Neuroimaging Solutions A Hands-on Training Course for Data

							OHBM 2009 Course

Neuroimaging Solutions

Neuroimaging Solutions: A Hands-on Training Course for Data Processing, Analysis & Interpretation using the LONI Pipeline
Presenters: John Van Horn, Cornelius Hojatkashani, Rico Magsipoc, Ivo Dinov & Arthur W. Toga Affiliation: Laboratory of Neuro Imaging (LONI), University of California, Los Angeles, CA 90095 Logistics: URL (TBD) Date: TBD Time: TBD Place: TBD URL: http://Pipeline.loni.ucla.edu Format: Two-hour hands-on training workshop providing concrete Neuroimaging Solutions to multitudes of modern brain mapping problems. The maximum number of participants is limited to 45. All participants are expected to bring WiFi-enabled laptops using one of these operating systems: Linux, Windows, Apple/Mac/BSD. Lead-instructor will guide the focus of the group discussion, according to the agenda. Supporting instructors will navigate between attendees and help with specific questions and problems. Each attendee will be provided with the LONI Pipeline Handbook. The goals of the workshop are to: • Introduce attendees to the suite of neuroimaging solutions provided by the LONI Lab; • Demonstrate the interactive, graphical and computationally-efficient construction of valuable data analysis protocols using the LONI Pipeline environment; • Train attendees in designing novel graphical workflows for analyzing neuroimaging data, and • Provide concrete examples of constructing and utilizing integrated heterogeneous data, tools and services across multiple platforms, organizations and data-types. Summary: The LONI Pipeline provides a graphical framework for development, maintenance and dissemination of neuroimaging data-analysis protocols. The Pipeline environment offers a scalable infrastructure for graphical integration of diverse, complex and heterogeneous tools for neuroimaging and brain mapping research. This course will provide the necessary training to employ any of the available Pipeline library of modules and workflows. Attendees will also learn how to design new pipeline workflows by using existent modules or by introducing new module descriptors. All LONI tools, including the Pipeline, are freely distributed on the web (http://www.loni.ucla.edu/Software/). Specific neuroimaging workflows (e.g., tissue segmentation, cortical surface processing, etc.) will be demonstrated hands-on during the workshop. Agenda: Part 1: Introduction to the LONI Pipeline • What is the LONI Pipeline? • How does the Pipeline work? • How to construct basic module-descriptions and pipeline-workflows? • Pipeline optimization, environment variables, variable transformations, etc.

http://Pipeline.loni.ucla.edu

1

OHBM 2009 Course Part 2: Neuroimaging Solutions - automated and robust neuroimaging pipelines for • Data management, DB connections • Skull-stripping and tissue classification • Automated Brain-Volume parsing • Cortical surface extraction • Volume-based (VBM) and Tensor-based (TBM) morphometry

Neuroimaging Solutions

Pipeline Features: • Data, Executable and Workflow Provenance: enables tracking data, workflow and execution history of all processes. • Intelligence: facilitates the automated construction of elaborate, functional and valid workflows. • Scripting: Pipeline workflows may be exported as makefiles or a bash scripts for external execution. • Remote File Browser: enables browsing & selection of files at remote servers • Server Status: provides information about the status of the connected Pipeline server(s). • Job Execution Monitor: Provide status information of job execution • Libraries: Server (remote) and Personal (client) Libraries of modules and workflows • Data Interfaces: Provide diverse functions for data, database and remote server data retrieval. • Heterogeneous Tool Integration: Enables the integration of data, tools and services from multiple sites, platforms and formats. • Server Failover: Pipeline server provides a failover functionality to support smooth job completion despite accidental server loss. Figures:
Figure 1: Jack – could you please enter here a snapshot of your multi-modal pipeline? Also please enter caption and send the workflow for inclusion in the Pipeline groups-of-modules library. Figure 2: A wavelet analysis of image registration (WAIR) Pipeline, which demonstrates the analysis protocol for assessing quality of image registration of 2 groups of 9 subjects each, 4 different warping techniques, 3 different wavelet shrinkage methods and 3 wavelet-based warp classification functionals. This entire workflow is available for use, expansion and validation to the community via the LONI Pipeline library.

http://Pipeline.loni.ucla.edu

2