Estimating ventilation/perfusion distributions from inert gas data: a Bayesian approach

In recent years, experimental techniques using inert gas studies have been developed for determining essentially continuous distributions of the ventilation/perfusion ratio over a lung. Mathematically, the technique requires the recovery of 'smooth' solutions of grossly underdetermined systems of linear equations. This paper investigates the use of Bayesian smoothing techniques and shows that, provided the true underlying distribution of ventilation/perfusion ratios is relatively smooth and well-behaved, it is possible to recover reliable estimates of such distributions.