We have applied a mathematical function to model sinus node recovery after atrial pacing at different cycle lengths in 13 autonomically denervated, anesthetized canines. This model provides "pure" indexes of sinoatrial conduction and of suppression of automaticity. We compared sinoatrial conduction computed from the model (SACTc) with sinoatrial conduction measured directly utilizing a catheter recording the sinus node electrogram (SACTd) to test the validity of the model. We used both the model and the sinus node electrogram to further define events after pacing. There was close agreement between SACTc and SACTd with correlation coefficients at pacing cycle lengths of 20 and 50% below resting of 0.86 (P < 0.0001) and 0.78 (P < 0.0001), respectively. We verified that the conventional sinus node recovery time is due both to suppression of automaticity and prolongation of sinoatrial conduction. Sinoatrial conduction time returned rapidly to normal over 1-3 beats, whereas automaticity recovered exponentially over 20-40 postpacing intervals. The model accurately describes postpacing phenomena and has the potential to provide a useful adjunct to the measurement of the sinus node electrogram for assessment of sinus node function in the experimental laboratory.