NeuroML is an open XML vocabulary for describing data driven neuron and nervous system simulations. I’ve used NeuroML and highly recommend its use for providing a standard way to exchange models of brain cells and brain systems.
A good place to start is the paper “NeuroML: A Language for Describing Data Driven Models of Neurons and Networks with a High Degree of Biological Detail” published in the June 2010 issue of PLOS Computational Biology.
On July 27, 2010 the Proceedings of the National Academy of Sciences of the USA (PNAS) published the paper entitled “Network architecture of the long-distance pathways in the macaque brain” by Dharmendra Modha (IBM Research – USA) and Raghavendra Singh (IBM Research – India).
The study used data from a half century of anatomical tracing data in nonhuman primates held in the Collation of Connectivity data on the Macaque brain (CoCoMac) online data repository.
By analyzing CoCoMac‘s full set of data, the authors’ were able to collate a connectivity network that included 383 brain regions with 6,602 connections. These are the relatively long distance connections that send signals from one part of the brain to another in a similar way that a telephone line sends a signal from one geographical location to another.
Each brain area was treated as a point. That means that each connection was treated as point-to-point and subtle anatomical complexities such as the particular three dimensional location within an area or kind of nerve cell targeted by a connection was ignored.
The level of abstraction used in this study is entirely understandable. It enabled the team to move forward with their new approach but ignoring anatomical details clearly has implications for the limits placed on using their results to understand signal processing carried out on the signals that arrive in each of the connected areas. Nevertheless, the future will no doubt bring methodologies to include increased target resolution within each area so that, for instance, the targeted cortical layer or even the targeted cell type may be taken into account.
This study is clearly important within the context of brain research using online data repositories. Using the strengths of a particular online research data repository – CoCoMac in this case – the authors were able to summarize a large volume of research carried out over the past 50 years, add potentially useful knowledge to the neuroscience domain, and unequivocally add useful knowledge and techniques to the neuroinformatics domain.
Neuroinformatics gains in at least two areas from the Modha and Singh paper. First, the resulting wiring diagram provides an excellent framework for summarizing brain connectivity and can provide a visual map for working with brain derived data. The CoCoMac data repository, for instance, might use a dynamically created diagram like the one in the paper’s Figure 1 as a portal into the sites data.
Second, the analysis described in “Network architecture of the long-distance pathways in the macaque brain” may be carried out dynamically online in real time if CoCoMac were to transform their data repository to be Semantic Web compatible.
In particular, using the Web Ontology Language (OWL) the logical operations described in the paper’s supplementary material could be carried out with standard Reasoners. It strikes me that the approach taken in this paper for understanding the data in CoCoMac would be an excellent match for creating a powerful application of Semantic Web technologies within the neurosciences.
Today in the United State alone taxpayers pay about $30 billion a year on biomedical research. Traditionally, once scientists completed and published a study the primary data that the publications were based on languished, accumulated dust, and were eventually lost.
It’s been clear for some time that everyone will get a better return on our health care and life sciences research budget if the data generated are available to everyone in a usable form across the Internet. This is the driving force behind a number of bioinformatics initiatives around the world and in the World Wide Web Consortium‘s effort to get linked biomedical data online and integrated into the Semantic Web.
This relatively new effort has resulted in a plethora of previously unobtainable research data being available to everyone through online repositories. At this stage the impact of the available data is anybodies guess although I think the impact will be huge.
Papers based on data from neuroscience research data repositories are beginning to appear. I will review some of these papers so that we may begin assessing the impact of placing neuroscience research data into publicly available online repositories.