Frequently Asked Questions
What is Neuroinfo?
Neuroinfo is a software package which was developed by MicroBrightField, Inc. to provide digital microscopy to intranets and the Internet. It provides a web server, and image server, an embedded database as an integrated package which is easy to customize and extend. The images you see are very large. They were created using a microscope with a motorized mechanism which scans many fields of view, and combines them into a large montage which we call a digital slide. Images may represent as much as 80 Gigabytes of uncompressed image data. The system provides access to any field of view at any magnification from any web browser instantly. The web site neuroinformatica.com is an example of the functionality that Neuroinfo provides.Neuroinfo features include
How does the Neuroinfo viewer display huge slide files so quickly?
Neuroinfo acts as an image server, with a Java-based slide viewer technology to serve ultra-high-resolution images rapidly and interactively on the Web. The viewer requests image data from the server one field of view at a time. As the viewer pans and zooms, additional tiles of the image are transferred. Virtual slides can be extremely large; file sizes of tens, even hundreds of megabytes, are impractical to download for viewing. The Neuroinfo technology stores these large images in a pyramidal file format, which stores multiple resolutions from the original source resolution of the microscope lens used for acquisition, down to a macro view of the entire slide or a thumbnail only a few kb in size. Each resolution level is cut into many small tiles. This enables navigating to any region of the slide at any magnification quickly.
Why is a system like Neuroinfo important?
The advent of virtual slide technology has allowed for the widespread sharing of information contained on glass microscope slides. This is a significant technological advance, as previously the only way to share information from slides was with photomicrographs that only contained a single field of view at a single magnification. Virtual slide technology allows remote viewers to examine slides with a functionality similar to that of a microscope; users can pan to regions of interest and zoom in and out freely. Advantages to virtual slides include remote access to slides, access to slides without a microscope, and the ability to see macro overviews of slides not attainable with a traditional microscope. However, the virtual slide cannot stand alone. Information about how the slide was prepared is essential to an understanding of the material; for example, information about the stain used, species, organ, scaling information, and orientation of the tissue section is germane to any information that is to be gleaned from examination of the slide. It is necessary is to present the digital slides along with enough detail about what the slide represents to give the viewer the feeling that they are in the 'presence of an expert'.
How can Neuroinfo be used in teaching?
Neuroinfo can be used to set up teaching web sites for students of histology and pathology, providing not only virtual slides and an interactive web-based viewer, but also information about the slide, instructional notes, links to sources of further information, and annotations overlaid on the slide image.
The University of Iowa is using MicroBrightField Technology to present a student with a slide that does not contain any annotations and is asked to explore and interpret the material. The student makes their own annotations to indicate the locations of the diseased regions of the tissue (if there are any.) The teachers can later verify that the student caught all signs of disease.
How can Neuroinfo be used in research?
Neuroinfo is also a powerful tool for scientific collaboration and cataloguing of slide materials. Remote workgroups can be set up allowing multiple researchers to simultaneously view slide material, add annotations to the slide, and write comments to one another. Using the bioinformatics database that is a part of Neuroinfo, slide material can be quantified, analyzed and compared to other sets of data as a means to support or negate hypotheses about the experimental results contained in the slides.
Neuroinfo can be used to create atlases that show the actual tissue at a cellular level, and demark regions that correspond to a controlled nomenclature. Atlases of many more species and variations will soon be available to the scientific community. These atlases can incorporate changes through the developmental stages of a species and side-by-side views of different developmental stages can easily be created.
How is your technology any different than the online atlases that exist today?
Other atlasing web sites have a wealth of photomicrographs and descriptions. Neuroinfo is unique in the integration of the very-large-image digital microscopy, annotation features, group-oriented exploration and documentation of features and the ability to include additional information including controlled nomenclatures. Our technology can store images that have cellular detail. It doesn't just store one field of view, but an entire slide at full magnification. The contents of Neuroinfo are authored using the web interface by authorized collaborators anywhere on the Internet. Since the entire slide is available the microscopic material can be explored and features discovered in a manner analogous to a traditional microscope.
Isn't it true that data, and not images lead to progress in science?Traditionally, yes. A researcher's published data and conclusions are the cornerstone of scientific progress. Neuroinfo enables the sharing of primary data when it is required. A paper in the PDF or HTML format can be created which links from black and white images to the full color full resolution microscopic material. Any question that the researcher was measuring the appropriate cells in the correct region can be conclusively verified. Other researchers with other specialties may notice and quantify phenomena that they see on the slide, adding to the wealth of information in Neuroinfo, and available to the scientific community. The microscopic material becomes the gateway to a wealth of information about a specific region of a brain whether it's at the level of gross anatomy or at the cellular level.
How does sharing large image sets lead to new discoveries in neuroscience research?
Quoting the NIH Draft Statement on Sharing Research Data:
"There are many reasons to share data from NIH-supported studies. Sharing data reinforces open scientific inquiry, encourages diversity of analysis and opinion, promotes new research, makes possible the testing of new or alternative hypotheses and methods of analysis, supports studies on data collection methods and measurement, facilitates the education of new researchers, enables the exploration of topics not envisioned by the initial investigators, and permits the creation of new data sets when data from multiple sources are combined. By avoiding the duplication of expensive data collection activities, the NIH is able to support more investigators than it could if similar data had to be collected de novo by each applicant." (NIH 2002. NIH announces draft statement on sharing research data. http://grants.nih.gov/grants/guide/notice-files/NOT-OD-02-035.html.)
The Report on Neuroinformatics from the Global Science Forum Neuroinformatics Working Group of the Organization for Economic Co-operation and Development (June 2002) states:
"A very large number of researchers working on many thousands of projects
world-wide are accumulating these data. But measurements made by individual
groups in the context of distinct research projects are often difficult to share
in a form easily exploited by the rest of the neuroscience community. Journal
publications - the current method of sharing data - present results in a highly
condensed format of representative data or average values. The community
urgently needs a way of sharing primary data." (23. The Report on
Neuroinformatics from the Global Science Forum Neuroinformatics Working Group of
the Organization for Economic Co-operation and Development (June 2002))
How are Virtual Slides created?
The partner technology to Neuroinfo is the Virtual Slice Module in the Neurolucida and Stereo Investigator software products from MicroBrightField. This Virtual Slice module allows for automated image acquisition of an entire microscope slide at any magnification available on the microscope. The software drives the motorized stage to acquire all fields-of-view contained in the region of interest, then seamlessly stitches the fields into a single image montage. Using innovative technology from Zoomify, the images are compressed and stored in the pyramidal format described above. If acquisition systems are not available, MicroBrightField offers slide scanning services. MicroBrightField does not use its own proprietary file formats for its virtual slides; all slide files are industry standard formats viewable by 3rd party software.
Can permission to view slides be restricted?
Neuroinfo handles authentication and permissions. Permission groups let the designated administrators create groups of users that have permission to view and edit the virtual slides and database. These users can be granted permission on a view-only basis, or can be granted permission to view, edit, and make entries in the database.
How is the Neuroinfo database organized?
Data about slides, groups of slides, and contributors is stored in a relational database. The database architecture scales to track new kinds of information easily. The information is stored as small chunks of data, each of which correspond to a carefully crafted Definition. The different small chunks of data can be linked to each other to form groups or to one another in any combination, allowing for the creation of rich sets of connected data. All textual data is also indexed for powerful text searching.
Example: Adding A Category In order to categorize slides by ethnicity of the subject, then a definition would be created using the web interface for "Ethnicity" and then this could be populated with all the different ethnic groups. Ethnicity would automatically show up on the search page as a filter, and each ethnic group would have a page that shows all of the corresponding slides. If one researcher enters "White" and later another enters "Caucasian", there is an administrative tool to remove the unwanted duplicate, and move all associations to the correct value.
There are three distinct types of definitions that Neuroinfo uses for all contents and allows you to create.
What classification scheme do you use to index serial sections?
We support any indexing scheme by means of our innovative database structure. If the image represents pixels in an X,Y metric coordinate system, each slide also stores the Z position of the section. Additional stereotaxic parameters and relationship to gross structures are easily specified and immediately available for categorizing and searching.
What curatorial mechanism exists to ensure data quality?
Every modification of the database updates a timestamp associated with the data. At the moment, we show recently modified database contents in a list on the home page. These can be reviewed and edited by a system administrator, or a collaborator who has editing privileges for the project the data belongs to. Mechanisms such as sending e-mail to a curator whenever data is edited can be provided.
Can Neuroinfo provide for the Bioinformatics need for my entire project?
Definitions of bioinformatics and neuroinformatics are broad and varied. The central figure in bioinformatics is the biologist, and like a microscope the computer is now an essential tool for the biologist's understanding and research. Neuroinfo's unique capabilities will have a role in the inevitable revolution of access to microscopic material and the understanding of that material over the next decade.
Neuroinfo provides the foundation for a rich integration of digital microscopy, metadata necessary for the correct interpretation of the microscopic material, and all analysis results of the microscopic material. Resources from external databases and XML files can easily be incorporated and links to relevant genetics databases can be created. External web sites and mediator applications can be written in almost any programming language to query Neuroinfo using the XML-RPC protocol.
What skills do I need to tailor Neuroinfo to my needs?
Changing the look of the web pages can be done easily with any html editor. We recommend Dreamweaver from Macromedia. Skills in graphic design are also desirable and can have a dramatic impact on the enthusiasm surrounding your site.
Since Neuroinfo provides a simple interface to the SQL database, no database expertise is needed. The Neuroinfo database is seeded with the most common categories likely to be used, including: staining technique, tissue, species, and section plane. New categories can be easily added from the administrative interface, and new items can be added to a category at any time. An innovative feature allows for the combination of redundant entries, so that a structure with several different names can be found with a single search of the database (for example, the word "putamen" can be combined with "lenticular nucleus" and "basal ganglia"). A basic facility with a computer is all that is needed to set up your own custom database.
Can I use the Neuroinfo viewer in my own institutional or personal web page?
Yes. The viewer can be used with existing web sites. The viewer does require an image server, and there are several methods of integrating resources to enable virtual microscopy on your pages. We provide templates of database access from external web sites written in ASP, and PHP to do text searches, browsing contents by category, and virtual slide viewing and annotation.
How do you handle the massive volume of data that these large slides require?
One of the largest sets of teaching material that is currently available to anyone in the world is the teaching collection at the University of Iowa. The collection consists of about 650 slides stored in the FlashPIX format which uses JPEG compression internally. The entire collection takes about 70 GB of hard drive space in its compressed form. In just the last year, the price of a hard drive that can store 70GB of image information has dropped to about $80 US. A high performance web server that uses a RAID array to speed access and provide fault tolerance and salability might cost $3,000. Images are single files, and can be copied from computer to computer just like any other computer file.
Is Neuroinfo actively being developed?
MicroBrightField, Inc. was recently awarded an NIH SBIR grant which will fund the continued development of Neuroinfo. MicroBrightField's customers drive the priority of new features and we will go out of our way to develop functionality necessary for a particular project.
How much does Neuroinfo cost?
Neuroinfo is available from MicroBrightField, Inc. at a low one time cost plus a yearly maintenance contract. The cost is very low compared to the cost of having a software developer reinvent the technology. Day to day use of Neuroinfo does not require any technical support staff or programming skills. If you do have a programming staff and would like to extend Neuroinfo for your needs the Neuroinfo database layer makes complex modifications easy. Maintenance includes the customization of web page functionality for specific purposes by MicroBrightField's staff.
What are the system requirements for Neuroinfo?
Neuroinfo will run on any machine running Windows 2000, Linux or Solaris. The power of the machine controls the number of simultaneous users you need to serve images to at one time. A 1GHz server with 256MB of RAM will handle about 12 simultaneous users all panning across virtual slides at once. Since users typically zoom in and examine the features of one location before moving on to another location, dozens of users can use a low end server at once. We highly recommend (and will configure and resell) Dell servers with 1GB of RAM, a 2.8GHz processor and a RAID array for fast disk access.
When a customer needs to handle hundreds of simultaneous viewers at once it will probably require more than one image server. For example, a class of 160 students needs to do some work just before (or during) exams. We want to ensure that the students can all see one field of view in two seconds. The server will have to be able to serve about four thousand tiles per minute. One high end server can service about 800 FlashPIX tiles per minute or about 1600 Zoomify tiles per minute. (Zoomify version 106 files are optimized for quick serving.) You will need three to five servers to handle the peak load. We provide the mechanism to delegate the image serving process to multiple machines. Since the peak load only happens a few times a semester, regular lab computers can be set up as supplemental image servers when necessary.
What is load balancing?
Load balancing is a technology that utilizes multiple servers to share work. When one server receives incoming requests it intelligently routes requests to the least busy server. To the outside world, it appears that a single server handles the work, but in reality, several machines in the background may be performing the processing. Neuroinfo can perform load balancing for you if you install it on several servers.
Can I create 3D models of the serial sections on Neuroinfo?
Not Directly. MicroBrightField, Inc. also creates Neurolucida which has a 3D Solid Modeling module. You can see some of the results of the solid modeling module in the Neurolucida Gallery. Using Neurolucida in conjunction with virtual microscopy makes the 3D modeling process as well as quantification of your serial reconstruction very efficient.
What aspects of Neuroinfo are open to the community and which are proprietary to MicroBrightField?
All data stored using Neuroinfo is available to anyone with sufficient permissions using a variety of data formats and software tools. The images that store the microscopic material are stored in formats that can always be converted to standard nonproprietary formats. In special cases, a source code escrow can be established in the event that MicroBrightField can no longer support the product. The only proprietary parts of Neuroinfo are the viewer, the image server and the database layer which make it easy to specialize and extend. Even the internals of Neuroinfo such as the embedded database and the tools which create the charts on the Image Server Statistics are all open source tools. The .jsp pages that format and present your data are also provided in full and we encourage you to modify them for your specialty.
Can dial-up modem users access the digital microscope?
Yes-- if they are patient. A good modem connection gives about 48,000 bits per second. One tile of one image is typically 80,000 bits, and one field of view is about nine tiles... so a modem user might have to wait about 18 seconds for one small field of view. There is no technical problem to solve to support modem users, modems are simply limited in their transfer rate. Any viewer with DSL or cablemodem will load single fields of view within about two seconds. Using the technology on a Local Area Network gives lightening fast panning and zooming.
How are annotations stored (what format?)
Annotations are stored in an XML-based format which describe the parameters specified when the label is drawn. Coordinates are stored in microns so that lengths and areas can be calculated without any conversion. When a boundary contour is drawn, there are several parameters which pertain not only to what is drawn, but what the region indicates, and how it behaves. Other labels such as arrows store the range of magnifications at which the arrow should be drawn. An arrow drawn at the low magnification gross view isn't useful when viewing the region at the cellular level. The architecture of the viewer allows any number of annotation tools to be added in the future. The annotation format is designed to grow as new annotation tools are needed.
Can I use your annotation capabilities on my collection of JPEG files?
Yes. The viewer can load and pan around .jpg files by loading the entire image at once. The annotation and quantification tools are identical to those used for virtual slides. This enables researchers who already have large collections of photomicrographs to use the collaboration and quantification capabilities of Neuroinfo.
Does it work with Zoomify?
Yes. We are completely compatible with Zoomify file formats, viewers and annotations systems. Neuroinfo provides a complete environment for people interested in Zoomify's flash-based viewer and annotation system. The annotator will even store annotations in our embedded database. We are continuing to use Java technology for our own web-based microscopy development. Neuroinfo would replace Tomcat in the zoomify instructions.
Does it work on the Macintosh?
Yes and no. The Neuroinfo server software will run slowly on Mac OSX 10.2.2 w/ Java 1.4.1. The Java team at Apple are actively working on improving Java support, and we expect that OS X will make a good server someday, but it is not supported yet.
People who are viewing Neuroinfo content from a Macintosh will have no problems if they are running OS X version 10.2 (or later) with Java version 1.3 (or later) and the Mozilla 1.2 (or later) web browser. Mac users who are using Microsoft Internet Explorer will be able to view slides, but annotations will not control the viewer properly because IE does not support LiveConnect on the Macintosh. Viewers will see a yellow link at the top of every Neuroinfo page that encourages using the Mozilla browser for this reason.
|Click here to obtain a link to this page|