Brain Information Group White Paper
The Information Infrastructure Needs of Neuroscience Research:
Opportunities and Issues of Implementation
Executive Summary
In the spring of 2003, the leadership of the Society for Neuroscience met with the Directors of the NIH's neuro-related Institutes to examine areas of mutual interest in which the Society could usefully serve the neuroscience community and the strategic interests of the Institutes. One endeavor was selected for further evaluation: the information infrastructure needs of neuroscience research. Within this challenging opportunity, three specific sets of tasks were identified: 1) assessment of the breadth, depth and accessibility of current neuroscience databases; 2) the opportunities for, and benefits of, integration of these and future neurodatabases; 3) the requirements and continuing support options for maintenance and refinement of an information infrastructure for the neuroscience community.
With support from the Wadsworth Foundation, the SfN Council under the leadership of President Huda Akil appointed a Task Force, the Brain Information Group (BIG) to consider these issues. Three meetings were held, two during the summer of 2003, and one just prior to the 2003 Annual Meeting. This White Paper reflects the discussions, agreements, and concerns expressed at these meetings, and suggests some pathways, opportunities, and requirements for incorporating these discussions and concerns into the functions of the Society for Neuroscience.
Initial discussions among the BIG revealed that an unexpectedly large number of detailed databases for neuroscience data already exist funded through a number of sources, most notably the Human Brain Project. Those known to the BIG serve communities of investigators ranging from individual laboratories to communities sharing focus on specific molecular entities (such as olfactory receptors or G-protein coupled receptors), to species-focused groups (for example, C. elegans and Drosophila), or region-focused groups (such as hippocampus or cerebellum) to broader issues of chemical neuroanatomy or physiology. However, while these databases are often well known within those particular communities, they are not well known to the general population of neuroscientists, nor, with few exceptions, were they initially designed to facilitate searching and comparisons across databases by the general neuroscience community.
According to the BIG discussions, integrated neuroscience databases can provide SfN members and the broader community with several immediately evident opportunities. As the methods of data collection operating at all levels of the current major experimental strategies in the neurosciences result in ever increasing collections of data, it becomes more and more difficult to gain awareness and maintain incorporation of the new data within any individual's working concepts and paradigms. As a result, new data may go unrecognized, inconsistencies between datasets go unchallenged, and unsupportable hypotheses on which experimental data were collected remain overlooked. A broad spectrum of integrated, organized, collated repositories of extant peer-reviewed and un-reviewed data and claims drawn from those data would benefit the community by providing interested scholars the tools to explore and compare the growing masses of data. Such data repositories and analytical tools could accelerate the pace and enhance the quality of the neuroscience research enterprise.
Following initial discussions, three goals were set for the BIG Task Force
- To inventory neuroscience relevant databases and demonstrate for SfN's Council and other funders what currently exists in the public domain;
- To identify strategies by which the Society for Neuroscience could promote broader and more integratable information infrastructural tools to place past, present and future neuroscience data in the public domain;
- To consider creation of a database portal describing the currently accessible publicly available neuroscience databases, and assess interest of the membership in expanding this portal and meeting needs not being currently addressed.
Many such discussions in the past have foundered over the diversity of terms employed by different groups of neuroscientists to designate brain structures or functions, over the issue of quality control for data being included or excluded, and over the diversity of equipment required to devise a system with universal access for database access, and display. While the emergence of the Internet and Web-based tools for access to databases together with generally available high speed Internet access have diminished concerns over equipment requirements for sharing and access, concerns over variable terminologies, and the quality control over data to be included or excluded persist.
After further discussions, and examination of existing terminology translations, the BIG concluded that concerns over nomenclature variations from database to database are approachable through the National Library of Medicine's Unified Medical Language System (UMLS). The UMLS at this time would appear to offer a basis to expand the neuroscience information infrastructure effort and a platform on which to develop broader neuroscience-specific ontologies. A expansive consideration of the terminology and ontology issue would include the keyword systems employed to index the content of abstracts, articles, reviews and other neuroscience data, including those used in abstracts for the annual meeting, and articles published in The Journal of Neuroscience.
The BIG proposed a demonstration project that could display the current spectrum of publicly accessible databases of neuroscience information in order to probe the interests of the Society membership. Such a demonstration project would also serve to probe the willingness of Society leadership to engage the issue of creation and maintenance of integrated neuroscience databases for the long-term. A subgroup of BIG (David Van Essen, Luis Marenco, Daniel Gardner, Maryann Martone, and Gordon Shepherd) then generated a prototype database gateway.
The BIG reported to President Huda Akil that they were prepared to recommend: 1) conversion of the BIG Task Force to an Ad Hoc Committee on Neuroinformatics to express the leadership's acceptance of these and other longer term opportunities; or 2) creation of a standing committee which could then interact with the Program Committee, the Publications Committee, and the Society's existing and future 'publications' to enhance the dissemination and employment of the data of the neurosciences. In response, at the final formal BIG meeting, President Akil announced that based on the interim report from the BIG, Council had agreed to move immediately to create a Committee on Neuroinformatics and had endorsed an effort to apply for funds to support the initial gateway to neuroscience databases (Neuroscience Database Gateway, NDG).
On April 13, 2004, Marty Saggese, Executive Director of the Society for Neuroscience announced that the Society had received a contract from NIH (funded by NIMH, NINDS, and NIDA) to support the opening of the planned NDG and to determine membership interests in expanding the role of the Society in information infrastructure under the leadership of the Committee on Neuroinformatics. This White Paper will appear with the opening of the NDG (http://www.sfn.org/ndg) . The Society's new Committee on Neuroinformatics will monitor the utilization of the databases and the feedback from the membership in an associated survey of NDG users when the Committee embarks on their development of an agenda for the Society in neuroinformatics.
The BIG concludes that as we enter the electronic era, it is becoming critical to communicate information in new ways. Linked, interoperable databases are central to the future of knowledge dissemination. Successful implementation of integrated databases of neuroscience information could benefit theoretical and experimental scholars, educators, funders and the general community by establishing effective access to current activities and emerging knowledge profiles of what has been reported and what has been learned. Gaps in the indices of reported observations could illuminate important neglected areas for future research opportunities.
The convening power of the Society provides a unique opportunity to develop community-based standards for data ontologies, for review and inclusion standards for the data that would populate neuroscience databases, for the selections of linkages to, and integration of, information across databases, and for the compilation of supplemental data files to augment conventional publications. The BIG encourages the Committee on Neuroinformatics to coordinate their efforts with both the Program Committee and the Editors of The Journal of Neuroscience to evolve a mutually acceptable comprehensive ontology of neuroscience index terms. For neuroscientists to be open about contributing their data to the database and to rely on the data within it to plan their future research, significant barriers will have to be overcome. An important undertaking for the Committee on Neuroinformatics will be to assess the feasibility to develop opportunities and incentives to lower the threshold for data sharing. The Society for Neuroscience Committee on Neuroinformatics could play an important role in helping to meet the recently mandated objective of the NIH to enforce data sharing.
Through the Committee on Neuroinformatics, the Society for Neuroscience could promote development of a network of information repositories to span both clinical as well as basic neuroscience, facilitate access to information about the myriad brain diseases, expedite the translation of basic research discoveries into improved diagnosis, treatment and prevention, and thereby serve to advance the pace, depth, and quality of neuroscience research of the future, with tangible benefits to society. Related longer term opportunities include the potential for enhanced interactions with neuroscience-focused charitable organizations, and neuroscience research funders of the future, as well as the larger issues of the dissemination, integration, and analysis of new and legacy neuroscience information and concepts. Such infrastructural enhancements could serve many purposes including improved communication with legislators and the public about the relevance and importance of neuroscience research and greater exposure of young scientists to the exciting opportunities in the field of neuroscience.
Background
In the spring of 2003, the leadership of the Society for Neuroscience, through efforts initiated by President Huda Akil and then President-elect Story Landis, met with the Directors of the neuro-related NIH Institutes to find areas of mutual interest in which the Society could usefully serve the neuroscience community. (When subsequently appointed Director of NINDS, Dr. Landis, a supporter of the effort described below, resigned as president-elect, while the effort described here was still underway.) After extensive discussion, one area selected for further evaluation was the information infrastructure needs of neuroscience research. Within this challenging opportunity, three specific sets of tasks were identified: 1) a Web gateway to currently existing neuroscience-related databases; 2) the opportunities for, and benefits of, integration across these and future neurodatabases; 3) the requirements and continuing support options for maintenance and refinement of a portal to information currently accessible in neuroscience databases and knowledge bases for the neuroscience community.
To date, there have been several efforts supported by individual NIH institutes and by multiple other federal agencies to create various classes of brain databases. These efforts include the pioneering efforts of the Human Brain Project (NIMH, NIDA, NIAAA, NASA, NSF and other funders), the strategic approaches being undertaken by the Biomedical Informatics Research Network (BIRN; funded by NIH's NCRR) and other programs, including the National Library of Medicine. Most of these databases focus on particular species (e.g., C. elegans, Drosophila, rat, mouse, human, non-human primates), or on particular approaches (e.g., three-dimensional imaging of brain regions or pathways, molecular anatomy, and the like). Despite the richness of these active databases, they do not constitute a coherent and integrated entity, but instead span different platforms of computers, database management systems, with distinct scientific nomenclatures and their emerging ontologies, and differing research tools for recovery and mining of their contents. Even more important, the broad neuroscience community remains largely unaware of what data are available or how to access them.
To confront these challenges and determine what role the Society for Neuroscience could play should it choose to do so, an Ad Hoc Task Force on Brain Information (Brain Information Group [BIG]; see Appendix for members) was formed to survey the current state of the field and to generate a strategic plan for how best to proceed. The effort was enabled through a grant from the Wadsworth Foundation of Seattle, Washington. This White Paper reflects the discussions, agreements, and concerns expressed at these meetings, and suggests some pathways, opportunities, and requirements for incorporating these discussions and concerns into the functions of the Society for Neuroscience.
BIG Meetings
Three BIG meetings were held:
Chicago, July 28-29, 2003.
Initial discussions among the BIG revealed an unexpectedly large number of detailed databases for neuroscience data that already exist. Those known to BIG serve communities of groups ranging from individual laboratories to communities sharing focus on specific molecular entities (such as olfactory receptors or G-protein coupled receptors) to species-focused groups ( for example, C. elegans and Drosophila) to broader issues of chemical neuroanatomy or physiology. However, while these databases are often well known within particular communities, they are not well known to the general population of neuroscientists, nor, with few exceptions, were they initially designed to facilitate searching and comparisons across databases by the general neuroscience community.
According to the BIG discussions, integrated neuroscience databases would serve to benefit SfN members and the broader community in several immediately evident opportunities. As the methods of data collection operating at all levels of the current major experimental strategies in the neurosciences result in ever increasing collections of data, it becomes more and more difficult to gain awareness and maintain incorporation of the new data within any individual's working concepts and paradigms. As a result, new data may go unrecognized, inconsistencies between datasets go unchallenged, and unsupportable hypotheses on which experimental data were collected remain overlooked. An integrated, organized, collated repository of extant peer-reviewed and un-reviewed data and claims drawn from those data could benefit the community by providing interested scholars the tools to explore and compare the growing masses of data. Such data repositories and analytical tools could accelerate the pace and enhance the quality of the neuroscience research enterprise.
Therefore at this initial meeting, three specific goals for the BIG activity were recognized:
- To inventory neuroscience relevant databases and demonstrate for SfN's Council and other funders what currently exists in the public domain;
- To identify strategies by which the Society for Neuroscience could promote broader and more integratable information infrastructural tools to place past, present and future neuroscience data in the public domain;
- To consider creation of a database portal describing the currently accessible publicly available neuroscience databases, and assess interest of the membership in expanding this portal and meeting needs not being currently addressed.
Such discussions have been held many times in the past, and have generally foundered over the diversity of terms employed by different groups of neuroscientists to designate brain structures, over the issue of quality control for data being included or excluded (some arguing for or against the inclusion of non-peer reviewed data), and over the need for multiple media to be employed to provide data incorporation and access for graphical, textual, and time series data (e.g., audio, video, and electrophysiological). Prior barriers have also reflected the diversity of equipment needed for database deployment, for data acquisition, database access, and data display.
Concerns over nomenclature variations from database to database were seen to be approachable through employment of the National Library of Medicine's Unified Medical Language System (UMLS). At this meeting, the BIG resolved at its second planned meeting to explore existing effective approaches to time series data, and the existing expertise in the development of complex sequence analysis and comparison algorithms that have proven essential in deriving the multiple existing genomic databases and their curation, annotation and mining. Lastly, the emergence of the Internet and Web-based tools for access to databases together with generally available high speed internet access have diminished concerns over equipment requirements for sharing and accessing public databases.
Washington, DC, September 9-10, 2003
As planned, the second meeting welcomed additional experienced neuroscientists. Douglas Bowden (University of Washington) described his achievements in employing the Unified Medical Language System to organize and translate terms employed for anatomic entities across mammalian brains from the perspective of regions defined for the primate (human and non-human) brain. Daniel Gardner (Weill Cornell Medical School) described his achievements in creating and maintaining a common data model system for sharing electrophysiological data. That system contains meta-data descriptions to define species, structures investigated, methods employed, and experimental conditions, as well as tools for data deposition and data analysis. Richard Mural, (Celera, Inc.), reflected on some of the conditions for high speed data collection and analysis used in genomic database construction, and some of the difficulties he could envision that any integration of databases for the neurosciences would have to solve.
In the discussions that followed, there was recognition that while the BIG participants were committed to continue their efforts in developing data collection and analysis tools for their own and the communities' research programs, the interests of the larger neuroscience community in such efforts have yet to be ascertained. Furthermore, any desire, let alone willingness of the broader community to endorse a role for the Society in the nurturing of the effort to create an integrated information infrastructure for the neurosciences were also unknown.
The BIG then began to consider the opportunities for a demonstration project that would display the current spectrum of publicly accessible databases of neuroscience information in order to probe the interests of the Society. Such a demonstration project would also serve to probe the willingness of Society leadership to engage the issue for the long-term. A subgroup of BIG (David Van Essen, Luis Marenco, Daniel Gardner, Maryann Martone, and Gordon Shepherd) was encouraged to generate a prototype database gateway.
Should the Society resolve to promote a general information infrastructure for service to the research of the membership, it would then be possible to envision longer term opportunities: the potential for service to neuroscience-focused charitable organizations, to neuroscience research funders of the future, as well as the larger issues of the dissemination, integration, and analysis of new and legacy neuroscience information and concepts. Such an infrastructure could serve many purposes ranging from congressional education to funding prioritization to education in the neurosciences.
For these longer term expressions of interest, the BIG reported to President Huda Akil that they were prepared to recommend: 1) conversion of the BIG Task Force to an Ad Hoc Committee on Neuroinformatics to express the leadership's acceptance of these and other longer term opportunities; or 2) creation of a committee which could then interact with the Program Committee, the Publications Committee, and the Society's existing and future publications to enhance the dissemination and employment of the data of the neurosciences.
New Orleans, November 8, 2003
BIG Members and advisers gathered for several hours to assess the status of efforts to gather data on existing neuroscience databases whose creators were willing and able to provide access to the community. The group agreed on the desirability of expanding the prototype version into a Society-sponsored gateway to neuroscience databases of three categories: 1) those containing substantial neuroscience-related experimental data that is publicly accessible; 2) those containing neuroscience knowledge; 3) databases of tools, tool registries, and links to other broad neuroscience portals of information (e.g., Human Brain Project, BIRN, BrainInfo sites), including keywords and other attributes of these linked databases.
At this meeting, President Akil announced that based on an interim report from the BIG, Council had agreed to move immediately to create a Committee on Neuroinformatics. It was also noted that the Society leadership had endorsed an effort to apply for funds to support the initial gateway to neuroscience databases (Neuroscience Database Gateway, NDG), and to bring to the attention of the neuro-related NIH institutes and other funders the opportunities such an information infrastructural effort might provide. At a meeting with those Directors, initial support for such funding applications was given.
On April 13, 2004, Marty Saggese, Executive Director of the Society for Neuroscience announced that the Society had received a Phase I NDG contract award from NIH in the amount of $74K to cover the period from April-August, 2004, with a phase II award for approximately the same amount to cover the period from August-December, 2004. These funds would support the opening of the planned portal and determine membership interests in participating in extending or refining such efforts to fund the Committee on Neuroinformatics initial explorations.
A starting Neuroscience Database Gateway (NDG)
Gordon Shepherd and Luis Marenco from the Yale SenseLab had already created a list of current public neuroscience databases at their site http://senselab.med.yale.edu/senselab/. BIG's review of the excellent SenseLab site reveals many of the properties that would be desired for a database of the currently available, neuro-databases. A uniform template was devised by the BIG database gateway subcommittee (Van Essen, Marenco, Gardner, Martone, Shepherd) and was sent to creators of some 75 known neuroscience databases to describe the content of their database in the gateway. Each database organizer has been challenged to create a file describing its database's metadata, including defined access protocols so that interoperability could eventually occur. This file will reside within each Web site so that any user who comes to a site can access the metadata. The assessment of user accesses by the Committee on Neuroinformatics will indicate the types of databases particular users find useful. Database organizers are free to decide what the appropriate metadata for their site are.
The database gateway subcommittee also formulated a Web-based survey that will be used to assess the current views of SfN members on neuroinformatics issues. The survey questions deal with perspectives on 1) the NDG site itself and whether its capabilities should be expanded; 2) Neuroscience Databases (public/shared) and Software Tools; and 3) data management issues within individual laboratories.
The announcement of the Neuroscience Database Gateway (http://www.sfn.org/ndg) to the SfN membership will occur with the release of this document. Members will also be encouraged to respond to the survey, which will be accessible from the NDG site. The Society's new Committee on Neuroinformatics will monitor the presentations of the databases and the feedback from the membership as they embark on their development of an agenda for the Society in neuroinformatics.
Creating this type of demonstration experiment was considered not only to be helpful in identifying the future hurdles of any larger scale project but would also help to assess the membership's interests about such a project and its capabilities. The success of a small-scale demonstration interoperability project would help validate a larger project and its inherent costs, especially if the demonstration receives positive feedback from the membership.
How can/should the Society for Neuroscience be involved in the development of neuroscience information infrastructure?
Should the Society for Neuroscience maintain the Neuroscience Database Gateway and expand its capabilities? Having the Society solicit members' interest in participation is an essential first step in documenting the reality of the potential opportunities envisioned by the BIG Task Force for integrated, searchable data repositories.
Should the Society for Neuroscience become engaged in promoting the development and continued evolution of common ontologies that facilitate searching within and across neuroscience databases? The convening power of the Society would offer the opportunity to develop community-based standards for data ontologies, for review and inclusion standards, and for the compilation of supplemental data files to augment conventional publications.
However, any such database effort will need to be sustained and maintained for a significant period to allow participants to understand the opportunities for full-scale benefits to individual members. In order to maximize potential benefits to the membership, the effort will require maintenance for currency, corrections, and additions, and will therefore be an ongoing need, not a one-time tool creation effort. Use of this NDG tool will in some ways require the reeducation of neuroscientists in where they turn to gain their data awareness, and how to make data collected available to others for replication, validation, and re-analysis. Thus, there will need to be an explicit commitment made by the Society, and this expanded mission may exceed the membership's views of what the Society should undertake. Through recent surveys, membership has endorsed public education as well as political action as being within the Society's mission. Membership endorsement of BIG's proposals will therefore be very important. Society leadership is also critical to coordination of the information infrastructure with The Journal of Neuroscience and to communicating issues to the public and to the membership. In the past, scientific societies have established journals to communicate scientific knowledge. As we enter the electronic era, it is becoming critical to communicate information in new ways. Linked, interoperable databases are the future of knowledge dissemination. These strategic considerations may thus be viewed as initial items for the starting agenda for the Committee on Neuroinformatics.
Does the UMLS 'linkage' mechanism satisfy current immediate interoperability concerns?
As the BIG effort transitions to the SFN Committee on Neuroinformatics, significant effort will need to be focused on the use of the UMLS as a solution to the terminology and ontology issue. Given the breadth of nervous systems under investigation would require means for continued expansion of the vocabulary, addressing terminological linkages, spatial linkages, and spatial coordinate systems. One of the most significant database concerns is that there is no single neuroscience vocabulary when ideally there should be. The UMLS of the National Library of Medicine group builds the vocabulary. Over time, new terms and term-relationships are incorporated into other vocabularies within UMLS. Clearly, it will be important for the Committee on Neuroinformatics to develop good working relationships with this UMLS group.
For example, some important queries to be posed would include: Does UMLS allow different cortical partitioning schemes? Is it hospitable to multiple descriptions/partitions for the same area? If there is a new concept introduced, when will it get its own unique and distinguishable identifier? When would data negating the entity be applied, and how would past links of the hypothesis be expunged? Could multiple differing, nested, non-overlapping, definitions of any entity share the same string but be distinguished by unique identifiers such that they will be related but not identical? How can the Committee on Neuroinformatics help create semantic networks in which different points of view of the same data are tied together in unique ways? Although strictly speaking, the UMLS is not a controlled vocabulary, the system has a unique 'preferred term' for every concept. Database creators are free to adopt those terms as a controlled vocabulary (see [1, 2] for examples). Nevertheless, the UMLS at this time would appear to offer a means to begin the information infrastructure effort and a pathway to develop other ontologies.
A broader consideration of the terminology and ontology issue would include the keyword systems employed to index the content of abstracts, articles, reviews and other neuroscience data. A consistent but expandable, nested, hierarchical ontology organizing the entities and concepts of research in the neurosciences could benefit theoretical and experimental scholars, educators, funders and the general community by establishing a consistent index to current activities and emerging knowledge profiles of what has been reported and what has been learned. Gaps in such indices of reported observations could illuminate important neglected areas for future research opportunities. Thus, the Committee on Neuroinformatics could cooperate effectively with both the Program Committee and the Editors of The Journal of Neuroscience in the development of such an ontology of neuroscience index terms.
At some point in the not too distant future, the Committee on Neuroinformatics should begin to engage with the leadership of the Society's journal to sense their willingness to endorse and enforce inclusion of peer-reviewed data sets in appropriate information repositories, when these are available, and the development of incentives for authors and reviewers to utilize the tools for analysis of already deposited data for replications and inconsistencies. Support for neuroscience databases at The Journal has already been discussed [3] .
Concerns of the data providers
For neuroscientists to be open about contributing their data to the database and to rely on the data within it to plan their future research, significant barriers will have to be overcome. Investigators have a significant time and money investment in collecting their data. The best approach for individual investigators must be seen as the best approach for the community and the field of neuroscience: your data may be useful to others in ways that will also benefit you and the way you approach your project. The fMRI consortium database at Dartmouth College has provided an important and apparently successful model. Data appearing in the peer reviewed articles of the Journal of Cognitive Neuroscience are accessible, and awards are given for data mining of those data leading to novel insights in the data pool, often outside the interests of the original authors [4, 5] .
An important undertaking for the Committee on Neuroinformatics will be to assess the feasibility to develop opportunities and incentives to lower the threshold for data sharing. Would The Journal of Neuroscience or any journal reporting neuroscience data be willing to offer free publication to those authors who provide their complete body of data in a manner suitable for entry into a publicly searchable database and be prepared to provide the original author a chance to comment on any papers making use of their banked data? Data sharing agreements through a journal or a Society-sponsored data repository will be recognized as a more formal socio-scientific commitment, but would of course not preclude the sorts of informal sharing of unpublished data that are today the norm.
The Society's Committee on Neuroinformatics could be a community voice to establish any future database user rules to define the limits or processes for how contributed data in the infrastructure's repository are to be used. For example, 'data download rules' might carry an explicit agreement by the downloader to abide by a set of Society-endorsed specific terms of data sharing. Another extreme prototypic 'rule' could be that the first publication of a completely new data-derived finding has to include the original data producers as co-authors. Perhaps the deliberations and feedback from membership will lead the Committee on Neuroinformatics to recommend an "intermediate space," a partial data-sharing stage on the way to completely open access. It will be very important for a researcher to include information in any publication on willingness to re-purpose without permission or desire for coordination for all shared data. User agreements will also have to contain warnings: caveat emptor, especially for data that are being deposited before peer-review, if such a step is to be permitted. A group of investigators currently funded by the Human Brain Project have recently proposed some guidelines and principles for neuroscience data sharing, including protection of the original author's intellectual property, the potential re-analyses of those data, and the beneficial effects they see emanating from such cooperative data sharing. [6]
Clinical databases pose important additional challenges (e.g., human subject protection and privacy issues), and may be beyond the province of the Society for Neuroscience to engage without collaboration with the societies of the clinical neurosciences. This interface could provide an impetus for Society leadership to initiate outreach to those societies. Eventually, the forward and backward translation of clinical problems and preclinical discoveries must be a part of the neuroscience information infrastructure. The Society for Neuroscience Committee on Neuroinformatics could eventually lead to an information repository that would be the hub of a human genome brain database with annotations for specific systems of innervated tissue disorders.
How can SfN engage funders and the community for long term support?
Long-term funding from government agencies may not materialize. As the paylines for new grants are reduced by the conclusion of the budget doubling annual incremental increases, pressures will mount to maintain individual grant numbers and projects, although the support of the NIH Directors of the main neuro-related Institutes for large scale neuroscience efforts has just been publicly asserted [7] . A negative attitude toward funding community infrastructure could nevertheless be anticipated in the minds of peer reviewers: "You received funding to generate data, why do you need more money to maintain the data as well?" Thus, for the sake of survival, any community-directed databases may need new and different funding mechanisms. So long as individuals hold responsibility for maintaining their own databases, who is to be held responsible for the sharing of such data in distributed but linked publicly accessible databases? While NIH has decreed that data sharing for grants receiving more than $500,000 in direct annual support will be mandatory as of October 2003, no mechanisms to cover the costs of such data sharing have been determined, and no mechanisms for the actual sharing have been proposed. Thus, the Society for Neuroscience Committee on Neuroinformatics could play an important and potentially fundable role in meeting this objective of data sharing.
The ideal funding model should be multi-sourced. Neurological and psychiatric charitable foundations could be envisioned as potentially funding part of this effort. Ordinarily, public charities and the donor public in general want money to go directly to research or outreach/support services/patient support. If the Committee on Neuroinformatics can demonstrate the importance of the neuroscience informatics infrastructure, the Society could become an important intermediary for such charitable funding, and in turn a source of reliable information on the topics which are now or could in the future provide significant new information in the charity's areas of interest.
Conclusions
The BIG concludes that as we enter the electronic era, it is becoming increasingly critical to communicate information in new ways. Linked, interoperable databases are central to the future of knowledge dissemination.
1. Successful implementation of integrated databases of neuroscience information would benefit the entire neuroscience community. Such databases would benefit theoretical and experimental scholars, educators, funders and the general community by establishing effective access to current activities and emerging knowledge profiles of what has been reported and what has been learned. Gaps in the indices of reported observations could illuminate important neglected areas for future research opportunities.
2. The convening power of the Society provides a unique opportunity to lead in the evolution and utilization of integrated neuroscience databases. Through its broad and active membership, public meetings, and dedicated committees, the Society for Neuroscience can lead the development of community-based standards for data ontologies, for review and inclusion standards for the data that would populate neuroscience databases, for the selections of linkages to, and integration of, information across databases, and for the compilation of supplemental data files to augment conventional publications.
3. For neuroscientists to be open about contributing their data to databases and to rely on the data within them to plan their future research, significant barriers will have to be overcome. An important undertaking for the Committee on Neuroinformatics will be to assess the feasibility to develop opportunities and incentives to lower the threshold for data sharing. The Society for Neuroscience Committee on Neuroinformatics could play an important role in meeting the recently mandated objective of the NIH to enforce data sharing.
4. The BIG encourages the Committee on Neuroinformatics to coordinate their efforts with both the Program Committee and the Editors of The Journal of Neuroscience to evolve a mutually acceptable comprehensive ontology of neuroscience index terms, and to cooperate in the inducement of the membership to contribute to the information infrastructure to be developed.
5. Through the Committee on Neuroinformatics, the Society for Neuroscience could envision longer term opportunities. This could include promoting the development of a network of information repositories that span clinical as well as basic neuroscience, facilitate access to information about the myriad of neurological and psychiatric diseases and disorders, expedite the translation of basic research discoveries into improved clinical diagnosis and treatment, thereby accelerating the pace, depth, and quality of neuroscience research of the future, with tangible benefits to society. Related long-term opportunities include the potential for enhanced interactions with neuroscience-focused charitable organizations and neuroscience research funders of the future, as well as the larger issues of the dissemination, integration, and analysis of new and legacy neuroscience information and concepts. Such infrastructural enhancements could serve many purposes, including improved communication with legislators and the public about the relevance and importance of neuroscience research and greater exposure of young scientists to the exciting opportunities in the field of neuroscience.
References:
1. Bowden, D.M. and R.F. Martin, NeuroNames Brain Hierarchy. Neuroimage, 1995. 2 (1): p. 63-83.
2. Bowden, D.M. and M.F. Dubach, NeuroNames 2002. Neuroinformatics, 2003. 1 (1): p. 43-59.
3. Shepherd, G.M., Supporting Databases for Neuroscience Research. J. Neuroscience, 2002. 22 (5): p. 1497.
4. Van Horn, J.D., J.S. Grethe, P. Kostelec, J.B. Woodward, J.A. Aslam, D. Rus, D. Rockmore and M.S. Gazzaniga, The Functional Magnetic Resonance Imaging Data Center (fMRIDC): the challenges and rewards of large-scale databasing of neuroimaging studies. Philos Trans R Soc Lond B Biol Sci, 2001. 356 (1412): p. 1323-39.
5. Van Horn, J.D., S.T. Grafton, D. Rockmore and M.S. Gazzaniga, Sharing Neuroimaging Studies of Human Cognition. Nature Neuroscience, 2004. 7 (5): p. 473-481.
6. Gardner, D., A.W. Toga, G.A. Ascoli, J.T. Beatty, J.F. Brinkley, A.M. Dale, P.T. Fox, E.P. Gardner, J.S. George, N. Goddard, K.M. Harris, E.H. Herskovits, M.L. Hines, G.A. Jacobs, R.E. Jacobs, E.G. Jones, D.N. Kennedy, D.Y. Kimberg, J.C. Mazziotta, P.L. Miller, S. Mori, D.C. Mountain, A.L. Reiss, G.D. Rosen, D.A. Rottenberg, G.M. Shepherd, N.R. Smalheiser, K.P. Smith, T. Strachan, D.C. Van Essen, R.W. Williams and S.T. Wong, Towards effective and rewarding data sharing. Neuroinformatics, 2003. 1 (3): p. 289-95.
7. Insel, T.R., N.D. Volkow, S.C. Landis, T.-K. Li, J.F.B. Jr and P. Sieving, Limits to growth: why neuroscience needs large-scale science. Nature Neuroscience, 2004. 7 (5): p. 426-427.
Appendix
Brain Information Group (BIG) Members
Chair:
Floyd Bloom
Chairman, Department of Neuropharmacology
The Scripps Research Institute
10550 North Torrey Pines Road
La Jolla, CA 92037
858-784-9730
fbloom@scripps.edu
Liaison to SfN Council:
David Van Essen
Washington University School of Medicine
Department of Anatomy & Neurobiology
660 South Euclid Avenue
St Louis, MO 63110
314-362-7043
vanessen@v1.wustl.edu
Liaison to Wadsworth Foundation:
Sarah J. Caddick
Executive Director
Wadsworth Foundation
3601 Fremont Ave North
Seattle, WA 98103
Members:
Huda Akil
University of Michigan
Mental Health Research Institute
Ann Arbor, MI 48109
734-763-3770
akil@umich.edu
Mark S. Boguski
Vulcan, Inc
505 Fifth Avenue South
Seattle, WA 98104
206-342-2581
markbo@vulcan.com
Douglas M. Bowden
University of Washington
National Primate Research Center
Box 357330
Seattle, WA 98195
206-543-2456
dmbowden@u.washington.edu
Daniel Gardner
Cornell Medical College
Department of Physiology
1300 York Avenue
New York, NY 10021
212-746-6373
DAN@APLYSIA.MED.CORNELL.EDU
Gwen A. Jacobs
Montana State University
Center for Computational Biology
Bozeman, MT 59717-3505
406-994-7334
gwen@cns.montana.edu
Edward G. Jones
University of California , Davis
Center for Neuroscience
Davis, CA 95616
530-757-8747
ejones@ucdavis.edu
Luis Marenco
Center for Medical Informatics
Yale University School of Medicine
333 Cedar Street
PO Box 208009
New Haven, CT 06520 -8009
203-737-2985
luis.marenco@yale.edu
Maryann E. Martone
University of California, San Diego
Department of Neuroscience (M008)
Gilman Drive
San Diego, CA 92093-0608
619-534-8332
mmartone@ucsd.edu
Richard J. Mural
Director, Scientific Content and Analysis
Celera Genomics
45 W. Gude Drive
Rockville, MD 20850
240-453-3665
richard.mural@celera.com
Gordon M. Shepherd
Yale University School of Medicine
Department of Neurobiology
333 Cedar Street
New Haven, CT 06510
203-785-4336
gordon.shepherd@yale.edu
John D. Van Horn
Dartmouth College
Center for Cognitive Neuroscience
6162 Moore Hall
Hanover, NH 03755
603-646-2909
john.d.van.horn@dartmouth.edu
Robert W. Williams
University of Tennessee Health Science Center
Department of Anatomy and Neurobiology
875 Monroe Avenue
Memphis, TN 38163
901-448-7018
rwilliam@nb.utmem.edu
|
