Heterogeneity of ventricular repolarization is associated with the development of life-threatening ventricular arrhythmias. Temporal heterogeneity of repolarization may be manifest in an individual beat (spatial heterogeneity) or in a sequence of beats (dynamic heterogeneity). Spatial inhomogeneity of repolarization throughout the myocardium may be expressed electrocardiographically as dispersion of repolarization durations computed in simultaneously recorded leads. The beat-to-beat changes in the repolarization pattern (duration and/or amplitude) may account for a dynamic (time-dependent) dimension of heterogeneity, occasionally seen as T-wave alternans. A visual detection of heterogeneous repolarization is a time-consuming, observer-dependent, and frequently inaccurate process. Therefore, we developed computer algorithms designed to detect automatically (1) dispersion of repolarization and (2) nonvisible T-wave alternans from digitally recorded (1,000 Hz) X, Y, and Z electrocardiogram leads. This automatic approach was subsequently tested in 10 patients with idiopathic long QT syndrome and in 10 age-matched normal subjects. Long QT syndrome patients presented with significantly higher indices of heterogeneity in comparison with the control subjects; the dispersion of repolarization was 44 +/- 11 and 13 +/- 6 ms, respectively (P < .01), and T-wave alternans index was 0.40 +/- 0.37 and 0.03 +/- 0.06, respectively (P < .01). Simultaneous evaluation of spatial (dispersion of repolarization) and dynamic (T-wave alternans) aspects of repolarization provides new insight into heterogeneity of electrical recovery after myocardial depolarization. The automatic detection of repolarization dispersion and T-wave alternans in digital electrocardiogram recordings provides a practical method to evaluate heterogeneity of repolarization and may be useful for stratifying patients at risk of ventricular arrhythmias.