The hidden structure of neuropsychology: text mining of the journal Cortex: 1991--2001

The stated mission of Cortex is "the study of the inter-relations of the nervous system and behavior, particularly as these are reflected in the effects of brain lesions on cognitive functions." The purpose of this paper is to explore the relationship between the stated mission and the executed mission as reflected by the characteristics of papers published in Cortex. In addition, we examine whether the results and conclusions of an analysis of this kind are affected by the level of description of the published papers.

A) Identify characteristics of contributors to Cortex; B) Identify characteristics of those who cite Cortex; C) Identify recurring themes; D) Identify the relationships among the recurring themes; E) Compare recurring themes and determine their relationships to the mission of Cortex; F) Identify the sensitivity of these results to the level of description of the Cortex papers used as the source database. G) Compare Cortex characteristics with those of Neuropsychologia, another Europe-based international neuropsychology journal.

Text mining (extraction of useful information from text) was used to generate the characteristics of the journal Cortex. Bibliometrics provided the Cortex contributor infrastructure (author/ organization/ country/ citation distributions), and computational linguistics identified the recurring technical themes and their inter-relationships. Citation mining (the integration of citation bibliometrics and text mining) was used to profile the research user community. Four levels of published article description were compared for the analysis: Full Text, Abstract, Title, Keywords.

Highly cited documents were compared among Cortex, Neuropsychologia, and Brain, and a number of interesting parametric trends were observed. The characteristics of the papers that cite Cortex papers were examined, and some interesting insights were generated. Finally, the document clustering taxonomy showed that papers in Cortex can be reasonably divided into four categories (papers in each category in parenthesis): Semantic Memory (151); Handedness (145); Amnesia (119); and Neglect (66). It is concluded that Cortex needs to take steps to attract a more diverse group of contributors outside its continental Western European base if it wishes to capture a greater share of seminal neuropsychology papers. Further investigation of the critical citation differences reported in the paper is recommended.