Review of:
J Klaessens, J Thijssen, J Hopman, K Liem,
"Experimental verification of conditions for near infrared spectroscopy (NIRS),"
Technology and Health Care 11(1): 53-60, 2003.

This paper is of potential interest to computing scientists: simulation enthusiasts, computing-bent bioengineers and bioinformatics proponents observing blossoming phenotypic expression study (not explicitly a paper topic). It offers insight into in vitro-in vivo correspondence, issues as old as the earliest artificial models: what does a given in vitro configuration really reveal about life; are its results repeatable; and are there scientific responses (e.g. simulations) that firm up weak model-reality links?

The paper, specifically concerned about how a (in vitro) "tissue-equivalent Phantom" reflects on (potentially irreversibly damaging) ischemia in neonatal brain, is mainly about experiments: phantom adequacies and inadequacies; and angular placements of input probes and output recorders (to maximize (minimize) responses) inherent to (application of) continuous wave near infrared spectroscopy ((CW-)NIRS in seeking real life counterparts in a phantom (inserted "rods" of variant absorption and scattering coefficients). The following quotation on correspondences among simulations, the phantom, experimental conditions, and real life may help: Simulation at an optode separation of 90(degrees) angle would be optimal for detecting an inhomogeneity at 15 mm depth, i.e. the location of the periventricular white matter" (corresponding to sought in vivo anisotropies). An identified "spatial sensitivity path enclosing the rod position at maximum absorption found experimentally" provides a basis to postulate optimal in vivo placements.

The paper is well written and working through it helps see how computation may join theory and experiment in modern scientific inquiry. Should the study's main results (and predictions) hold up under further research, computing folks can enjoy the computing contributions.