et al. 2009)
Seemingly Unrelated Regression to pool information
across simultaneous models (SUR; Jahanshad et al.
More work required to analyse multiple data modalities
at once (e.g., anatomical MRI and DTI)
Brain Imaging and Behavior
not replicate in subsequent analysis. Clearly, although the
number of genes in the human genome is limited, an almost
unlimited number of candidate genes could be still tested
(Bishop 2013; Flint and Munafo 2013).
Even so, others have argued that this might be an obvious
but not in itself sufficient argument for needing very large
sample sizes, as it does not take into account the increase in
effect sizes enabled by careful selection of phenotypes (this is
an area of neuroimaging research in itself; see Table 1),
nor corroborative evidence from other sets of data, nor
independent replication of top hits—which is commonly
performed in GWAS studies. Clearly, an alternative way to
for a variant (the “convergent approach”). A final approach is
to focus on candidate variants with known molecular function,
andthere is a long tradition ofwork relating genetic variants in
the monoamine neurotransmitter pathways to risk for
psychiatric illness. For instance, lower hippocampal but larger
amygdala volumes have been associated with the long variant
of the serotonin transporter polymorphism inmajor depression
(Frodl et al. 2004, 2008), and variants in the 5-HT1Areceptor
gene have been related to amygdala volume in borderline
personality disorder (Zetzsche et al. 2008). Variants in the
Huntington’s disease gene have also been shown to affect
normal brain structure (Mühlau et al. 2012).
As the field began to come to terms with the sample sizes
needed to demonstrate or replicate a genome-wide hit,
alliances began to form and some studies reported genetic
associations with volumetric brain measures supported by
evidence from more than one cohort. Stein et al. (2011)
reported a variant associated with caudate volume in young
and old cohorts scanned on two continents at two different
field strengths. Hibar et al. (2013b) also reported a variant
associated with the volume of the lentiform nucleus on MRI
scans. In some of these reports, a meta-analysis approach was
used, to combine the evidence for genetic association across
cohorts, in a way that weights the cohorts by their sample size
or error variance.
Formation of the ENIGMA Consortium and first project
(ENIGMA1; Stein et al. 2012)
neuroimaging or large-scale genetics studies formed a
network called “Enhancing NeuroImaging Genetics through
Meta-Analysis” or ENIGMA. The goal of the effort was to
(see Fig. 2).
In Stein et al. (2012), the ENIGMA consortium, in
collaboration with another multi-site consortium (CHARGE;
Seshadri et al. 2010; Fornage et al. 2011; Bis et al. 2012;
Ikram et al. 2012) reported that the mean bilateral volume of
the hippocampus was significantly associated with the
intergenic variant rs7294919 (see Forest plots, Fig. 3).
The genotype at that locus was also nominally associated
with expression levels of the positional candidate gene TESC
in adult human temporal lobe tissue, based on the UCL
database of brain tissue resected during epilepsy surgeries.
TESC is currently not well studied, but it is known that it is
expressed during brain development in mice and chickens
(Bao et al. 2009). It encodes tescalcin, which interacts with
the Na+/H+exchanger (NHE1) (Baumgartner et al. 2004),
important in the regulation of intracellular pH, cell volume
and cytoskeletalorganization. In addition, intracranialvolume
HMGA2 gene that had been previously tied to height. The
CHARGE consortium had focused on elderly cohorts
recruited primarily for studies of cardiovascular health in old
age, whereas ENIGMA had aggregated data from cohorts
across the lifespan showing the genetic effect was strong,
happening over multiple stages of brain growth. Both efforts
had strengths—the number of cohorts in ENIGMA (23) was
larger than that of CHARGE, but some of the constituent
cohorts in CHARGE numbered over 2,800 individuals.
Practical considerations and opportunities
The first ENIGMA project demonstrated the feasibility of
discovering statistically significant effects of “single letter”
genomic differences in brain data worldwide, using imaging
and genetic data collected using diverse protocols. Novel,
harmonized data analysis and meta-analysis protocols were
vital to the success of this project.
First, all genotypic data was imputed to the HapMap3
reference dataset to correct for diversity of the genotyping
chips (this reference was subsequently updated to the 1,000
Genomes dataset for ENIGMA2, which began in May 2012).
The imputation protocols (detailed at enigma.ini.usc.edu) and
the standard reference datasets allow the consistent reporting
of genotypes at the same set of genetic loci across cohorts.
Imputation effectively adds prior knowledge to the data and
may in this way also increase the power of a study. Not all
Mexican-Americans and the NIMH cohort contains a
significant number of African-Americans. As in any GWAS,
population structure is taken into account during the statistical
modeling of associations, to ensure that differences in SNP
frequencies with ancestry are not picked up as spurious
associations. Also in the imputation step, the appropriate
reference populations are used for each individual, which
may in some cases be Yoruban or Hispanic, as well as the
CEU population that is used to represent Caucasians.
However, it is not a computationally trivial task to impute
very large samples of data from so many subjects. The
successive refinement of reference panels means that re-
Brain Imaging and Behavior
imputing the same dataset to the most current standards is
likely to further boost power.
Second, the exchange of data between the ENIGMA and
CHARGEconsortia,forthe Steinetal.(2012) study,involved
a reciprocal look-up rather than an exchange of the full meta-
analyzed data across the entire genome. This effort found that
the top hits of both consortia were in fact the same ones—
providing extremely high credibility to the hits. Even so, the
full genomic data were not exchanged or meta-analyzed; this
more powerful effort is currently underway.
Third, the effort to harmonize the analysis of imaging data
involved development of new quality control procedures.
Considerable time was devoted to checking outliers, testing
volumes, and checking the allele frequencies, genomic
inflation factors, and other statistical summaries of the cohort
data. The genetic and MRI analysis protocols used in
structures/ and the DTI analysis protocols are available at
In a project of such a scale, the need for adequate data
curation is paramount. Several measures were assessed to
determine whether structureshad values inthe expected range
for volume, hemispheric asymmetry, etc. In the end, the large
number of contributing authors and contributing sites made it
easier to identify sites that had outliers in their genotypic or
brain measurement data, partly because normative data were
becoming available from all the other sites. All relevant
population structure that might lead to ancestry differences
masquerading as true genetic associations to brain traits.
Among these tests, we demonstrated that different, widely-
used programs to measure regional brain volumes produced
consistent results on a broad range of cohorts worldwide,
based on data from young and old samples and mixtures of
both (Stein et al. 2012; see supplement comparing FSL and
FreeSurfer segmentations). It was not feasible to require all
sites to use one specific program to measure brain regions; as
might be predicted, no single algorithm performed best for
quantifying brain volumes across all datasets.
Fourth, ENIGMA did not use the “mega-analysis” model
adopted by the Psychiatric Genomics Consortium, where all
phenotypic and genotype data are sent to a centralized site for
analysis. Instead, ENIGMA uses a “meta-analysis” concept:
GWAS was run locally using pre-agreed covariates—often
with and without adjustments that may affect interpretation
(e.g., adjustments for overall head size). Subsequently, and
after quality control at each site, the p-values and regression
coefficients were combined and meta-analyzed in a way that
weights the results based on the sample sizes of each
Many cohorts within ENIGMA have restrictions on data
use and data access. Some preclude the sending of data out of
Fig. 2 ENIGMA founding sites. The first ENIGMA project (Stein et al.
2012) was initiated in 2009, by a consortium of research groups
worldwide involved in neuroimaging and genetics. Several existing
consortia and research networks are taking part, including IMAGEN,
EPIGEN, SYS, FBIRN, and ADNI. Many of these efforts pre-dated
ENIGMA and continue today; each conducts its own projects in addition
to their collaborative work within ENIGMA. ADNI collects data at 58
sites around the U.S.; for clarity, not all data collection sites are shown
here. Each symbol represents a site contributing to ENIGMA, as
of June 2013
Brain Imaging and Behavior
the lab where the data were collected. Some restrict the
sending of personally identifying information to any other
site. A founding goal of ENIGMA was to not require cohort
data to be shared outside the center that collected it, to avoid
creating ethical and legal issues for the study sites. Although
data sharing is a laudable goal in science, a more pragmatic
analyze summaries of the resulting data that lack personally
identifying information. This also encourages maximum
participation as each site retains fiduciary responsibility for
growing perception of the community that data sharing is one
fundamental building block of reproducibility in science, and
a rapidly expanding number of imaging datasets are being
data acquisition are being developed concurrently (Poline et al.
However, there are also disadvantages associated with the
perform meta-analyses for more complex and potentially
more informative analyses such as (1) polygenic scoring,
which determines how much of the phenotypic variance can
be explained by common SNPs in aggregate (Purcell et al.
2009), (2) structural equation modeling to demonstrate
evidence for the endophenotype concept (Kendler and Neale
2010), and (3) stepwise linear or ridge regression, to identify
the causal variant at an associated locus (McCarthy et al.
2008; McCarthy and Hirschhorn 2008). Not being able to
perform interaction analyses (e.g., disease by SNP) is also a
limitation, although this could be overcome as ENIGMA’s
disease related working groups expand.
In the aftermath of ENIGMA1, several investigators
requested access to the GWAS meta-analyzed results, and
the meta-analyzed data were made available online through
an interactive website named ENIGMA-Vis (Novak et al.
2012; http://enigma.ini.usc.edu/enigma-vis/). This interface
allows a user to input any gene (or SNP) that they are
interested in, and query its effects on a wide variety of brain
measures. The current dataset also allows targeted studies of
individual genes, enabling research groups, even if they are
Fig. 3 Forest plots from the ENIGMA1 study (adapted from Stein et al.
2012). Forest plots are a graphical display designed to illustrate the
relative strength of an effect in different cohorts. In the left panel, we
show the effect of the genetic variant at rs7294919 on the hippocampal
volume, in a range of cohorts in ENIGMA. In ADNI, for example, the
confidence interval on the effect overlaps zero, which means that there is
no evidence to reject the hypothesis of no effect, if only that cohort were
considered. The “ENIGMA Discovery” line combines the effects of all
cohorts above it. At the bottom of the figure, the meta-analysis of all
effects above the line includes data from another large consortium,
CHARGE, and several replication samples. The area of each square is
proportional to the study’s weight in the meta-analysis. The right panel
shows a similar plot for the effect on intracranial volume of the common
genetic variant at rs10784502. It is not necessary for the effect to be
detected in all cohorts for the meta-analysis to support the effect. The
abbreviations denote the names of the different cohorts in ENIGMA
(pleasesee Stein et al.2012, fordetails). [Adapted, with permission, from
Stein et al., Nature Genetics, April 15 2012]
Brain Imaging and Behavior
not involved in human studies, to assess the possible impact
on the brain of genes they are studying (via the ENIGMA-Vis
In one study, Bulayeva and colleagues (Bulayeva et al.
2013) performed a linkage analysis of psychosis and mental
retardation in population isolates in remote areas of Dagestan
and Chechnya. They were able to implicate the top hits in
ENIGMA1 in these illnesses, supporting a psychiatric effect
of the genetic variant in human populations. Erk et al. (2013)
also studied rs7294919—ENIGMA1’s top hit for association
with hippocampal volume, and found that it related to
behavioral differences and different patterns of brain
activation in memory tasks. Work by Ming Li and colleagues
(Lietal.2013) alsofoundSNPsinthe candidate geneCREB1
of which was rs6785) and were also associated with measures
of hippocampal volume and function.
Current projects of the ENIGMA Consortium
ENIGMA2 is a follow-on study from ENIGMA1, in which
genome-wide association analysis (Hibar et al. 2013a). These
volumes are known to be moderately to highly heritable, with
thalamus (0.80) and caudate nucleus (0.88) and lowest for the
left nucleus accumbens (0.44; den Braber et al. 2013). The
structures are also implicated in a wide range of psychiatric
and degenerative brain disorders, making it crucial to identify
genetic and environmental factors that may influence them.
Preliminary meta-analysis is underway which will be
followed by functional characterization of the key hits, with
a full report to be submitted soon. A range of functional
evaluations is envisaged over the short and long term, ranging
from in silico assessments of eQTL data (for instance), all the
way to new functional experiments in neuronal cell lines or
animal models. This kind of follow up study is crucial, given
the difficulties of obtaining robust functional variants from
ENIGMA-CHARGE genome-wide meta-analysis
Researchers from the ENIGMA and CHARGE consortia have
recently joined forces to perform a genome-wide meta-analysis
of hippocampal and intracranial volume using updated versions
and ENIGMA will be meta-analyzing GWAS results from the
much more densely sampled 1,000 Genomes reference set;
data is being meta-analyzed. In addition, the ENIGMA project
will be contributing updated results from many new samples
who have joined the effort since the completion of the
ENIGMA1 project. This analysis is currently ongoing.
ENIGMA-DTI is a Working Group developing a
harmonized protocol for analyzing DTI data for GWAS
meta-analysis (Kochunov et al. 2012). Diffusion tensor
imaging offers a range of measures that reflect the
microstructure of both white and gray matter (by probing the
diffusion profile of the water molecules). It is also possible to
reconstruct macroscopic structures such as tracts, using the
tensor’s directional information. Understanding the genetic
factors underlying the connections of the brain is one of the
especially true since the inherent plasticity of the developing
brain allows for the remodeling of connectivity based on
environmental influences. Interestingly though, heredity of
diffusion weighted measures remains quite high, suggesting
that factors related to brain growth, development, and
plasticity are also highly heritable (White et al. 2013).
A major theme in prioritizing brain phenotypes for large-
scale genetic analyses is the presence of significant
heritability, indicating that a proportion of individual variance
in the phenotypes can be explained by genetic variation. The
effect size of individual variants cannot be inferred from the
heritability and higher heritability does not translate to a
higher likelihood of a positive GWAS finding. However,
phenotypes whose heritability is not significantly different
from zero may not be good candidates for GWAS analysis
because of the lack of variance due to additive genetic factors.
Jahanshad et al. (2013a) screened a number of regions of
interest across multiple cohorts, finding high heritability for
most major white matter pathways and consistent heritability
across pedigree-based and twin-based samples (Fig. 4;
Kochunov et al. 2012; Jahanshad et al. 2013a). A variety of
DTI parameters could be measured reliably among
individuals, and genetic factors explained a substantial
proportion of the variance for the most commonly used DTI
Penke et al. (2010) showed strong correlations among
measures of DTI-derived fractional anisotropy (FA) from a
range of major white matter tracts in the Lothian Birth Cohort
1936 (LBC1936), a group of community-dwelling subjects in
their seventies. Applying principal component analysis, a
general white matter integrity factor was found that explained
around 45 % of the individual differences in FA across eight
tracts, and was significantly correlated with processing speed.
Lopez et al. (2012) conducted a GWAS on this general white
matter integrity factor on 535 subjects of the LBC1936 study
and found suggestive genome-wide association with SNPs in
ADAMTS18 and LOC388630. Initial studies suggest that a
proportion of the variance in fiber integrity can be predicted
from common variants (Kohannim et al. 2012; Jahanshad et
al. 2012, 2013c; Thompson and Jahanshad 2012; Braskie et
al. 2012; Sprooten et al. 2013).
Genome-wide screens of connectome data have also been
Brain Imaging and Behavior
by their heritability, prior to entering the genome-wide
screening phase (Jahanshad et al. 2013b; Thompson et al.
2013). ENIGMA has not yet attempted a multi-site genetic
study of brain connectivity. Agreement on a protocol depends
on ongoing harmonization of DTI analysis across ENIGMA
sites, which is being finalized by the ENIGMA-DTI working
group (Jahanshad et al. 2013a). The harmonized DTI analysis
will also allow a multi-cohort examination of the association
between white matter integrity and general cognitive ability
Penke et al. 2012). Also underway is a harmonization of
cortical segmentation, which is being studied empirically by
the ENIGMAworking groups that focus on psychiatric illness
ENIGMA Disease Working Groups
An implicit goal of ENIGMA is to see whether genetic
variants impact the brain in a way that affects disease risk. In
fact, a number of imaging genetics papers have identified
genetic variants that appear to affect the brain and behavior;
Halletal. (2006) and McIntoshetal. (2008) reporteda variant
in the neuregulin 1 gene that was associated with abnormal
cortical function, altered white matter integrity, and with
psychosis,and there are manyother examplesof variantswith
Around a third of the data in ENIGMA is from patients with
psychiatric illness, so once ENIGMA1 was complete, a large
volume of new data couldbe brought tobear onthe questionof
braindifferences ina varietyof disorders.Tomake connections
sense to prioritize brain measures with robust case–control
differences. Of course, robust case–control differences do not,
in themselves, imply that the same genetic variants influencing
the phenotype will be the same as those associated with disease
risk. A more informed way to rank brain measures for genetic
screening is to use the endophenotype ranking value (ERV;
Glahn et al. 2012), which aims to rank biomarkers in terms of
their promise as endophenotypes for any heritable illness.
The ERV balances the strength of the genetic signal for the
endophenotype and the strength of its relation to the disorder
of the illness (hi2), the square-root of the heritability of the
endophenotype (he2), and their genetic correlation (ρg):
In schizophrenia, for example, decades of studies have
reported morphometric differences in patients versus controls,
framework to rank brain measures in order of their effect sizes
for case–control differences; these effects could also be further
weighted based on their genetic correlation with the illness, to
give anotherranking. Nor is it expected thatthe disease should
have identical effects on the brain in all cohorts; the variety of
participating cohorts in ENIGMA makes it possible to dig
deeper into medication-related, or even geographic or
demographic factors to explain why brain differences vary so
drastically across different studies (see Fig. 5 for the locations
of ENIGMA sites in disease-related working groups).
In a meta-analysis of data from 1,136 patients and 1,401
controls, the ENIGMA-Schizophrenia Working Group found
that hippocampal volume gave among the highest effect sizes
for any subcortical brain structuraldifference inschizophrenia
% heritable (100*h2)
Fig. 4 A meta-analysis of tract-wise heritability, by the ENIGMA-DTI
working group, showed most tracts in the brain are moderately to highly
heritable across cohorts of different ethnicities, even though they were
imaged with different parameters. The “skeleton” of the white matter,
reconstructed using a widely used DTI analysis program called “tract-
based spatial statistics” (TBSS; Smith et al. 2006), is shown in purple for
reference. The corticospinal tract (in light blue) was the least heritable
in an initial GWAS of DTI-FA measures. Other methods for phenotype
selection and prioritization are summarized in Table 1
Brain Imaging and Behavior
(Turner et al. 2013; van Erp et al. 2013). In a related meta-
analysis of structural MRI data from 1,022 patients and 1,415
controls, the ENIGMA-Bipolar Disorder Working Group
found consistent differences across the subcortical regions,
but in a different pattern than that characteristic of
schizophrenia (Hibar et al. 2013a). There were significant
reductions in the bilateral thalamus, hippocampus, and
amygdala in patients diagnosed with bipolar disorder. In
general, the trend revealed a decrease in subcortical volumes
throughout the brain in patients with bipolar disorder. This
work is clinically important, as the alterations in limbic and
some cortical regions are thought to underlie some of the
affective symptoms in bipolar disorder; even so, the source
of many of the subcortical and cortical differences in the
disorders has beenamatterofdebate,andformanystructures,
morphometric findings have not always been consistent. This
work is still ongoing, with a total sample of 4,729 subjects
as of November 2013).
In addition, following the model of the Schizophrenia
and Bipolar Disorder and working groups, an ENIGMA-
Major Depressive Disorder (MDD) Working group was
recently initiated. It already includes structural MRI data
An additional working group on attention deficit/
hyperactivity disorder is currently being formed, ENIGMA-
ADHD, which will analyze data from more than 1,500 cases
(children and adults) with the disorder and over 2,000 controls.
An even more recent extension is a developing ENIGMA-
Addiction working group that will analyze data relevant to
addiction phenotypes including case-control comparisons
across a variety of substances, and going beyond this to
examine the influence of comorbidities, gender and stages of
disorder. The ENIGMA meta-analytic approach will be used
to aggregate data from case-control and developmental
cohorts to examine the relative contribution of various genetic
and brain correlates to risk for early onset substance misuse,
transition to regular use, susceptibility to dependence, and
individual differences in relapse vulnerability. We will also
examine variants common to all addictive behaviors, and
attempt to identify those that might be specific to
homogeneous addiction profiles. To date, a pooled sample
of 8,000 participants with neuroimaging, genetics and
addiction phenotypes has been identified; their data will be
re-analyzed to address these questions.
Clearly, these studies aggregate data from cohorts that are
heterogeneous in terms of duration of illness, disease etiology,
medication history, demographics, exposure to potentially
neuroprotective substances such as lithium, and many other
Fig. 5 Locations of the ENIGMA Working Groups. After ENIGMA’s
firstproject was completed (ENIGMA1; Stein et al.2012), largeamounts
of brain imaging data had been analyzed from patients with a variety of
psychiatric disorders. Working groups (WGs) were formed to understand
the effects on the brain of bipolar disorder, major depressive disorder
(MDD), schizophrenia, by pooling and comparing data from many
neuroimaging centers. These groups are open to any researchers who
have collected MRI scans from patients with these illnesses. No genetic
data is needed to join. In fact, most projects study factors that might
influence how these disorders affect the brain—medications, geographic
factors, and the age and gender of the patient. A further working group
focuses on diffusion tensor imaging, which assesses white matter
integrity; current projects relate DTI measures to individual differences
incognition andgenetic make-up.Theinstitutionsintheworkinggroups,
as of June 2013, are shown on the map (see color key, inset)
Brain Imaging and Behavior
factors. But part of the richness of the ENIGMA efforts is that
they afford sufficient power to begin to see which of these
factors—including geographic factors, perhaps—affect disease
expression in the brain and the universality,or otherwise,of the
biomarkers of brain disease. Based on these findings, a key
direction for ENIGMA is to see how genes that affect the brain
relate to genes that affect risk for psychiatric illness (identified
to identify shared biology between brain characteristics and
disease. A collaboration between ENIGMA and the PGC is
now underway to study these relationships.
One limitation of ENIGMA to date is that the sites
contributing to the meta-analysis include a clinically somewhat
heterogeneous group of population-based studies, and case–
control cohorts from multiple psychiatric and neurological
diseases. ENIGMA’s working groups on psychiatric disorders
are meta-analyzing representative datasets from schizophrenia,
bipolar, depressed, ADHD, autism, OCD, 22q deletion
syndrome, HIV, addiction, and other cohorts worldwide.
Eventually, cross-disorder comparisons should be able to
and how they depend on genetic variation, and diagnostic
criteria, and in duration of illness, medication history, ethnicity,
and in other demographics. Some studies include patients with
co-morbid conditions deliberately excluded by other studies.
The diversity of psychiatric cohorts in ENIGMA suggests a
second wave of analyses to understand cohort-specific factors
that might account for, or contribute to, the heterogeneity in
results across sites. Recent work by the Psychiatric Genomics
Consortium Cross-Disorders working group has identified
considerable genetic overlap between several major disorders,
at the level of common genetic variants (Lee et al. 2013).
ENIGMA may be able to do the same from a neuroimaging
perspective, to determine if genetic factors implicated in
different disorders account for some of the cross-disorder
differences in the brain imaging meta-analyses.
Future directions and caveats
purpose. The fact that so many investigators are actively
involved makes it possible to benefit from the combined
resources and talents of all participants for “crowd-sourcing”
discovery. Also, the sample sizes involved—unprecedented
for a neuroimaging study—alleviate some of the concerns
about underpowered studies and unreliable findings (Button
et al. 2013). In addition, apart from identifying genetic
variants, another important role for the ENIGMA consortium
is to help understand how GWAS-derived genetic variants for
behavioral phenotypes influence the brain. Exploring the
effects of disease risk alleles on brain measures can help us
understand the brain systems affected, and at which stage—
and also whether the effects are pervasive or selective
(de Geus 2010).
Much of this overview of the history and future efforts of
ENIGMA highlights its relevance to studies of disease,
focusing on psychiatric and neurodegenerative disorders such
as Alzheimer’s disease. Even so, only around a third of the
ENIGMA data comes from patients with psychiatric illness,
and much can be learned about the genetic factors that drive
normal variation in the general population. A great deal of
fundamental information on the biology of the human brain
can be discovered from efforts such as ENIGMA, irrespective
of whether it has a direct relevance to any specific disease.
There are limitations to a study like ENIGMA despite its
strengths. The first is that many other types of genetic or
epigenetic variation other than GWAS are important—rare
variants, CNVs, expression and methylation analyses are all
crucial; they simply have not yet been evaluated through
ENIGMA, but that is likely to change in the future. In recent
genome-wide complex trait analyses (“GCTA” analyses;
Yang et al. 2011; Lee et al. 2011), Wray, Visscher and their
colleagues have shown that GWAS data may account for a
when the individual predictive value of a given locus or SNP
Asa basic principle ofgenetics, anoverall largeheritability
does not guarantee locus specific heritability, but recent
discoveries have surprised some geneticists in supporting the
that accounted for about 5 % of the variance in height, large
studies have now demonstrated SNPs can account for around
45 % of the variance in height (for which the overall
heritability is around 80 %). Also, common causal variants
may account for around 23 % of the risk for schizophrenia
(Lee et al. 2013) and up to 60 % of the risk for autism (Klei et
al. 2012). Given the polygenic architecture of these disorders,
these results suggest that more individual SNP associations
will be detected for each disorder, as sample sizes increase.
Some authors have emphasized that similar kinds of
“polygene” scores to the ones used to predict illness risk can
be used to explain variation in related phenotypes, such as
cognition (Davies et al. 2011), structural MRI (Holmes et al.
2012), neural activation on functional MRI (Whalley et al.
2013a) and white matter integrity (Whalley et al. 2013b).
Even so, some geneticists argue that rarer variants are in
some cases more worthy of study than GWAS as they tend to
have larger effects that are more easily validated functionally.
GWAS has been successful when used for QTL localization,
but the over-arching goal of complex disease genetics is to
identify the causal genes and functional variants responsible
for phenotypic variation. So some have argued that GWAS is
useful for detecting the signal from common variants, but that
few GWAS results have turned into functional variants or
Brain Imaging and Behavior
genes (across all disease domains). In principle, ENIGMA
could be used to study rare variants as well, but a range of
complementary approaches are always necessary.
Second, the geographical diversity of ENIGMA is higher
than that of most neuroimaging studies, but there may be
the generalization of results to all human populations. This
limitation is not specific to ENIGMA, and its consortium
structure could be expanded geographically to assess ethnic
differences in factors associated with disease and their
relationship to brain structure. Ethnic differences in genetic
effects are well known: the meta-analysis of the effects of the
APOE risk gene for Alzheimer’s disease shows a difference
between European and Asian populations. So far, people in
Africa and Asia are under-represented on the ENIGMA map;
we aretherefore keentoinclude samples from thesecontinents.
Third, the need for multivariate analysis has long been
known in the neuroimaging field—no serious neuroscientist
would predict diagnosis from a single pixel or voxel in an
image—the multivariate pattern of signals is paramount.
Methods to access and recover the maximum amount of
pertinent information in an image dataset are only just in their
infancy in the fields of genetics and neuroimaging—partially
due to the lack of sufficient data to test competing methods,
until recently (Liu et al. 2009; Stein et al. 2010; Hibar et al.
2011b; Vounou etal. 2010; Le Floch et al. 2012; Rosenblatt et
al. 2013; Meda et al. 2010, 2012).
Fourth, the discovery of a GWAS hit—in ENIGMA or any
other GWAS study—is the beginning of a long road of
discovery, especially if the finding is intergenic or in a gene
ofunknownfunction. Some genomic screens ofanatomicalor
structural connectivity data have implicated genes such as
SPON1 (Jahanshad et al. 2013b) and FRMD6 (Ryles et al.
2012) that were discovered in later case–control studies to be
risk genes for AD (Hong et al. 2012; Sherva et al. 2013).
Functional validationofgenetic variantsreliablyimplicatedin
large scale studies will be the way we learn new biological
processes and further our understanding of risk for psychiatric
Fifth, ENIGMA has started by analyzing those phenotypes
that are easily measured in a standardized way. Brain images
can be analyzed in more sophisticated ways than traditional
morphometric methods; shape analysis, for example, has long
been used to characterize features of brain structure, including
cortical complexity, curvature, fractal dimension, spectral
content, and other indices. Also, the reporting of regional
summaries from DTI data clearly does not exploit all of the
available information in the data—DTI can be analyzed using
automatic whole-brain tractography to reveal the brain’s fiber
patterns and measure connectivity, and even the topology of
these brain networks. Despite the ever-expanding landscape
of brain features that can be studied, the large samples
required for genetic analysis have motivated the study of the
simplest brain measures first. Undoubtedly the future will
hold large scale genetic studies of brain connectivity,
anatomical shapes, functional networks, and features that are
as yet unknown and undiscovered.
As a final limitation, ENIGMA analyses have been
cross-sectional to date, rather than longitudinal. Genetic
predictors of brain changes over time have substantial
importance for clinical trial enrichment (e.g., TREM2,
which harbors variants that predict rates of brain atrophy
over time; Rajagopalan et al. 2013; see Kohannim et al.
2013, for an approach to boost power in drug trials by
multi-locus genetic profiling; see Andreasen et al. 2012,
for an approach examining epistatic relationships
between multiple genes and progressive brain tissue loss
Are fewer subjects needed with imaging?
An open question is whether GWAS meta-analyses really do
require fewer subjects with imaging than they do when
behavior is the target of study. In 2009, before ENIGMA
are expensive to collect does not change the power
calculations.” (N. Martin, pers. commun., 2009). But more
recently, Rose and Donohoe (2013) performed an empirical
analysis of effect sizes in genetic studies of cognitive and
neuroimaging traits in schizophrenia, and found evidence
supporting the efficiency of using imaging traits. However,
some evidence does suggest that imaging traits may have
intermediate effect sizes when compared to phenotypes
theoretically closer or farther away from the underlying
biology. The percent variance explained in gene expression
GWASs (often called eQTL studies) for the top SNP hits are
well above 10 % of the variance in the expression of a
particular gene (Stranger et al. 2007). The percentage of trait
variance in hippocampal volume explained by the top genetic
variant in ENIGMA1 was 0.27 % (Stein et al. 2012) although
independent cohorts willbe requiredto estimate the explained
variance in the population at large. Finally, the top hit in one
genome-wide analysis of traits correlated with cognition—
such as educational attainment—explained 0.02 % of the
variance (Rietveld et al. 2013).
It is interesting to compare the effect of ENIGMA’s top
(Barnes et al.,2009), patientswith Alzheimer’s disease have a
24.1 % mean hippocampal volume deficit. Patients with mild
cognitive impairment have a 15.3 % deficit relative to
cognitively healthy controls (calculated from Table 9 in
Leung et al. 2010). In psychiatric disorders, the hippocampal
volume deficit is typically reported to be smaller. Although
hippocampal volume deficits are one of the more robust MRI
findings in schizophrenia, they are not always evident in
Brain Imaging and Behavior
single site studies (Shenton et al. 2001). The ENIGMA
Disease Working Groups now suggest a mean hippocampal
volume deficit, relative to controls, of around 3.6 % in
schizophrenia and 2.9 % in bipolar disorder—depending, of
course, on cohort-specific factors (medication, duration of
illness, etc.). By contrast, in ENIGMA1, the rs7294919
genetic variant was associated with a hippocampal volume
decreased by 47.6 mm3or by 1.2 % of the average
hippocampal volume per risk allele. Bearing in mind that the
cause of the effects is quite different, the effect of ENIGMA’s
top hippocampal SNPs on hippocampal volume is
approximately 5 % of the mean effect of Alzheimer’s disease,
around one-third of the effect of schizophrenia, and around
40 % of the effect of bipolar disorder. If such effect sizes are
typical, sample size requirements will generally be larger in
genetic association studies than in neuroimaging studies of
disease effects on the same phenotypes, but not vastly larger.
To some researchers, these preliminary observations
suggest an expected pattern of effect sizes, whereby GWAS
for cognitive traits may have top hits with smaller effect sizes
than those for imaging traits, which in turn may tend to have
smaller effect sizes than expression traits. Also, now that
consistent hits have been identified in ENIGMA1 and
ENIGMA2, it should be possible to estimate effect sizes for
the same hits in new replication samples, to understand what
sample sizes are sufficient to detect them. Conversely, some
have argued that the available data on genetic effect sizes at
different levels of neuroscience are currently too limited to
draw conclusions from only the three data points cited here.
Clearly, we do not yet have sufficient information to say
whether imaging genetics effects sizes will always be larger
than effect sizes for cognitive/behavioral phenotypes. One
way to begin to address this question might be to quantify
effects of some specific functional SNP at three different
related levels, such as eQTL data from the hippocampus,
hippocampal volume on MRI, and a behavior that is tightly
dependent on hippocampal function; this may show the
appropriate differential effect sizes in one framework.
A further interesting angle is the follow-up of the top
ENIGMA1 hits in ethnically isolated cohorts of Dagestan
and Chechnya by Bulayeva et al. (2013). Genetic isolates
can be valuable for studying any human phenotypes; required
heterogeneous outbred populations due to the genetic
homogeneity of these isolates. This has long been noted by
classical statistical and population geneticists (e.g., Falconer
1960; Neel 1992; Bulayeva et al. 2005). As genetic isolates
have a high rate of traditional endogamy and inbreeding, it is
not possible to perform GWAS, but genome-wide linkage
analysis is possible (Sheffield et al. 1998). In particular, the
analysis of isolated populations makes it less challenging to
study polygenic disorders by reducing the number of loci
possibly involved in the disorder.
ENIGMA now focuses on genetic analysis of
neuroimaging measures, but psychiatric diagnosis is
important as well. ENIGMA does not simply relate imaging
data to genetics; many of its working groups study the
relationship between imaging measures and diagnosis. So, in
a sense, psychiatric diagnosis is also a key target of study. For
example, recent GWAS and follow-up studies have provided
common risk factor for bipolar disorder and schizophrenia
(Cichon et al. 2011; Mühleisen et al. 2012). Both studies
found that the A allele of SNP rs1064395 is a risk-mediating
allele and that rs1064395 influences risk to a broad psychosis
phenotype. To identify a putative mechanism, Schultz et al.
(2013) tested whether the risk allele has an influence on brain
structure. In patients with schizophrenia, they found a
significant association with higher folding in a right lateral
occipital region and at a trend level for the left dorsolateral
prefrontal cortex. Controls did not show an association. The
findings suggest a role of NCAN in visual processing and top-
down cognitive functioning. Both major cognitive processes
are known to be disturbed in schizophrenia. Another GWAS
study by Rietschel et al. (2012) identified a risk factor for
schizophrenia in a chromosomal region harboring the genes
AMBRA1/CHRM4/DGKZ/MDK (rs11819869). In an
independent follow-up analysis, they found that healthy
carriers of the risk allele T showed altered activation in the
subgenual cingulate cortex during a cognitive control task.
This brain region is a critical interface between emotion
regulation and cognition, which are structurally and
functionally abnormal in schizophrenia and bipolar disorder.
The recent successes of psychiatric GWAS have unearthed
a vast resource of findings when sample sizes became
very large (Ripke et al. 2011, for the Schizophrenia
Psychiatric Genome-Wide Association Study (GWAS)
Consortium; Sklar et al. 2011, for the Psychiatric
GWAS Consortium Bipolar Disorder Working Group;
Cichon et al. 2011, for the MooDS consortium Bipolar
Disorder Group; Rietschel et al. 2012, for the MooDS
consortium Schizophrenia Group; Cross-Disorder Group
of the Psychiatric Genomics Consortium et al. 2013). In
fact, it was the need for large samples that encouraged
the ENIGMA groups to work together, realizing that
scans would be severely limited.
A secondquestioniswhichofthe recentlyreportedGWAS
findings are true or credible if they were discovered in small
cohorts. ENIGMA focuses on meta-analyses, but some have
argued that smaller individual studies—when replicated—
may discover interesting (and strong) associations. Paus et
al. (2012) noted that they had run a successful GWAS in only
around 300 participants. Also, some GWAS signals may be
important in the context of some individual cohorts, but may
be washed out in meta-analyses. A more skeptical line of
Brain Imaging and Behavior
argument is that if we need 20,000+ samples to detect an
effect, most likely what we see has very small effect size.
Even so, some recent GWAS analyses appear to have picked
up variants with strong effects—e.g., that roughly double
disease risk (e.g., in TREM2), and these effects have been
confirmed insmaller cohorts bycomparing brain imagesfrom
carriers and non-carriers (Rajagopalan et al. 2013).
One testable hypothesis is that GWAS would be more
powerful and efficient if we select imaging phenotypes in a
principled way; if this is true, it may be possible to perform
GWAS with consistently replicated hits, that do not require
tens of thousands of subjects to reject the null hypothesis. In
fact, the ENIGMAworking groups each aim to find heritable
brain measuresthatmaximize case–control differentiation, for
studies of disease. This will help us to prioritize the future
targets of study for influential genes. Although ENIGMA is
data-driven, that does not mean that we cannot use patterns in
the findings to design more targeted approaches that prioritize
phenotypes and genetic loci for follow-up analyses.
One such analysis (Desrivieres et al. 2013) evaluates pre-
selected genes that are expressed in the brain and change in
their expression throughout brain development. By narrowing
the search space to genes that are likely to play a role—and
whose functions have more chance of being understood—the
neuroscience and medicine. This must be balanced with the
knowledge that approximately 88 % of GWAS hits are in
intergenic regions (Hindorff et al. 2009) and almost all genes
are expressed at some location in the brain at some period of
the lifespan (www.brainspan.org).
ENIGMA, to date, has used a mass-univariate analysis,
where each trait (or brain measure) is considered on its own,
and each genetic variant is considered on its own. Recent
multivariate analyses can cluster voxels in the brain—or
SNPs on the genome—to empower analyses, sometimes with
both forms of clustering occurring at once (Hibar et al. 2011a;
Thompson et al. 2013). Some of these multivariate analysis
methods have been used to detect significant hits in image-
wide genome-wide searches in cohorts of under 1,000
subjects (Ge et al. 2012; Chen et al. 2012b; Jahanshad et al.
2013b). In most analyses, multivariate refers to condensing
both methods and joint methods are emerging. Multivariate
methods can be quite sophisticated mathematically. Some
genetics and twin designs. Chiang and colleagues (Chiang et
al. 2011, 2012), for example, computed the “cross-trait cross-
twin correlation” between all pairs of voxels in an image, to
pull out “image clusters” with common genetic determination
(see also Chen et al. 2011, 2012a). Others have incorporated
principled way to mine high-dimensional datasets, boosting
power for any subsequent GWAS.
A further line of work studies the “interactome”: it is now
possible to search pairs or sets of SNPs for interaction effects
in images (Hibar et al. 2013c) and some have argued that this
is the norm for mechanisms of gene action, and the context of
other genetic variants should be included in the analyses
(Hariri and Weinberger 2003). For instance, Roffman et al.
(2008) were the first to show functional MRI evidence of
epistasisinschizophrenia.Their findingswere consistentwith
epistatic effects of the COMT and MTHFR polymorphisms
on prefrontal dopamine signaling, suggesting that in
schizophrenia patients, the MTHFR 677 T allele exacerbates
prefrontal dopamine deficiency. Andreasen et al. (2012) used
interacting with one another and predicting a continuous
outcome measure that is a biologically meaningful phenotype
(“intermediate phenotype”) for schizophrenia: changes in
brain structure occurring after the onset of the illness.
Expanding the study of interactions to the whole genome,
Kam-Thong et al. (2012) presented a whole genome analysis
of epistasis (SNP-by-SNP) on several traits, including
hippocampal volume. Pandey et al. (2012) reported a pathway
analysis, which includes information from the network of
genes for pathway analysis. Clearly no genetic variant acts
alone; however, multiple comparisons increase exponentially
In conclusion, when the next round of ENIGMA studies
has been completed, there will be a scaffolding of credible
genetic variants on which to build and test new models of
macroscale brain development, the implications of those
variantsfor neuropsychiatricdisease,anda new basis toprobe
the genetic architecture of the living human brain.
Conflict of interest
The authors declare that they have no conflict of
Open AccessThis article is distributed under the terms of the Creative
Commons Attribution License which permits any use, distribution, and
reproduction in any medium, provided the original author(s) and the
source are credited.
Almasy, L., Gur, R. C., Haack, K., Cole, S. A., Calkins, M. E., Peralta, J.
M., et al. (2008). A genome screen for quantitative trait loci
influencing schizophrenia and neurocognitive phenotypes. The
American Journal of Psychiatry, 165, 1185–1192.
Alzheimer’s Association. (2013). The Worldwide ADNI. URL: http://
Accessed 2 June 2013.
Brain Imaging and Behavior
Amunts, K., Schleicher, A., Burgel, U., Mohlberg, H., Uylings, H. B., &
Zilles, K. (1999). Broca’s region revisited: cytoarchitecture and
intersubject variability. The Journal of Comparative Neurology,
Andreasen, N. C., Wilcox, M. A., Ho, B. C., Epping, E., Ziebell, S.,
Zeien, E., et al. (2012). Statistical epistasis and progressive brain
change in schizophrenia: an approach for examining the
relationships between multiple genes. Molecular Psychiatry,
Ashburner, J., Andersson, J. L., & Friston, K. J. (1999). High-
dimensional image registration using symmetric priors.
NeuroImage, 9, 619–628.
Baaré, W. F., Hulshoff Pol, H. E., Boomsma, D. I., Posthuma, D., de
Geus, E. J., Schnack, H. G., et al. (2001). Quantitative genetic
modeling of variation in human brain morphology. Cerebral
Cortex, 11, 816–824.
Bakken, T. E., Roddey, J. C., Djurovic, S., Akshoomoff, N., Amaral, D.
G., Bloss, C. S., et al. (2012). Association of common genetic
variants in GPCPD1 with scaling of visual cortical surface area in
humans. Proceedings of the National Academy of Sciences of the
United States of America, 109, 3985–3990.
et al. (2009). Expression and evolutionary conservation of the
tescalcin gene during development. Gene expression patterns :
GEP, 9, 273–281.
Batouli, S. A. H., Sachdev, P. S., Wen, W., Wright, M. J., Suo, C., Ames,
D., et al. (2012). The heritability of brain metabolites on proton
magnetic resonance spectroscopy in older individuals. Neuroimage,
Barnes, J., Bartlett, J. W., van de Pol, L. A., Loy, C. T., Scahill, R. I.,
in Alzheimer's disease. Neurobiol Aging, 30(11), 1711–23.
Baumgartner, M., Patel, H., & Barber, D. L. (2004). Na(+)/H(+)
exchanger NHE1 as plasma membrane scaffold in the assembly of
signaling complexes. American Journal of Physiology - Cell
Physiology, 287, C844–C850.
perspectives. International Review of Neurobiology, 84, 167–184.
Biffi, A., Anderson, C. D., Desikan, R. S., Sabuncu, M., Cortellini, L.,
Schmansky, N., et al. (2010). Genetic variation and neuroimaging
measures in Alzheimer disease. Archives of Neurology, 67, 677–
M., et al. (2012). Common variants at 12q14 and 12q24 are
associated with hippocampal volume. Nature Genetics, 44, 545–
Bishop, D. (2013). Blog posting: http://deevybee.blogspot.co.uk/2013/
01/genetic-variation-and-neuroimaging-some.html. Accessed 5
trait loci: King Harvest has surely come. Current Opinion in
Genetics & Development, 14, 233–240.
Blokland, G. A., de Zubicaray, G. I., McMahon, K. L., & Wright, M. J.
(2012). Genetic and environmental influences on neuroimaging
phenotypes: a meta-analytical perspective on twin imaging studies.
Twin Research and Human Genetics: the Official Journal of the
International Society for Twin Studies, 15, 351–371.
Bralten, J., Arias-Vásquez, A., Makkinje, R., Veltman, J. A., Brunner, H.
SORL1 with hippocampal volume in young, healthy adults. The
American Journal of Psychiatry, 168(10), 1083–1089. doi:10.1176/
Braskie, M. N., Jahanshad, N., Stein, J. L., Barysheva, M.,
McMahon, K. L., de Zubicaray, G. I., et al. (2011). Common
Alzheimer’s disease risk variant within the CLU gene affects
Braskie, M. N., Jahanshad, N., Stein, J. L., Barysheva, M., Johnson, K.,
McMahon, K. L., et al. (2012). Relationship of a variant in the
NTRK1 gene to white matter microstructure in young adults.
Journal of Neuroscience, 32, 5964–5972.
Brouwer, R. M., Mandl, R. C., Peper, J. S., van Baal, G. C., Kahn, R. S.,
Boomsma, D. I., et al. (2010). Heritability of DTI and MTR in nine-
year-old children. Neuroimage, 53(3), 1085–1092.
Genetics Part B: Neuropsychiatric Genetics, 132B(1), 76–84.
Bulayeva, K., Glatt, S., Walsh, C., Gurgenova, F., Berdichevets, I.,
variants at 12q24.21-q24.32 found in genetic isolate with
aggregation of unspecific mental retardation. Open Journal of
Genomics, 2, 3.
Button, K. S., Ioannidis, J. P., Mokrysz, C., Nosek, B. A., Flint, J.,
Robinson, E. S., et al. (2013). Power failure: why small sample size
undermines the reliability of neuroscience. Nature Review
Neuroscience, 14(5), 365–376. doi:10.1038/nrn3475.
Calhoun, V. D., Liu, J., & Adali, T. (2009). A review of group ICA for
fMRI dataand ICA for joint inference ofimaging, genetic, and ERP
data. Neuroimage, 45(1 Suppl), S163–S172. doi:10.1016/j.
CARDIoGRAMplusCD4_Consortium, Deloukas, P., Kanoni, S.,
Willenborg, C., Farrall, M., Assimes, T. L., et al. (2013). Large-
scale association analysis identifies new risk loci for coronary artery
disease. Nature Genetics, 45, 25–33.
Carrillo, M. C., Bain, L. J., Frisoni, G. B., & Weiner, M. W. (2012).
Worldwide Alzheimer’s Disease Neuroimaging Initiative.
Alzheimer’s & Dementia : the Journal of the Alzheimer’s
Association, 8, 337–342.
Fennema-Notestine, C., et al. (2011). Genetic influences on cortical
regionalization in the human brain. Neuron, 72, 537–544.
Chen, C. H., Gutierrez, E. D., Thompson, W., Panizzon, M. S., Jernigan,
human cortical surface area. Science, 335, 1634–1636.
Chen, J., Calhoun, V. D., Pearlson, G. D., Ehrlich, S., Turner, J., Ho, B.
C., et al. (2012b). Multifaceted genomic risk for brain function in
schizophrenia. NeuroImage, 61, 866–875.
Chen, J., Calhoun, V.D., Pearlson, G.D., Perrone-Bizzozero, N., Sui, J.,
Turner, J.A., et al. (2013). Guided exploration of genomic risk for
gray matter abnormalities in schizophrenia using parallel
independent component analysis with reference. NeuroImage (in
Chiang, M. C., McMahon, K. L., de Zubicaray, G. I., Martin, N. G.,
Hickie, I., Toga, A. W., et al. (2011). Genetics of white matter
development: a DTI study of 705 twins and their siblings aged 12
to 29. NeuroImage, 54, 2308–2317.
Chiang, M. C., Barysheva, M., McMahon, K. L., de Zubicaray, G. I.,
Johnson, K., Montgomery, G. W., et al. (2012). Gene network
effects on brain microstructure and intellectual performance
identified in 472 twins. Journal of Neuroscience, 32, 8732–8745.
Cichon, S., Mühleisen, T. W., Degenhardt, F. A., Mattheisen, M., Miró,
X., Strohmaier, J., et al. (2011). Genome-wide association study
identifies genetic variation in neurocan as a susceptibility factor for
Cohen, R. A., Harezlak, J., Schifitto, G., Hana, G., Clark, U., Gongvatana
A., et al. (2010). Effects of nadir CD4 count and duration of human
antiretroviral therapy era. Journal of neurovirology, 16(1), 25–32.
Brain Imaging and Behavior
3D intersubject registration of MR volumetric data in standardized
Talairach space. Journal of Computer Assisted Tomography, 18,
Corder, E. H., Saunders, A. M., Strittmatter, W. J., Schmechel, D. E.,
Gaskell, P. C., Small, G. W., et al. (1993). Gene dose of
apolipoprotein E type 4 allele and the risk of Alzheimer’s disease
in late onset families. Science, 261, 921–923.
Smoller, J. W., Craddock, N., Kendler, K., Lee, P. H., Neale, B. M.,
et al. (2013). Identification of risk loci with shared effects on five
major psychiatric disorders: a genome-wide analysis. Lancet, 381,
Davies, G., Tenesa, A., Payton, A., Yang, J., Harris, S. E., Liewald, D., et
al. (2011). Genome-wide association studies establish that human
16(10), 996–1005. doi:10.1038/mp.2011.85.
Davies, G., Harris, S.E., Reynolds, C.A., Payton, A., Knight, H.M.,
Liewald, D.C., et al. (2012). A genome-wide association study
implicates the APOE locus in nonpathological cognitive ageing.
Molecular Psychiatry (in Press).
De Geus, E. J. C. (2010). From genotype to EEG endophenotype: a route
for post-genomic understanding of complex psychiatric disease?
Genome Medicine, 2, 63. http://genomemedicine.com/content/2/9/
Den Braber, A., Bohlken, M.M., Brouwer, R.M., van ’t Ent, D., Kanai,
R., Kahn, R.S., et al. (2013). Heritability of subcortical brain
measures: a perspective for future genome-wide association studies.
NeuroImage (in press).
Desrivieres, S., et al. (2013). Single nucleotide polymorphism in the
neuroplastin locus associates with cortical thickness and verbal
Ecker, C., Ginestet, C., Feng, Y., Johnston, P., Lombardo, M. V., Lai, M.
C., et al. (2013). Brain surface anatomy in adults with autism: the
relationship between surface area, cortical thickness, and autistic
symptoms. JAMA Psychiatry, 70, 59–70.
Elias-Sonnenschein, L. S., Viechtbauer, W., Ramakers, I. H., Verhey, F.
R., & Visser, P. J. (2011). Predictive value of APOE-ε4 allele for
progression from MCI to AD-type dementia: a meta-analysis.
Journal of Neurology, Neurosurgery and Psychiatry, 82(10),
Erk, S., Meyer-Lindenberg, A., Opitz von Boberfeld, C., Esslinger, C.,
Schnell, K., Kirsch, P., et al. (2011). Hippocampal function in
healthy carriers of the CLU Alzheimer’s disease risk variant.
Journal of Neuroscience, 49, 18180–18184.
Erk, S., Meyer-Lindenberg, A., Schmierer, P., Grimm, O., Tost, H.,
Mühleisen, T., et al. (2013). Functional impact of a recently
identified quantitative trait locus for hippocampal volume with
genome-wide support. Translational Psychiatry (in press).
Esslinger, C., Walter, H., Kirsch, P.,Erk, S., Schnell, K., Arnold, C., etal.
(2009) Neural mechanisms of a genome-wide supported psychosis
variant. Science, 324(5927), 605.
Estrada, K., Styrkarsdottir, U., Evangelou, E., Hsu, Y. H., Duncan, E. L.,
Ntzani, E. E., et al. (2012). Genome-wide meta-analysis identifies
56bone mineral density loci and reveals 14 loci associated with risk
of fracture. Nature Genetics, 44, 491–501.
Evans, A. C. (2006). The NIH MRI study of normal brain development.
NeuroImage, 30, 184–202.
Falconer, D. S. (1960). Introduction to quantitative genetics. Edinburgh:
Oliver and Boyd.
Fischl, B., van der Kouwe, A., Destrieux, C., Halgren, E., Segonne, F.,
Salat, D. H., et al. (2004). Automatically parcellating the human
cerebral cortex. Cerebral Cortex, 14, 11–22.
Flint, J., & Munafo, M. R. (2013). Candidate and non-candidate
genes in behavior genetics. Current Opinion in Neurobiology,
Flint, J., Greenspan, R. J., & Kendler, K. S. (2010). How genes influence
behavior. Oxford, NY: Oxford University Press.
Fornage, M., Debette, S., Bis, J. C., Schmidt, H., Ikram, M. A., Dufouil,
C., et al. (2011). Genome-wide association studies of cerebral white
matter lesion burden: the CHARGE consortium. Annals of
Neurology, 69, 928–939.
Fornito, A., Zalesky, A., Bassett, D. S., Meunier, D., Ellison-Wright, I.,
Yücel, M., et al. (2011). Genetic influences on cost-efficient
organization of human cortical functional networks. Journal of
Neuroscience, 31, 3261–3270. doi:10.1523/jneurosci.4858-10.
Frackowiak, R. S. J. (1997). Human brain function. San Diego:
Friston, K. J., Holmes, A. P., Poline, J. B., Grasby, P. J., Williams, S. C.,
Frackowiak, R. S., et al. (1995). Analysis of fMRI time-series
revisited. NeuroImage, 2, 45–53.
Frodl, T., Meisenzahl, E. M., Zill, P., Baghai, T., Rujescu, D., Leinsinger,
G., et al. (2004).Reduced hippocampal volumes associated with the
long variant of the serotonin transporter polymorphism in major
depression. Archives of General Psychiatry, 61(2), 177–183.
Frodl, T., Koutsouleris, N., Bottlender, R., Jäger, M., Born, C.,
Mörgenthaler, M., et al. (2008). Reduced gray matter brain volumes
are associated with variants of the serotonin transporter gene in
major depression. Molecular Psychiatry, 13(12), 1093–1101.
Ge, T., Feng, J., Hibar, D. P., Thompson, P. M., & Nichols, T. E. (2012).
Increasing power for voxel-wise genome-wide association studies:
the random field theory, least square kernel machines and fast
permutation procedures. NeuroImage, 63, 858–873.
Glahn, D. C., Thompson, P. M., & Blangero, J. (2007). Neuroimaging
endophenotypes: strategies for finding genes influencing brain
structure and function. Human Brain Mapping, 28, 488–501.
Glahn, D. C., Winkler, A. M., Kochunov, P., Almasy, L., Duggirala, R.,
Carless, M. A., et al. (2010). Genetic control over the resting brain.
Proceedings of the National Academy of Sciences of the United
States of America, 107, 1223–1228.
Glahn, D. C., Curran, J. E., Winkler, A. M., Carless, M. A., Kent, J. W.,
Jr., Charlesworth, J. C., et al. (2012). High dimensional
endophenotype ranking in the search for major depression risk
genes. Biological Psychiatry, 71, 6–14.
Glover, G. H., Mueller, B. A., Turner, J. A., van Erp, T. G., Liu, T. T.,
Greve, D. N., et al. (2012). Function biomedical informatics
research network recommendations for prospective multicenter
functional MRI studies. Journal of Magnetic Resonance Imaging,
Goldman, D. (2012). Our genes, our choices : how genotype and gene
interactions affect behavior. London: Elsevier/Academic Press.
Gollub,R., Shoemaker, J.M., King,M.,White,T., Ehrlich, S.,Sponheim,
S., et al. (2013) The MCIC collection: a shared repository of multi-
modal, multi-site brain image data from a clinical investigation of
schizophrenia. Journal of NeuroInformatics. PMID: 23760817.
Gottesman, I. I., & Gould, T. D. (2003). The endophenotype concept in
psychiatry: etymology and strategic intentions. The American
Journal of Psychiatry, 160, 636–645.
Hall, J., Whalley, H. C., Job, D. E., Baig, B. J., McIntosh, A. M., Evans,
K. L., et al. (2006). A neuregulin 1 variantassociatedwith abnormal
cortical function and psychotic symptoms. Nature Neuroscience,
Hariri, A. R., & Weinberger, D. R. (2003). Imaging genomics. British
Medical Bulletin, 65, 259–270.
Harold, D., Abraham, R., Hollingworth, P., Sims, R., Gerrish, A.,
Hamshere, M. L., et al. (2009). Genome-wide association study
identifies variants at CLU and PICALM associated with
Alzheimer’s disease. Nature Genetics, 41, 1088–1093.
Hibar, D. P., Kohannim, O., Stein, J. L., Chiang, M. C., & Thompson, P.
M. (2011a). Multilocus genetic analysis of brain images. Frontiers
in Genetics, 2, 73.
Brain Imaging and Behavior
Hibar, D. P., Stein, J. L., Kohannim, O., Jahanshad, N., Saykin, A. J.,
Shen, L., et al. (2011b). Voxelwise gene-wide association study
(vGeneWAS): multivariate gene-based association testing in 731
elderly subjects. NeuroImage, 56, 1875–1891.
Hibar, D. P., van Erp, T. G. M., Turner, J. A., Haukvik, U. K., Agartz, I.,
et al. (2013a). Meta-analysis of structural brain differences in
bipolar disorder: the ENIGMA-Bipolar Disorder Project. Seattle,
Hibar, D. P., Stein, J. L., Ryles, A. B., Kohannim, O., Jahanshad, N.,
Medland, S. E., et al. (2013b). Genome-wide association identifies
geneticvariants associated with lentiform nucleusvolumeinN=1345
Hibar, D., Stein, J.L., Jahanshad, N., Toga, A.W., McMahon, K.L., de
Zubicaray, G.I., et al. (2013c). Exhaustive search of the SNP-SNP
interactome identifies replicated epistatic effects on brain volume.
MICCAI 2013, Nagoya, Japan, Sept. 22–26 2013 (in press).
Hindorff, L. A., Sethupathy, P., Junkins, H. A., Ramos, E. M., Mehta, J.
P., Collins, F. S., et al. (2009). Potential etiologic and functional
implications of genome-wide association loci for human diseases
and traits. Proceedings of the National Academy of Sciences of the
United States of America, 106, 9362–9367.
Hollingworth, P., Harold, D., Sims, R., Gerrish, A., Lambert, J. C.,
Carrasquillo, M. M., et al. (2011). Common variants at ABCA7,
MS4A6A/MS4A4E, EPHA1, CD33 and CD2AP are associated
with Alzheimer’s disease. Nature Genetics, 43(5), 429–435. doi:
Holmes, A. J., Lee, P. H., Hollinshead, M. O., Bakst, L., Roffman, J. L.,
Smoller, J. W., et al. (2012). Individual differences in amygdala-
medial prefrontal anatomy link negative affect, impaired social
function, and polygenic depression risk. Journal of Neuroscience,
Amouyel, P., et al. (2012). Genome-wide and gene-based
association implicates FRMD6 in Alzheimer disease. Human
Mutation, 33, 521–529.
Hulshoff Pol, H. E., Schnack, H. G., Mandl, R. C., Brans, R. G., van
Haren, N. E., Baare, W. F., et al. (2006). Gray and white matter
density changes in monozygotic and same-sex dizygotic twins
discordant for schizophrenia using voxel-based morphometry.
Neuroimage, 31(2), 482–488.
Ikram, M. A., Fornage, M., Smith, A. V., Seshadri, S., Schmidt, R.,
Debette, S., et al. (2012). Common variants at 6q22 and 17q21 are
associated with intracranial volume. Nature Genetics, 44, 539–544.
Ioannidis, J. P. (2005). Why most published research findings are false.
PLoS Medicine, 2, e124.
Jack, C. R., Jr. (2012). Alzheimer disease: new concepts on its
neurobiology and the clinical role imaging will play. Radiology,
Jahanshad, N., Zhan, L., Bernstein, M.A., Borowski, B.J., Jack, C.R.,
Toga, A.W., et al. (2010). Diffusion tensor imaging in seven
minutes: determining trade-offs between spatial and directional
resolution. In: Biomedical Imaging: From Nano to Macro (pp.
1161–1164). 2010 IEEE International Symposium on Biomedical
Jahanshad, N., Kohannim, O., Hibar, D. P., Stein, J. L., McMahon, K. L.,
de Zubicaray, G. I., et al. (2012). Brain structure in healthy adults is
Jahanshad, N., Kochunov, P., Sprooten, E., Mandl, R.C., Nichols, T.E.,
Almasy, L., et al. (2013a). Multi-site genetic analysis of diffusion
images and voxelwise heritability analysis: A pilot project of the
ENIGMA-DTI working group. Neuroimage, 81, 455–469.
discovers SPON1 gene variant influencing dementia severity.
Proceedings of the National Academy of Sciences of the United
States of America, 110, 4768–4773.
Jahanshad, N., Nir, T. M., Toga, A. W., Jack, C. R., Bernstein, M. A.,
Weiner, M. W., & Thompson, P. M. (2013c). Seemingly Unrelated
Regression empowers detection of network failure in dementia.
Neurobiol Aging. In Press (SI: Neuroimaging Biomarkers in
Jenkinson, M., Bannister, P., Brady, M., & Smith, S. (2002).
Improvedoptimizationfor therobust andaccuratelinear registration
and motion correction of brain images. NeuroImage, 17(2),
Jenkinson, M., Beckmann, C. F., Behrens, T. E., Woolrich, M. W., &
Smith, S. M. (2012). FSL. NeuroImage, 62, 782–790.
Jones,E.G.,& Mendell,L.M. (1999). Assessingthe decadeofthe brain.
Science, 284, 739.
Agartz, I., et al. (2009). A common MECP2 haplotype associates
with reduced cortical surface area in humans in two independent
United States of America, 106, 15483–15488.
Kam-Thong, T., et al. (2012). GLIDE: GPU-based linear regression for
detection of epistasis. Human Heredity, 73(4), 220–236.
Kendler, K. S., & Neale, M. C. (2010). Endophenotype: a conceptual
analysis. Molecular Psychiatry, 15, 789–797.
Klei, L., Sanders, S., Murtha, M., Hus, V., Lowe, J., Willsey, A., et al.
(2012). Common genetic variants, acting additively, are a major
source of risk for autism. Molecular Autism, 3 (1). doi:10.1186/
Kochunov, P., Glahn, D. C., Lancaster, J. L., Winkler, A. M., Smith, S.,
Thompson, P. M., et al. (2010). Genetics of microstructure of
cerebral white matter using diffusion tensor imaging. NeuroImage,
Kochunov, P., Jahanshad, N., Sprooten, E., Thompson, P., McIntosh, A.,
Deary, I., et al. (2012) Genome-wide association of full brain white
matter integrity—from the ENIGMA DTI working group. In:
Organization of Human Brain Mapping, Beijing, China.
Reese, A. H., et al. (2012). Predicting white matter integrity from
multiple common genetic variants. Neuropsychopharmacology :
official publication of the American College of
Neuropsychopharmacology, 37, 2012–2019.
Kohannim, O., Hua, X., Rajagopalan, P., Hibar, D.P., Jahanshad, N.,
Grill, J., et al. (2013). Multilocus genetic profiling to empower drug
trials and predict brain atrophy, Neuroimage-Clinical (in press).
Koten, J. W., Jr., Wood, G., Hagoort, P., Goebel, R., Propping, P.,
Willmes, K., et al. (2009). Genetic contribution to variation in
cognitive function: an FMRI study in twins. Science, 323, 1737–
Kremen, W. S., Prom-Wormley, E., Panizzon, M. S., Eyler, L. T., Fischl,
on the size of specific brain regions in midlife: the VETSA MRI
study. NeuroImage, 49(2), 1213–1223. doi:10.1016/j.neuroimage.
Lambert, J. C., Heath, S., Even, G., Campion, D., Sleegers, K., Hiltunen,
M., et al. (2009). Genome-wide association study identifies variants
at CLU and CR1 associated with Alzheimer’s disease. Nature
Genetics, 41(10), 1094–1099.
Lango Allen, H., Estrada, K., Lettre, G., Berndt, S. I., Weedon, M. N.,
Rivadeneira, F., et al. (2010). Hundreds of variants clustered in
genomic loci and biological pathways affect human height.
Nature, 467, 832–838.
Le Floch, Guillemot, V., Frouin, V., Pinel, P., Lalanne, C.,
Trinchera, L., et al. (2012). Significant correlation between a
set of genetic polymorphisms and a functional brain network
revealed by feature selection and sparse Partial Least Squares.
NeuroImage, 63(1), 11–24.
Brain Imaging and Behavior
Lee, S. H., Wray, N. R., Goddard, M. E., & Visscher, P. M. (2011).
Estimating missing heritability for disease from genome-wide
association studies. Am J Hum Genet, 88, 294–305.
Lee, S. H., Ripke, S., Neale, B. M., Faraone, S. V., Purcell, S. M., Perlis,
R. H. et al. (2013). Genetic relationship between five psychiatric
disorders estimated from genome-wide SNPs. Nature Genetics, 45,
al. (2009). Alzheimer’s disease neuroimaging initiative: a one-year
follow up study using tensor-based morphometry correlating
degenerative rates, biomarkers and cognition. NeuroImage, 45,
Leung, K. K., Barnes, J., Ridgway, G. R., Bartlett, J. W., Clarkson, M. J.,
Macdonald, K., et al. (2010). Automated cross-sectional and
longitudinal hippocampal volume measurement in mild cognitive
impairment and Alzheimer's disease. Neuroimage, 51(4), 1345–59.
Li, M., Luo, X.J., Rietschel, M., Lewis, C.M., Mattheisen, M., Muller-
Chinese for CREB1 SNPs and their implications in gene expression
regulation, hippocampal structure andfunction,andbipolar disorder
susceptibility. Molecular Psychiatry (in press).
Liu, J., Pearlson, G., Windemuth, A., Ruano, G., Perrone-Bizzozero, N.
I., & Calhoun, V. (2009). Combining fMRI and SNP data to
investigate connections between brain function and genetics using
parallel ICA. Human Brain Mapping, 30(1), 241–255.
Liu, C. C., Kanekiyo, T., Xu, H., & Bu, G. (2013). Apolipoprotein E and
Alzheimer disease: risk, mechanisms and therapy. Nature Reviews
Neurology, 9(2), 106–118. doi:10.1038/nrneurol.2012.263.
Logue, M. W., Schu, M., Vardarajan, B. N., Buros, J., Green, R. C., Go,
R. C., et al. (2011). A comprehensive genetic association study of
Lopez, L. M., Bastin, M. E., Munoz Maniega, S., Penke, L., Davies, G.,
Christoforou, A., et al. (2012). A genome-wide search for genetic
influences and biological pathways related to the brain’s white
matter integrity. Neurobiology of Aging, 33, 1847.e1–14.
al. (2011). Apolipoprotein E genotype is associated with temporal
and hippocampal atrophy rates in healthy elderly adults: a tensor-
Mazziotta, J. C., Toga, A. W., Evans, A., Fox, P., & Lancaster, J. (1995).
A probabilistic atlas of the human brain: theory and rationale for its
development. The International Consortium for Brain Mapping
(ICBM). NeuroImage, 2, 89–101.
McCarthy, M. I., & Hirschhorn, J. N. (2008). Genome-wide association
studies: potential next steps ona genetic journey. Human Molecular
Genetics, 17, R156–R165.
McCarthy, M. I., Abecasis, G. R., Cardon, L. R., Goldstein, D. B., Little,
complex traits: consensus, uncertainty and challenges. Nature
Reviews Genetics, 9, 356–369.
McIntosh, A. M., Moorhead, T. W., Job, D., Lymer, G. K., Muñoz
Maniega, S., McKirdy, J., et al. (2008). The effects of a neuregulin
1 variant on white matter density and integrity. Molecular
Psychiatry, 13(11), 1054–1059.
Meda, S. A., Jagannathan, K., Gelernter, J., Calhoun, V. D., Liu, J.,
Stevens, M. C., et al. (2010). A pilot multivariate parallel ICA study
to investigate differential linkage between neural networks and
genetic profiles in schizophrenia. (2010). NeuroImage, 53(3),
C., Calhoun, V. D., et al. (2012). A large scale multivariate parallel
ICA method reveals novel imaging-genetic relationships for
Alzheimer’s disease in the ADNI cohort. NeuroImage, 60(3),
Mühlau, W., Winkelmann, J., Rujescu, D., Giegling, I., Koutsouleris,
N., Gaser, C., et al. (2012). Variation within the Huntington’s
disease gene influences normal brain structure. PLoS One, 7(1),
Mühleisen, T. W., Mattheisen, M., Strohmaier, J., Degenhardt, F.,
Priebe, L., Schultz, C. C., et al. (2012). Association between
schizophrenia and common variation in neurocan (NCAN), a
genetic risk factor for bipolar disorder. Schizophrenia Research,
Naj, A. C., Jun, G., Beecham, G. W., Wang, L. S., Vardarajan, B. N.,
Buros, J., et al. (2011). Common variants at MS4A4/MS4A6E,
CD2AP, CD33 and EPHA1 are associated with late-onset
Alzheimer’s disease. Nature Genetics, 43, 436–441.
Neel, J. (1992). Minority populations as genetic isolates: the
interpretation of inbreeding results. In A. H. Bittles & D. F.
Roberts (Eds.), Minority populations: Genetics demography and
health. London: The MacMillan Press.
& Toga, A. W. (2012). EnigmaVis: online interactive visualization
Genetics through Meta-Analysis (ENIGMA) consortium. Twin
Research and Human Genetics, 15, 414–418.
Nugent, A. C., Luckenbaugh, D. A., Wood, S. E., Bogers, W., Zarate, C.
A., Jr., & Drevets, W. C. (2012). Automated subcortical
segmentation using FIRST: Test-retest reliability, interscanner
reliability, and comparison to manual segmentation. Human Brain
O'Donovan, M. C., Craddock, N., Norton, N., Williams, H., Pierce, T.,
Moskvina, V., Nikolov, I. et al. (2008). Identification of loci
associated with schizophrenia by genome-wide association and
follow-up. Nature Genetics, 40(9), 1053–1055.
Pandey, A., Davis, N. A., White, B. C., Pajewski, N. M., Savitz, J., &
Drevets, W. C. (2012). Epistasis network centrality analysis yields
pathway replication across two GWAS cohorts for bipolar disorder.
Translational Psychiatry, 2, e154. doi:10.1038/tp.2012.80.
Pasaniuc, B., Rohland, N., McLaren, P. J., Garimella, K., Zaitlen, N., Li,
H., et al. (2012). Extremely low-coverage sequencing and
imputation increases power for genome-wide association studies.
Nature Genetics, 44(6), 631–635.
Patenaude, B., Smith, S. M., Kennedy, D. N., & Jenkinson, M. (2011). A
Bayesian model of shape and appearance for subcortical brain
segmentation. NeuroImage, 56(3), 907–922.
Paus, T., Bernard, M., Chakravarty, M. M., Davey Smith, G., Gillis, J.,
Lourdusamy, A., et al. (2012). KCTD8 gene and brain growth in
adverse intrauterine environment: a genome-wide association study.
Cerebral Cortex, 22(11), 2634–2642. doi:10.1093/cercor/bhr350.
Pengas, G., Pereira, J. M., Williams, G. B., & Nestor, P. J. (2009).
Comparative reliability of total intracranial volume estimation
methods and the influence of atrophy in a longitudinal semantic
dementia cohort. Journal of Neuroimaging, 19(1), 37–46. doi:10.
Penke, L., Munoz Maniega, S., Murray, C., Gow, A. J., Valdes
Hernandez, M. C., Clayden, J. D., et al. (2010). A general factor
in healthy older people. Journal of Neuroscience, 30(7559), 7674.
Penke, L., Maniega, S. M., Bastin, M. E., Hernandez, M. C. V., Murray,
C., Royle, N. A., et al. (2012). Brain white matter integrity as a
neural foundation for general intelligence. Molecular Psychiatry,
Peper, J. S., Brouwer, R. M., Boomsma, D. I., Kahn, R. S., & Hulshoff
Pol, H. E. (2007). Genetic influences on human brain structure: a
review of brain imaging studies in twins. Human Brain Mapping,
28(6), 464–473. Review.
Pierson, R., Johnson, H., Harris, G., Keefe, H., Paulsen, J. S., Andreasen,
N. C., et al. (2011). Fully automated analysis using BRAINS:
AutoWorkup. NeuroImage, 54(1), 328–336.
Brain Imaging and Behavior
Hanke, M., et al. (2012). Data sharing in neuroimaging research.
Frontiers in Neuroinformatics, 6, 9.
Potkin, S. G., & Ford, J. M. (2009). Widespread cortical dysfunction in
schizophrenia: the FBIRN imaging consortium. Schizophrenia
Bulletin, 35, 15–18.
Potkin,S.G.,Guffanti,G., Lakatos,A., Turner, J.A., Kruggel,F.,Fallon,
J. H., et al. (2009a). Hippocampal atrophy as a quantitative trait in a
genome-wide association study identifying novel susceptibility
genes for Alzheimer’s disease. PLoS One, 4, e6501.
Potkin,S. G.,Turner, J.A.,Guffanti, G., Lakatos,A., Torri,F., Keator, D.
B., et al. (2009b). Genome-wide strategies for discovering genetic
influences on cognition and cognitive disorders: methodological
considerations. Cognitive Neuropsychiatry, 14, 391–418.
Sullivan, P. F., et al. (2009). Common polygenic variation contributes
to risk of schizophrenia and bipolar disorder. Nature, 460, 748–752.
risk gene carriers lose brain tissue faster, Letter to the Editor. New
England Journal of Medicine (in press).
Rietschel, M., Mattheisen, M., Degenhardt, F., Kahn, R. S., Linszen, D.
region on chromosome 11 and schizophrenia in large samples from
Europe. Molecular Psychiatry, 17(9), 906–917.
Rietveld, C.A., Medland, S.E., Derringer, J., Yang, J., Esko, T., Martin,
N.W., et al. (2013). GWAS of 126,559 individuals identifies genetic
variants associated with educational attainment. Science (in Press).
Rimol, L. M., Agartz, I., Djurovic, S., Brown, A. A., Roddey, J. C., Kahler,
A. K., et al. (2010). Sex-dependent association of common variants of
microcephaly genes with brain structure. Proceedings of the National
Academy of Sciences of the United States of America, 107, 384–388.
Ripke, S., Sanders, A. R., Kendler, K. S., Levinson, D. F., Sklar, P.,
Holmans, P. A., et al. (2011). Genome-wide association study
identifies five new schizophrenia loci. Nature Genetics, 43(10),
Roffman, J. L., Gollub, R. L., Calhoun, V. D., Wassink, T. H., Weiss, A.
P., Ho, B. C., et al. (2008). MTHFR 677C -> T genotype disrupts
prefrontal function in schizophrenia through an interaction with
COMT 158Val -> Met. Proceedings of the National Academy of
Sciences of the United States of America, 105(45), 17573–17578.
Rose, E. J., & Donohoe, G. (2013). Brain vs behavior: an effect size
comparison of neuroimaging and cognitive studies of genetic risk
for schizophrenia. Schizophrenia Bulletin, 39(3), 518–526. doi:10.
Rosenblatt, J., Benjamini, Y., Bogomolov, M., Stein, J. L., & Thompson,
P. M. (2013). vGWAS revisited: A novel and powerful approach to
voxelwise genomewide association studies. Seattle, WA: OHBM.
Ryles, A. B., Kohannim, O., Toga, A. W., Jack, C. R., Jr., Weiner, M.
W., McMahon, K. L., et al. (2012). Alzheimer’s disease risk
variant in the FRMD6 gene is associated with altered brain
structure: Opposite Effects in 740 elderly and 755 young
adults. New Orleans, LA: Society for Neuroscience Annual
Saykin, A. J., Shen, L., Foroud, T. M., Potkin, S. G., Swaminathan, S.,
Kim, S., et al. (2010). Alzheimer’s Disease Neuroimaging Initiative
and plans. Alzheimer’s & Dementia : the Journal of the Alzheimer’s
Association, 6, 265–273.
Schnack, H. G., van Haren, N. E., Brouwer, R. M., van Baal, G. C.,
Picchioni, M., Weisbrod, M., et al. (2010). Mapping reliability in
multicenter MRI: voxel-based morphometry and cortical thickness.
Schultz, C.C., Mühleisen, T.W., Nenadic, I., Koch, K., Wagner, G.,
Schachtzabel, C., et al. (2013). Common variation in NCAN, a risk
factor for bipolar disorder and schizophrenia, influences local
cortical folding in schizophrenia. Psychological Medicine.
Manuscript accepted for publication.
Schumann, G., Coin, L. J., Lourdusamy, A., Charoen, P., Berger, K. H.,
Stacey, D., et al. (2011). Genome-wide association and genetic
functional studies identify autism susceptibility candidate 2 gene
(AUTS2) in the regulation of alcohol consumption. Proceedings of
the National Academy of Sciences of the United States of America,
108(17), 7119–7124. doi:10.1073/pnas.1017288108.
Seshadri, S., Fitzpatrick, A. L., Ikram, M. A., DeStefano, A. L.,
Gudnason, V., Boada, M., et al. (2010). Genome-wide
analysis of genetic loci associated with Alzheimer disease.
JAMA, the Journal of the American Medical Association,
Sheffield, V. C., Stone, E. M., & Carmi, R. (1998). Use of isolated inbred
human populations for identification of disease genes. Trends in
Genetics, 14(10), 391–396.
Shenton,M.E.,Dickey, C.C.,Frumin,M.,& McCarley, R. W.(2001). A
review of MRI findings in schizophrenia. Schizophr Res, 49(1–2),
Sherva, R., Tripodis, Y., Bennett, D.A., Chibnik, L.B., Crane, P.K., de
Jager, P.L., et al (2013). Genome-wide association study of the rate
of cognitive decline in Alzheimer’s disease. Alzheimer’s &
Dementia : the Journal of the Alzheimer’s Association.
Shokouhi, M., Barnes, A., Suckling, J., Moorhead, T. W., Brennan, D.,
Job, D., et al. (2011). Assessment of the impact of the scanner-
related factors on brain morphometry analysis with Brainvisa. BMC
Medical Imaging, 11, 23. doi:10.1186/1471-2342-11-23.
Silver, M., Janousova, E., Hua, X., Thompson, P. M., Montana, G., &
Alzheimer’s Disease Neuroimaging Initiative. (2012). Identification
imaging phenotypes with sparse regression. NeuroImage, 63(3),
Sklar,P.,Ripke, S.,Scott,L.J.,Andreassen,O. A.,Cichon,S.,Craddock,
N., et al. (2011). Large-scale genome-wide association analysis of
bipolar disorder identifies a new susceptibility locus near ODZ4.
Nature Genetics, 43(10), 977–983. doi:10.1038/ng.943.
E., Mackay, C. E., et al. (2006). Tract-based spatial statistics:
voxelwise analysis of multi-subject diffusion data. NeuroImage,
Speliotes, E. K., Willer, C. J., Berndt, S. I., Monda, K. L., Thorleifsson,
G., Jackson, A. U., et al. (2010). Association analyses of 249,796
individuals reveal 18 new loci associated with body mass index.
Nature Genetics, 42, 937–948.
C., Hall, J., et al. (2013). White matter integrity as an intermediate
phenotype: exploratory genome-wide association analysis in
individuals at high risk of bipolar disorder. Psychiatry Research,
Stefansson, H., Ophoff, R. A., Steinberg, S., Andreassen, O. A., Cichon,
S., Rujescu, D., et al. (2009). Nature, 460, 744–747.
Stein, J. L., Hua, X., Lee, S., Ho, A. J., Leow, A. D., Toga, A. W., et al.
(2010). Voxelwise genome-wide association study (vGWAS).
NeuroImage, 53, 1160–1174.
Zubicaray, G. I., et al. (2011). Discovery and replication of
dopamine-related gene effects on caudate volume in young and
elderly populations (N=1198) using genome-wide search.
Molecular Psychiatry, 16(927–937), 881.
Stein, J. L., Medland, S. E., Vasquez, A. A., Hibar, D. P., Senstad, R. E.,
Winkler, A. M., et al. (2012). Identification of common variants
associated with human hippocampal and intracranial volumes.
Nature Genetics, 44, 552–561.
Stranger, B. E., Nica, A. C., Forrest, M. S., Dimas, A., Bird, C. P.,
Beazley, C., et al. (2007). Population genomics of human gene
expression. Nature Genetics, 39, 1217–1224.
Brain Imaging and Behavior
Talairach, J., Tournoux, P., & Missir, O. (1993). Referentially oriented Download full-text
cerebral MRI anatomy: an atlas of stereotaxic anatomical
correlationsforgray and white matter. Stuttgart: G. Thieme Verlag.
Thompson, P. M., & Jahanshad, N. (2012). Ironing out
neurodegeneration: is iron intake important during the teenage
years? Expert Review of Neurotherapeutics, 12, 629–631.
Huttunen, M., et al. (2001). Genetic influences on brain structure.
Nature Neuroscience, 4, 1253–1258.
Thompson, P.M., Ge, T., Glahn, D.C., Jahanshad, N., Nichols, T.E.
(2013). Genetics of the connectome. Neuroimage.
Thorgeirsson, T. E., Geller, F., Sulem, P., Rafnar, T., Wiste, A.,
Magnusson, K. P., et al. (2008). variant associated with nicotine
dependence, lung cancer and peripheral arterial disease. Nature,
452(7187), 638–642. doi:10.1038/nature06846.
Turner, J. A., Smyth, P., Macciardi, F., Fallon, J. H., Kennedy, J. L., &
Potkin, S. G. (2006). Imaging phenotypes and genotypes in
schizophrenia. Neuroinformatics, 4(1), 21–49.
Turner, J. A., Hibar, D. P., Rasmussen, J., Andreassen, O., Haukvik, U.
K., Agartz, I., et al. (2013). A prospective meta-analysis of
subcortical bain volumes in schizophrenia via the ENIGMA
consortium. Seattle, WA: OHBM.
D.I.,& Hulshoff Pol,H.E.(2013).Geneticcontroloffunctionalbrain
network efficiency in children. European Neuropsychopharmacology,
23, 19–23. doi:10.1016/j.euroneuro.2012.06.007.
van Erp, T. G., Saleh, P. A., Huttunen, M., Lonnqvist, J., Kaprio, J.,
Salonen, O., et al. (2004a). Hippocampal volumes in schizophrenic
twins. Archives of General Psychiatry, 61(4), 346–353.
van Erp, T., Cannon, T., Tran, H., Wobbekind, A., Huttunen, M.,
Lönnqvist, J., et al. (2004) Genetic influences on human brain
morphology. IEEE International Symposium on Biomedical
van Erp, T.G.M., Hibar, D.P., Rasmussen, J., Potkin, S., Ophoff, R.,
Andreassen, O., et al. (2013). A large-scale meta-analysis of
subcortical brain volume abnormalities in schizophrenia via the
ENIGMA consortium, Society for Biological Psychiatry (SOBP).
van Soelen,I. L., Brouwer, R.M.,Peper, J. S.,van Leeuwen, M.,Koenis,
M. M., van Beijsterveldt, T. C., et al. (2012). Brain SCALE: brain
structure and cognition: an adolescent longitudinal twin study into
the genetic etiology of individual differences. Twin Research and
Human Genetics, 15(3), 453–467. doi:10.1017/thg.2012.4.
Vounou, M., Nichols, T. E., & Montana, G. (2010). Discovering genetic
associations with high-dimensional neuroimaging phenotypes: a
sparse reduced-rank regression approach. NeuroImage, 53, 1147–
Vounou, M., Janousova, E., Wolz, R., Stein, J. L., Thompson, P. M.,
Rueckert, D., et al. (2012). Sparse reduced-rank regression detects
genetic associations with voxel-wise longitudinal phenotypes in
Alzheimer’s disease. NeuroImage, 60(1), 700–716.
Weiner, M. W., Veitch, D. P., Aisen, P. S., Beckett, L. A., Cairns, N. J.,
Green, R. C., et al. (2012). The Alzheimer’s disease neuroimaging
initiative: a review of papers published since its inception.
Alzheimer’s & Dementia : the Journal of the Alzheimer’s
Association, 8, S1–S68.
Whalley, H. C., Papmeyer, M., Sprooten, E., Romaniuk, L., Blackwood,
D. H., Glahn, D. C., et al. (2013a). The influence of polygenic risk
for bipolar disorder on neural activation assessed using fMRI.
Translational Psychiatry, 2, e130. doi:10.1038/tp.2012.60.
Whalley, H. C., Sprooten, E., Hackett, S., Hall, L., Blackwood, D. H.,
Glahn, D. C., et al. (2013b). Polygenic risk and white matter
integrity in individuals at high risk of mood disorder. Biological
White, T., & Gottesman, I. I. (2012). Brain connectivity and gyrification
as endophenotypes for schizophrenia: weight of the evidence.
Current Trends in Medicinal Chemistry, 12, 2393–2403.
White, T., Andreasen, N. C., & Nopoulos, P. (2002). Brain volumes and
surface morphology in monozygotic twins. Cerebral Cortex, 12,
White, T., El Marroun, H., Nijs, I., Schmidt, M. N., van der Lugt, A.,
Wielopolski, P., et al. (2013). Pediatric population-based imaging
and the Generation R Study: the intersection of developmental
neuroscience and epidemiology. European Journal of
Epidemiology, 28, 99–111.
T., et al. (2010). Cortical thickness or grey matter volume? The
importance of selecting the phenotype for imaging genetics studies.
NeuroImage, 53(3), 1135–1146. doi:10.1016/j.neuroimage.2009.
Woods, R. P., Mazziotta, J. C., & Cherry, S. R. (1993). MRI-PET
registration with automated algorithm. Journal of Computer
Assisted Tomography, 17, 536–546.
T. (2002). Genetic contributions to regional variability in human
Yang, J., Lee, S. H., Goddard, M. E., & Visscher, P. M. (2011). GCTA: a
tool for genome-wide complex trait analysis. American Journal of
Human Genetics, 88, 76–82.
Yang, J., Loos, R. J., Powell, J. E., Medland, S. E., Speliotes, E. K.,
Chasman, D. I., et al. (2012). FTO genotype is associated
with phenotypic variability of body mass index. Nature, 490,
Zetzsche, T., Preuss, U. W., Bondy, B., Frodl, T., Zill, P., Schmitt, G., et
al. (2008). 5-HT1Areceptor gene C-1019 G polymorphism and
amygdala volume in borderline personality disorder. Genes, Brain
and Behavior, 7(3), 306–313.
Zhan, L., Jahanshad, N., Ennis, D.B., Jin, Y., Bernstein, M.A., Borowski,
B.J., et al. (2012) Angular versus spatial resolution trade-offs for
diffusion imaging under time constraints. Human Brain Mapping.
Brain Imaging and Behavior