There is no one specific tool that constitutes solving for pathology dynamics. We utilize a combination of developed and traditional tools to accomplish this task. These techniques have goals which include examining relationships, interactions, feedback, and overall system stability. Additionally, a wide variety of tools are used to gather, extract, organize and quantify the data before the analysis even begins.
Developed Tools
- Dynamic meta-analysis (DMA): DMA is similar to traditional meta-analysis except that is explicitly includes time and implicitly includes interactions. In short, the algebraic meta-regression equation is converted into a set of time-varying differential equation(s). For more information, see Mitchell and Lee, 2012, Intech).
- Relational modeling: form of computational modeling that solely utilizes relationships and correlations to aggregate and simulate detailed, unknown or complicated mechanisms at a system-level. See Mitchell and Lee, 2009, J Neurotrauma.
- Relational analysis: an aggregation of correlative and complex systems techniques that uses multiple inter-relationships of system components to make predictions of system-level behavior. In particular, “landscapes” of component relationships are compared over varying conditions or at different time points. See Mitchell and Lee, 2007, N Neural Engineering for process details and Mitchell and Lee, 2009, J Neurotrauma for pathology example.
- Conceptual modeling: form of computational modeling that uses concepts and theories to build model mechanisms. For example of a conceptual model, see Shapiro and Lee, 2007, J Neurophysiology.
- Viewpoint aggregation: an analytical technique that combines different views of a pathological system, such as conceptual insights, experimental observations, refined/detailed mechanisms, and clinical treatment outcomes to produce a comprehensive, system-level view. For details see Mitchell, 2009, Georgia Institute of Technology.
Traditional Tools
- Multi-variate statistics: use of cross-correlation analysis, cluster analysis, hierarchial analysis, factor analysis, etc. to identify key trends and quanitfy system dimensionality.
- Mathematical and engineering dynamics: used to determine stability of the system, identify instability type, determine system gains and feedback control.
- Mechanistic modeling: used to examine mechanistic dynamics at the system component level.
- Databases and ontologies: used to organize published data, organize extracted and quantified data points as well as to provide hiearchial structure for browsing, querying, and searching.
- Bioinformatics and literature mining: used to comb through thousands of published articles to find the quantifiable data needed to examine the pathological system and its dynamics. We are always interested in trying new tools to help us collect and quanitfy data more efficiently.