Researchers in neuroscience are increasingly exploiting the possibility of combining program modules from different software packages into processing pipelines to create analysis solutions which are not possible with a single software package alone. CamBA (Cambridge Brain Activation) is a suite of programs developed in a Virtual Pipeline Laboratory that facilitates this process and, further, provides a run-time environment for end-users. In this way, CamBA is a forum for the exchange of ideas and actions between methodologists and neuroscientists.
CamBA seperates code services into three domains: Pipeline processes, Data table management, Batch processing engine
Virtual Pipeline Laboratory: CamBAfx
CamBAfx is the environment where pipelines take their data, run and deliver their output. There is a familiar spreadsheet-styled user interface for data table management, which collects and validates of all the data to be processed as well as tools for loading, customising and designing pipelines. When completed and read, multiple subject datasets are processed by the pipeline in batch mode, accommodating different batch processing paradigms simultaneously.
CamBAfx is built on Eclipse technology. Created by Object Technology Inc (now part of IBM, Armonk, New York, USA), Eclipse was originally created as an Integrated Development Environment or IDE (a Graphical User Interface with tools designed to assist programmers in writing programs). It pioneered the concept of an IDE that accepts and integrates diverse set of tools from different vendors. Since then, an integration framework has evolved that has generalised to allow users to build traditional software using a Rich Client Platform (RCP). It was made open source by IBM in 2001 and in 2004, the stewardship of the platform was passed to the independent Eclipse Foundation. It enjoys strong supports from the software industry including BEA System Inc (San Jose, CA, USA), Nokia, Intel and Wind River (Alameda, CA, USA). To date, it has been used to build software as diverse as IBM Workspace Client Technology, Nasa Maestro, Java IDE (JDT) and C/C++ IDE (CDT).
Software for processing neuroimages can be configured as chained modules; each module representing some action on the dataset: subject motion correction, standard space mapping, cross-correlations and so on. Through modularisation, code can be shared and reused. Modules can be mixed and matched to generate the best possible pipelines or experiment with the latest techniques. Bundled into CamBA are the following pipelines:
- Time-series (first-level) analysis of fMRI, standard space mapping and within-group effects
- Second-level linear modelling of within- and between-group effects
- Second level two-way ANOVA: main effects and interactions
- Estimation of spectral properties from fMRI time-series and standard space mapping
Statistical inference modules within CamBA are characterised by the use of permutation methods and spatially extended statistics with nominal type I error control and greater sensitive than voxel-based metrics.
Launched in September 2006, CamBA supports the Nifti data format, hyperlinked HTML output and over-the-net update/extension facilities.
Instructions (Detailed Instructions):
- Download by clicking on the image to the left, then select the download site closest to you.
- Open a command line window
- unzip camba-220.127.116.1120116-linux.tar.gz
- cd camba-18.104.22.16820116
- Download the User Guide to get started
If you have any suggestions or queries, contact John Suckling (js369 AT cam.ac.uk).