Transcript Child and Adolescent NeuroDevelopment Initiative Division of

“A Knowledge Environment for
Neuroimaging in Child Psychiatry”
R01 MH083320
Clinical Advisors Committee – 14.Feb.2012
Public Website:
www.nitrc.org/projects/candi_share
Agenda:
 Review of current status of grant objectives
 Brief overview of renewal objectives
 Discussion of 'community participation’
 Other discussion?
 Adjourn
CANDIShare - Aims
Aims:
• Aim 1: Use the existing resources NITRC to make a large set of MR image and
anatomic analysis data available. This will include:
• a. Image data - including structural and diffusion imaging at 1.5 and 3.0 T.
Each subject will include a comprehensive set of clinical, demographic and
behavioral measures.
• b. Segmentation – the segmentation results for ‘general segmentation’ and
parcellation (cortical and white matter).
• c. Creation and dissemination of static and dynamic probabilistic atlases
from specific subsets of these data for use in other segmentation and analysis
frameworks.
• Aim 2: Integrate the volumetric data derived from these images into the Internet
Brain Volume Database (IBVD), and use the IBVD system to catalogue the
representative literature of other published data from these similar age and
diagnostic categories of study.
• Aim 3: Create the prototype for a unified Child Psychiatry Neuroimaging Portal
that seamlessly integrates information across these resources and to select outside
resources (PubMed, BIRN, MRI Study of Normal Brain Development, etc.).
Promote the overall framework as a prototype to engender greater sharing of
information within the child psychiatric community
1R01 MH083320
http://www.nitrc.org/projects/candi_share
1R01 MH083320
CANDIShare - NITRC
CANDIShare - Release
Static Probability Atlases
Images and
Segmentation
Images only
1R01 MH083320
CANDIShare - Release
Individual
Data
PosNorm Space
1R01 MH083320
Linear
Non-linear
MNI152_T1_2mm Space
‘Static’
Probability
Atlases
1R01 MH083320
CANDIShare - Release
MNI152_T1_2mm Space
Linear
Probability per structure
…also released as a ‘.gca’ for use in Freesurfer
Non-linear
CANDIShare - Release
1R01 MH083320
CANDIShare - Release
1R01 MH083320
CANDIShare - Release
1R01 MH083320
Coming Soon (~2/15/12)
Get your CANDI here…
CANDIShare - Release
Also, on our servers and about to be released:
•Data fully organized for a second paper:
• Frazier, Jean A, et al. Cortical gray matter
differences identified by structural magnetic
resonance imaging in pediatric bipolar
disorder. ''Bipolar Disord''. 2005 Dec;
7(6):555-69
• Featuring ‘lobe’ and ‘cortical parcellation
unit’ segmentations
•Structure and Segmentation data for remaining
140 subjects, yet to be organized by paper.
1R01 MH083320
CANDIShare - Release
Importance of ‘Organization’ by Publication:
• Linkability!
• Reproducibility
1R01 MH083320
Dynamic Probability Server
Based upon:
•Search function of the publically available
CANDIShare NITRC-IR (XNAT)
•Subjects are pre-registered to target space
•Resultant subjects (age, gender, diagnosis, etc.)
are simply ‘averaged’ to generate structural
probability maps expressed in ‘raw’ form and as
‘.gca’ (for freesurfer)
Prototype:
Pediatric Neuroimaging Portal
Pediatric Neuroimaging Portal
Prototype:
•From one ‘launch pad’ can set numerous
‘contexts’:
• Age range, Diagnoses, gender, etc.
•Collect information about:
• Available imaging data (from CANDIShare
and others (i.e. NIH Pediatric Database, 1000
Functional Connectome, XNAT Central, etc.))
• Publications (PubMed)
• Volumes (IBVD)
• Genes (NCBI GeneDB)
Pediatric Neuroimaging Portal
Gene Example, Complex Query:
•Using PubBrain (http://www.pubbrain.org) to identify anatomic structural
prevalence in the literature, the ‘context’ of Pediatric Bipolar Disorder generates
the following ‘top 3’ anatomic contexts: prefrontal cortex, anterior cingulate, and
amygdala
•
•
•
From these ‘contexts’,
GeneDB generates:
•
Pediatric Bipolar
Disorder – 21
genes
•
Prefrontal cortex –
312 genes
•
Anterior cingulate
cortex – 76 genes
•
Amygdala - 254
genes
Find the common
genes!
Find Gene/Context
prevalence in the
literature!
BDNF, COMT, IL1B, SLC64A; DRD2, DRD4, HTR1A, NCAM1, SLC6A3
CANDIShare - future
 Competing renewal to be submitted (March 5, 2012)
 New Aims, in a nutshell:
 Generalize the sharing resource mechanism to others (we did it for
our data, we can help ‘you’ do it for yours…)
 More portal development: additional targets (‘activation’ BrainMap, SumsDB, ‘connectivity’ – CoCoMac, BAMS, etc.); interface
usability; data ‘collection’, etc.
 Discovery neuroscience, for pediatric neuropsychiatric populations,
using the integrated data and portal
 Validation – can we positively document data interoperability?
Where do CANDIShre ‘normals’ fall relative to PING and NIH
Ped DB ‘typicals’?
 Extension – do multiple types of analysis on the same data yield
consistent results?
 Novel findings and associations
 But, to do this, we NEED a bigger network of shared pediatric
psychiatric populations …
1R01 MH083320
CANDIShare - future
 But, to do this, we NEED a bigger network of shared pediatric
psychiatric populations…
 This MAY be where YOU come in…
 Do you have pediatric populations you are willing to make available?
 Would you be willing to be ‘beta-testers’ of our to-be-developed
YOUShare procedures (i.e. BJShare, BARRYShare, ELLENShare, JAYShare,
XShare, etc.) in order to begin to flesh out a coordinated and integrated
set of datasharing resources, that retain the lab and publication heritage
of the initial data collection and analysis efforts, but begins to magnify
the potential and impact of each of these datasets through their reuse
and greater integration?
 IF SO, what will it cost? Remember, retrospective datasharing is HARD
work…
1R01 MH083320
CANDIShare - Discussion
 Discussion…
 Thank you for your support and assistance on this!
1R01 MH083320