Weighted singular value distribution of RRI series applied to the characterization of heat intolerance in humans

A nonlinear scheme was used for the analysis of variability in the heart beat interval [R-R interval (RRI)] data to differentiate heat-intolerant humans from the heat tolerant. All subjects studied had previously suffered exertional heatstroke. Core temperature (T/sub re/) and electrocardiogram data from 11 heat-tolerant (HT) and 6 heat-intolerant (HIT) males were studied, the grouping being based on the distinguishing rate of rise in T/sub re/ versus time up to 39/spl deg/C during submaximal exercise. The RRI data were subjected to wavelet transformation and the transformed data were utilized to generate weighted singular value (WSV) distribution profiles. The normalized WSV profiles merged together for the HT subjects, but remained widely dispersed for the HIT subjects. From WSV profiles of five HT subjects a standard WSV template (w/sub t/) was constructed and with respect to w/sub t/ the cumulative square error (/spl epsiv/) for individual WSV profiles for a cohort of six (additional) HT and six HIT subjects was analyzed. In terms of /spl epsiv/, HT and HIT groups could be differentiated with the sensitivity and the specificity exceeding 83%. The strength of the WSV profiles in characterizing processes is also demonstrated using synthetic data.