This article compares different methods of scatterplot analysis to assess the optimal methodology. The scatterplot (Poincare plot) is a nonlinear heart rate variability method where a "return map" is constructed by plotting each current cycle against the previous beat (RR vs. RR(n-1)). Geometric analysis of the scatterplot allows short-term and long-term heart rate variability (HRV) to be assessed. A three-dimensional construct is also possible, where the third axis represents the density of values, at any given RR vs. RR(n-1) intersection. Topological methods of analysis can compute the density distribution function or compactness of a dataset. Scatterplots that otherwise appear very similar in the two-dimensional plot may be clearly differentiated using this approach. Correct characterization may improve the ability of scatterplot analysis to predict outcomes in cardiovascular disease. We have assessed two computational approaches that take account of scatterplot density, namely, the heart rate variability fraction and the compactness measure. Scatterplots were constructed from three double-blind and randomized placebo controlled studies conducted in a total of 49 healthy subjects. Single oral doses of antagonists (atenolol 50 mg [beta-1], propranolol 160 mg [beta-1 and beta-2], and ICI 118,551 25 mg [beta-2]) or agonists (xamoterol 200 mg [beta-1], salbutamol 8 mg [beta-2], prenalterol 50 mg [beta-1 and beta-2], and pindolol 10 mg [mainly beta-2] of the cardiac beta-adrenoceptor were studied. Salbutamol, pindolol, and xamoterol increased compactness and reduced HRV fraction significantly compared with placebo. However, when compared with the more conventional scatterplot parameters, these newer density methods were found to be less discriminating. An alternative approach to improve scatterplot discrimination, using the combination of several scatterplot features, is under investigation.