arrow

From NeuroLex

Revision as of 05:28, 6 May 2009 by Memartone (Talk | contribs)
Jump to: navigation, search

Contents

Why are we creating the Neurolex?

The Neurolex project is an outgrowth of the Neuroscience Information Framework (NIF) project. The NIF project is developing a practical framework for neuroscientists to find resources that are relevant to their research. Resources include databases, data, literature, software tools, research materials, services and training materials. Because neuroscience is a broad and diverse discipline, providing a unifying framework is a challenging task. Neuroscientists work at many scales, with many types of data and in many organisms. Unlike the genomics field, where everyone can tie data to a sequence, neuroscience has few organizing principles that span the multiplicity of systems and modalities in which we work. The main thing that ties neuroscience together is conceptual; neuroscientists tie the data they produce to terminology describing cells, anatomy, function and disorders of the brain. The NIF thus relies on terminology to provide the unifying framework for organizing and searching neuroscience resources. A common vocabulary is used to tag resources and information contained in the resources to make it easier for the NIF system to return relevant results. The more that neuroscientists use this common vocabulary when reporting their results and making new databases and resources, the easier it is for systems like the NIF to find and interpret them.

Unfortunately, the terminology used by neuroscientists when communicating with each other are ill defined and not usually suited for use within information systems. In order to make them useful, we need to turn them into something that can be used by a machine to perform the types of conceptual leaps used by neuroscientists when navigating through the wealth of data and tools available on the web. To use a simple example, neuroscientists have no difficulty in browsing through resources and recognizing that a database on Parkinson's disease and a database on Parkinsonian disorders are probably related to one another. Similarly, they know that a database with information on Purkinje cells is also related to an article on "Cells of the cerebellum". Information systems can't make these leaps unless we provide some formalization of the relationship between Parkinson disorders and Parkinson's disease; Purkinje cell and "Cells of the cerebellum." NIF has knowledge engineers working on creating these formal vocabularies, called ontologies, but these knowledge engineers need input from the community to ensure that all relevant concepts are represented and that they are defined clearly and reasonably. Because Wikis provide a relatively easy-to-use means for community participation, we created the Neurolex wiki.

How is the NeuroLex different from Wikipedia or any other Wiki-based encyclopedia for neuroscience?

The NeuroLex is interested in defining the meaning of concepts and not in providing general information about the concept. It is a dictionary rather than an encyclopedia. We want to make explicit the way in which a concept is defined so that it can be consistently applied to data annotation. We aren't interested in creating the "right" definition of a term like basal ganglia, but to allow the community to tell us the many ways in which it is defined. Thus, we have created a definition template and some policies that allows the user to provide the required information in a form that's easy for automated agents to understand.

How will the NeuroLex be used?

The NeuroLex will provide a resource for those annotating data or developing methods to search and integrate data based on the meaning of concepts. For the Neuroscience Information Framework project, the NeuroLex will provide the base set of concepts from which we are developing the NIFSTD ontologies. We hope that other groups will also use these base concepts to build their information resources. By using the same core concepts, we can easily connect these different resources together.

Do I need an account to contribute to the NeuroLex?

No, you do not need an account. We want the NeuroLex to provide a very low barrier entry point for you to contribute your knowledge. We view the NeuroLex as the marketplace of ideas for neuroscience concepts. However, if you do get an account, all of your contributions will be linked to you.

Is NeuroLex curated?

Yes, it is. We monitor the NeuroLex regularly. Some of the content of the NeuroLex is already being incorporated into our ontologies by knowledge engineers working on the NIF. We will certainly remove any content that is inappropriate (i. e., doesn't fall within our scope, is obscene or inflammatory). We will also update entries that have been incorporated into the NIFSTD ontologies so that the concept can recieve an appropriate version number. But the NeuroLex is meant to be a community resource and needs to evolve in a way that is useful to the community. So we will adjust our curation policies if necessary as we go.

What is a category page? What is a page?

The NeuroLex is built using something called the Semantic MediaWiki. We chose this platform because it makes it easier to go between the formal ontologies and the Wiki space. It can be a little confusing to get started and to understand its structure. We have created several video tutorials explaining the structure of the NeuroLex and how to edit it. These tutorials are available from the page, How to Contribute to the NeuroLex. We'll be adding more in the future.

How are categories related to one another in NeuroLex?

NeuroLex was designed as a single tree hierarchy, that is, a given category has one and only one parent category (super category) and one or more children categories. A category is related to its parent and children through a "is a type of" relationship, also called an "is a" relationship. So, a Neuron "is a type of" Cell; a "Purkinje cell" is a type of" Neuron. In order to make sure that each Category is related to its parent only through an "is a" relationship, NeuroLex has created some contrived categories, e.g., Regional part of brain; Cytoarchitectural part of hippocampal formation". We are preparing some additional documents to provide the rationale for this design, but for now, they provide useful buckets into which we can place information. The next phase will be more tricky, as we define other relationships between categories that are necessary to build a network of neuroscience concepts. The NeuroLex lets you define other relationships between Categories, e.g., "Is part of"; "Located in"; "Neurotransmitter". For a list of these properties, please see Special pages. When you are defining relationships between categories, you should make sure that you use the correct relationship. The easiest way to do this, is to create a sentence with a subject (Category 1) a predicate (relationship) and an object (Category 2). The Thalamus "is part of" the Diencephalon.

What are some best practices for defining my new term?

The primary purpose of terms in Neurolex is for the annotation of data. Therefore, the definition should provide the meaning of the term in a way that distinguishes it from other related terms (monothetic) so that users will know how to apply that term to their data. We favor the Aristotelian approach: A "is a type of" B "which has" C. B gives the parent class (see above); C gives one or more properties that distinguishes that class from all other related classes.

How do I add a new class?

You can find a quick start guide here.

Who can I contact with more questions?

Mailing list: neurolex at googlegroups.com

Maryann Martone: maryann at ncmir.ucsd.edu

Stephen Larson: slarson at ncmir.ucsd.edu

==Why are there so many different types of brain parts?

There are many different ways to divide up the brain. Although Neurolex takes a structural approach, i.e., we haven't yet defined neural systems like the visual system, we are still left with gross anatomy, cytoarchitecture, chemoarchitecture and connectivity, to name a few. Neurolex is concerned with building formal vocabularies for information systems. We therefore have to make sure that we create our ontologies as cleanly as possible, without mixing up all of these different types of parts. However, we don't expect a naive user to navigate all of these complexities. We therefore set up a few different types of categories into which users can deposit their brain parts. We are still trying to determine whether this is a useful exercise or not, but so far it seems to be working. We have started a page where we document our current categories (http://neurolex.org/wiki/Brain_parts_organization).



bookmark