Dear Editor:

We thank Dr. Kawada for his comment on our work [1] as he raises an important point in relation to the use of actigraphy in sleep assessment, namely, that the optimal Actiwatch setting for maximizing sleep/wake accuracy will be subject-dependent. As noted in our paper, we used the “default” settings for Actiwatch sensitivity provided by the manufacturer, which resulted in an overall accuracy that was comparable to the other instruments in the study, but which had lower wake specificity. It would have been helpful to have also included the results when the “high sensitivity” setting was used, which has the tendency to increase wake specificity at the expense of sleep sensitivity. In a related publication [2] where we reported on the results of Actiwatch and biomotion sensor in a sleep apnea population, the Actiwatch sleep sensitivity and wake specificity were 0.94 and 0.34 at the default setting, but changed to 0.90 and 0.45 at the high sensitivity setting. These findings are consistent with Kushida’s results at “high sensitivity” setting which gave 0.92 and 0.48 for the corresponding measures [3].

In short, we agree with Dr. Kawada’s commentary that researchers need to be aware of the impact of the actigraphy algorithm sensitivity settings on the results, and that in subjects with a high pre-test probability of wakefulness, one should consider the use of the “high sensitivity” settings in addition to the default setting. We also agree that further studies in clinically interesting populations such as those with sleep disordered breathing, periodic limb movements and primary insomnia should be carried out to validate the use of non-PSG sleep measurement devices.