Home

People

NeuroScholar

Proteus

Science Overview

Sangam Template

Demo Interface

Plan Execution

Quarterly Reports

 

 

 

 

 

Overview

Anxiety (or stress) disorders are the most common mental illness in America, affecting about 19.1 million American adults.  Some types of these disorders (Table 1) can be associated with other illnesses such as eating disorders, depression, or drug dependency. Many scientists devote considerable energy trying to understand the causal mechanisms underlying this important clinical problem, as is apparent from communities such as the Endocrine Society, which currently numbers 11,000 members in 80 countries (www.endo-society.org). Such scientists have obtained much data suggesting that anxiety disorders are caused by dysfunction within specific brain circuits, but the precise relationships between these circuits and the way in which they are recruited by stress signals is unclear.  Understanding this is critical for treating stress disorders.

Table 1. Anxiety Disorders, One-Year Prevalence (Adults)1

Illness

Percent

Population

Estimate

(Millions)*

Any Anxiety Disorder

13.3

19.1

Panic Disorder

1.7

2.4

Obsessive-Compulsive Disorder

2.3

3.3

Post-Traumatic Stress Disorder

3.6

5.2

Any Phobia

8.0

11.5

Generalized Anxiety Disorder

2.8

4.0

*Based on 7/1/98 U.S. Census resident population estimate of 143.3 million, ages 18-54. Source: NIMH.

 

Fig 1 shows how stress triggers a  sequence of events within the brain.  First, highly tuned stress sensors transmit signals to distinct brain circuits, which transfer this information to various brain regions to activate selected genes within those regions. The signals are also sent to CRH neurons, a distinct population of nerve cells that respond to stress by releasing hormones and by turning on gene sequences to make new hormones. The relative simplicity of this figure belies the breadth and depth of data in this field. For example, the specific problem we tackle in this application is: Does every type of stress stimulus recruit the same set of brain circuits and activate the same genes, or do such circuits and genes vary across different stressors? An answer to this question helps clinicians and drug manufacturers to develop better treatments and drugs for stress disorders.  This question can be answered by integrating information from  each   stage of this sequence to assess     how stress machinery is  engaged for a particular stressor.

 

             We propose Sangam (a Tamil word for “a symposium, a meeting or a council of scholars”; also an Urdu/Hindi word denoting “river delta”) as an end-to-end system to address this challenge.  Sangam provides a what-oriented interface to facilitate the seamless, rapid integration of different data sources using the context provided by local databases of scientific data concerning brain responses to stress, as outlined in Fig 1. It provides a graphical interface for use by a neuroscientist working in this field.  Sangam demonstrates the utility of Web Services, WSE 2.0, Microsoft .NET, Proteus, and NeuroScholar for scientific client applications.

 

The broader impact of this activity is multi-fold:

1.   It demonstrates the feasibility of automatic integration of diverse data sources for a specific discipline.

 

2.   It produces the fundamental computer science research results that enable realistic prototypes.

 

3.   It provides a toolset that will be immediately useful to biomedical scientists within other disciplines.

 

4.    Perhaps most importantly, it could provide a catalyst for developers throughout biomedical informatics to build interoperability into their systems using Web Services.

 

                Interoperability has always been emphasized by funding agencies as very desirable and yet, very few online systems can run queries remotely (see the Society for Neuroscience's ‘database gateway’, http://big.sfn.org/NDG/site/ for a list of 79 examples). Since Sangam offers the possibility of providing seamless interoperability between Web Service enabled systems, and the technical overhead of implementing Web Services is small, our work could conceivably set a new standard for building interoperability into biomedical informatics systems.  We set an example by writing mediators that convert those data sets required by our eScience application into WSE compliant Web Services.