LONI Pipeline

LONI Pipeline
Name	LONI Pipeline
Developer	Laboratory of Neuro Imaging
Released	0 2003
Operating system	Cross-platform
Genre	Scientific workflow system
License	Open-source

LONI Pipeline. It is a free, open-source environment for designing, executing, and sharing complex scientific workflows, primarily in the field of neuroimaging. Developed by the Laboratory of Neuro Imaging at the University of Southern California, it provides a visual programming interface that enables researchers to chain together disparate data analysis tools without manual scripting. The system is widely used to automate and reproduce processing streams for magnetic resonance imaging (MRI) and other brain imaging data, facilitating large-scale, collaborative studies in neuroscience.

Overview

The software serves as a graphical workflow engine that abstracts the complexity of command-line tools and high-performance computing resources. It allows users to visually construct pipelines by connecting modular processing nodes, which represent individual executables or scripts from toolkits like FreeSurfer, FSL, and AFNI. This design promotes reproducibility and scalability in computational research, particularly for projects involving the Alzheimer's Disease Neuroimaging Initiative or the Human Connectome Project. By managing data provenance and execution across distributed resources, it has become integral to many studies published in journals such as NeuroImage and Human Brain Mapping.

Architecture and Design

Its architecture is client-server based, where a Java-based client provides the visual interface and a server manages workflow execution on local machines or computational grids. The system utilizes a XML-based pipeline description language to define the structure and parameters of workflows, ensuring they are portable and version-controlled. Processing modules are often wrappers for established neuroimaging software libraries, enabling integration with packages from the International Neuroinformatics Coordinating Facility community. The design emphasizes interoperability with resources like the XCEDE data model and supports execution on cluster computing environments such as Sun Grid Engine.

Functionality and Features

Key functionalities include drag-and-drop pipeline construction, automated data format conversion, and robust error handling for long-running analyses. It features advanced scheduling capabilities for distributing tasks across multi-core workstations or cloud computing platforms, optimizing the use of resources from institutions like the National Center for Supercomputing Applications. The environment maintains comprehensive provenance tracking, logging every processing step for audit trails required by funding agencies like the National Institutes of Health. Additional features support batch processing of large cohorts, conditional branching within workflows, and interactive visualization of intermediate results through integrated tools like ImageJ.

Applications and Use Cases

Its primary application is in automating preprocessing and analysis pipelines for structural MRI, diffusion MRI, and functional MRI data in both clinical and research settings. It has been extensively used in consortia such as the ENIGMA Consortium to harmonize analyses across hundreds of international sites studying conditions like schizophrenia and major depressive disorder. Other use cases include processing data from the ADNI for Alzheimer's disease biomarkers, streamlining tractography analyses in the Human Connectome Project, and facilitating machine learning applications on brain imaging datasets. The platform also supports genomics and multimodal integration projects that combine imaging with data from the Allen Institute for Brain Science.

Development and History

Initial development began in the early 2000s at the Laboratory of Neuro Imaging under the direction of Arthur W. Toga, with funding from the National Institute on Aging and the National Institute of Biomedical Imaging and Bioengineering. The first public release occurred in 2003, coinciding with the launch of major projects like the ADNI. Subsequent versions expanded connectivity to grid computing infrastructures supported by the TeraGrid and later the Extreme Science and Engineering Discovery Environment. Ongoing development is influenced by the needs of large collaborative initiatives, including the BRAIN Initiative, and continues to integrate emerging standards from the Neuroimaging Data Model community.

Related Software and Tools

It exists within a broader ecosystem of scientific workflow systems, including Taverna, Kepler, and Apache Airflow, though it is specialized for biomedical data. It directly interfaces with neuroimaging packages like SPM, CIVET, and MINC, and shares data models with the XNAT platform. Related visualization and quality control tools include FreeSurfer, TrackVis, and MRICron. The development team also maintains the LONI Image Data Archive for sharing processed datasets, and the software often complements resources from the NeuroDebian repository and the Nipype framework for Python-based pipeline development.

Category:Neuroimaging software Category:Scientific workflow systems Category:Free science software Category:University of Southern California