The Java Image Science Toolkit (JIST) for rapid prototyping and publishing of neuroimaging software

Abstract

Non-invasive neuroimaging techniques enable extraordinarily sensitive and specific in vivo study of the structure, functional response and connectivity of biological mechanisms. With these advanced methods comes a heavy reliance on computer-based processing, analysis and interpretation. While the neuroimaging community has produced many excellent academic and commercial tool packages, new tools are often required to interpret new modalities and paradigms. Developing custom tools and ensuring interoperability with existing tools is a significant hurdle. To address these limitations, we present a new framework for algorithm development that implicitly ensures tool interoperability, generates graphical user interfaces, provides advanced batch processing tools, and, most importantly, requires minimal additional programming or computational overhead. Java-based rapid prototyping with this system is an efficient and practical approach to evaluate new algorithms since the proposed system ensures that rapidly constructed prototypes are actually fully-functional processing modules with support for multiple GUI's, a broad range of file formats, and distributed computation. Herein, we demonstrate MRI image processing with the proposed system for cortical surface extraction in large cross-sectional cohorts, provide a system for fully automated diffusion tensor image analysis, and illustrate how the system can be used as a simulation framework for the development of a new image analysis method. The system is released as open source under the Lesser GNU Public License (LGPL) through the Neuroimaging Informatics Tools and Resources Clearinghouse (NITRC).

Figure 1

MIPAV (A) automatically detects the presence of the Plug-In Selector and provides a menu item to open the tool. The Plug-In Selector interface lists all detected plug-ins according to their programmer specified hierarchy (B). Once selected from this menu, a programmatically generated GUI appears for any tool (C) which can operate on any image accessible to MIPAV.

Figure 2

The Pipeline Layout Tool is accessible as a MIPAV plug-in or as an independent program. This interface detects and lists all available analysis tools and input/output interfaces on the left (A). These modules may be dragged into the visual programming interface on the right (B). When a particular module is selected, the options for that module appear in the parameter panel (C). These options may be manually adjusted or connected to “pins” on other modules to enable information flow. Cyclic loops are detected and disallowed.

Figure 3

The Process Manager may be started as a MIPAV plug-in, from within the Pipeline Layout Tool, or as an independent program. Once a layout is loaded, all tasks appear as rows in a large table. Individual tasks or the system as a whole may be started, stopped, or restarted (A). The status of each task is reported (B) in the table. When a user selects a task, all inputs and dependencies for the task are shown in the ancestor pane (C), and once a task has run, its outputs are shown in the result pane (D). Debugging and log information are accessible for running or finished tasks by clicking on a row.

Figure 4

All CRUISE modules were ported to use the JIST API. Rather than using shell scripts to bridge programs, the Layout Tool is used to setup dependent processing steps and validate that each program's output is passing a valid input to the next program (a). This pipeline extracts cortical surfaces from three-dimensional volumetric data in a fully automated manner. It computes local shape metrics on the surfaces which may be examined on an individual basis (column b) or registered and compared in group analyses (column c).

Figure 5

The JIST port of CATNAP combines all functionality available within the original implementation as well as superior file format support and ease of integration with fiber tracking modules (a). Different methods for data modeling, motion registration, or fiber tracking can be simply substituted and explored with the pipeline environment. A visualization of all fibers overlaid with the labeled, right cortical hemisphere is shown in (b) as a typical result which is achievable with this framework. The visualization was accomplished by generating colored vector graphics surfaces and streamlines (in “vtk” format) with JIST and rendering the results in ParaView.

Figure 6

JIST provided an ideal simulation framework for design of the CFARI fiber crossing analysis method. Each step in the simulation (a) could be constructed with modular units for simple reuse. In the simulation process, a noise-free biophysical model was constructed (b and top row in a). Noisy data were then simulated based on this model (c and center left in a). A discrete basis set was constructed for simulation (center right in a). Finally, the simulated data were projected onto the basis set to estimate a fiber crossing model (d and lower row in a).