Non-technical summary Cardiac arrhythmias occur when the heart is unable to beat rhythmically, and can be induced if the heart is forced to work in excess of its capacity. Here we examine how cells isolated from the atrial chamber of the human heart respond when their beating rate is increased. Our results show that there are differences in the ability of individual heart cells to respond rhythmically when they are forced to beat fast. Moreover, we find that mechanisms that regulate the level of calcium ions in the heart cells determine their ability to respond rhythmically. These results show that pharmacological control of calcium levels in heart cells might be useful in the treatment of people who suffer cardiac arrhythmias during excessive exercise.Abstract Irregularities in intracellular calcium on a beat-to-beat basis can precede cardiac arrhythmia, but the mechanisms inducing such irregularities remain elusive. This study tested the hypothesis that sarcoplasmic reticulum (SR) and L-type calcium channel activity determine the beat-to-beat response and its rate dependency. For this purpose, patch-clamp technique and confocal calcium imaging was used to record L-type calcium current (I Ca ) and visualize calcium in human atrial myocytes subjected to increasing stimulation frequencies (from 0.2 to 2 Hz). The beat-to-beat response was heterogeneous among a population of 133 myocytes, with 30 myocytes responding uniformly at all frequencies, while alternating and irregular responses were induced in 78 and 25 myocytes, respectively. Myocytes with uniform responses had the lowest frequency of calcium wave-induced transient inward currents (I TI ; 0.4 ± 0.2 min −1 ), I Ca density (1.8 ± 0.3 pA pF −1 ) and caffeine-releasable calcium load (6.2 ± 0.5 amol pF −1 ), while those with alternating responses had the highest I TI frequency (1.8 ± 0.3 min −1 , P = 0.003) and I Ca density (2.4 ± 0.2 pA pF −1 , P = 0.04). In contrast, the calcium load was highest in myocytes with irregular responses (8.5 ± 0.7 amol pF −1 , P = 0.01). Accordingly, partial I Ca inhibition reduced the incidence (from 78 to 44%, P < 0.05) and increased the threshold frequency for beat-to-beat alternation (from 1.3 ± 0.2 to 1.9 ± 0.2 Hz, P < 0.05). Partial inhibition of SR calcium release reduced the I TI frequency, increased calcium loading and favoured induction of irregular responses, while complete inhibition abolished beat-to-beat alternation at all frequencies. In conclusion, the beat-to-beat response was heterogeneous among human atrial myocytes subjected to increasing stimulation frequencies, and the nature and stability of the response were determined by the SR and L-type calcium channel activities, suggesting that these mechanisms are key to controlling cardiac beat-to-beat stability.