A Large Citation Network

The figure below is a snapshot of the field of neuropsychology circa 1975 (when I acquired the data) viewed as a matrix of communications among researchers, "communication" here defined as the citing of one worker by another in his published papers. The image represents a matrix of 1052 frequently-cited researchers in that field (columns) and the 16432 authors who cited them (rows) between 1970 and 1975.

A cell in the matrix contains a one if the row author cited the column author during the five-year period, zero otherwise. For display purposes I squashed the matrix vertically so that each pixel condenses information on two column authors and about 32 row authors. A dark pixel means that at least one citation occurred among the 2x32 authors; a white pixel means that no citations occurred.

(Text continued below images.)

	GREENGARD CARLSSON | | <------------------------------ 1052 cited authors------------------------>
^ | | | | | | | KANDEL______ | (2)---> 16432 citing authors | | | | | | | | | | | | | | | | | | | | (1)---> | | CARLSSON____ | GREENGARD~~~ | | | | | | | | v
	^ ^ | | | | | | <------ -----------> Zoom of center of major cluster

I used the so-called "bond-energy" clustering algorithm running on a supercomputer to rearrange the rows and columns until a maximum degree of adjacency between non-zero cells was achieved. A dark area on the diagonal represents intercommunication amoung a large number of authors, and therefore may be taken to be a "hot" area of research.

A couple of examples (marked 1 and 2 in the large image): Cluster (1) obviously dominates the matrix. It is concerned with neuropharmacology and brain metabolism. Cluster (2) comprises studies of brain architecture, electrophysiology and neural development.

OCTOBER, 2000 -- I have marked the columns and rows occupied by the three Nobel Prize winners in Physiology, Arvid Carlsson, Paul Greengard and Eric Kandel.

• Column names