To overcome the "parcellation problem" of
incompatible brain maps we have developed a mathematical approach named
"Objective Relational Transformation" (ORT). Combining methods from formal
logic, graph-theory and theoretical computer science, the main components
of ORT are:
new classifications for brain data and for relations between parcellation
schemes,
operationalised codings for the description of published brain data,
an algebra of coordinate-independent transformation rules,
graphtheoretical algorithms, formal languages and finite automata for optimisation
of data conversion.
ORT offers
objectivity by representing data in their original parcellation
schemes,
flexibility by offering conversion of data into freely chosen
brain maps,
reliability by using formal
algorithms in the mapping process,
transparency by fully documenting the criteria and opinions
on which the mapping process is based.
universality by making use of already published coordinate-independent
data.
We have implemented ORT within the relational
CoCoMac
databases. Introductions to ORT have been published, including a complete
description of the methodology (Stephan & Kötter 1998, 1999; Stephan
et al. 2000a). |