Background: Few studies have dealt with the behavior of the corrected (QTc) and uncorrected QT intervals during exercise and recovery. Hypothesis: Based on previously published dynamics of the QT interval during treadmill testing, this study attempted to reevaluate the computer‐proposed underlying mechanisms of these dynamics and to determine whether the so‐called memory phenomenon could be operative in some subjects without evidence of structural heart disease. Methods: This study included 42 unmedicated healthcare volunteers, 23 men and 19 women aged between 20 and 42 (mean 31.7) years. All had normal physical examinations, x‐rays, and transthoracic echocardiograms. The electrocardiograms were also normal with 12‐lead QT interval dispersions of <90ms. Results: During exercise and recovery, the behavior of the QT intervals permitted the categorization into two groups. In Group 1 (31/42; 73.8% of subjects) the uncorrected QT interval showed a biphasic pattern consisting of a gradual decrease during incremental exercise followed by a gradual increase during recovery. In contrast, the QTc interval had a triphasic pattern resulting from a slight increase during the early phase of exercise, a gradual decrease at the highest rates, and a final increase during recovery as the rate slowed to control values. In Group 2 (11/42; 36.2% of subjects) the behavior was considered as paradoxical since the uncorrected QT interval displayed in a triphasic pattern whereas the QTc interval yielded a tetraphasic pattern due to the fact that both showed a second decrease during recovery while the rate was decreasing. Conclusions: Analysis of dynamics behavior of the QTc and the uncorrected QT intervals during exercise showed that some normal subjects had a paradoxical behavior which, because of its temporal relation to the phases of exercise, could be an expression of the so‐called memory phenomenon.