IN THEIR COMMENTARY,¹ DRS. FERRI AND ZUCCONI ARGUE THAT THE OBSERVATIONS ON AUTONOMIC AND EEG CHANGES ACCOMPANYING PERIODIC (PLMS) and isolated leg movements in sleep (ILMS) that we recently published in SLEEP² are the “opposite” of their recent report.³ However, we believe that a direct comparison of this kind is problematic, because the two studies derived their results from different patient populations and with different analysis techniques. Whereas we looked at ILMS of patients with normal sleep studies, i.e., without restless legs syndrome (RLS),² Ferri et al. analyzed ILMS of patients with RLS.³ Differences between the 2 studies could therefore point to potentially important autonomic and EEG changes accompanying not only PLMS of patients with RLS but even their ILMS. This notion seems to be in agreement with the conclusions of Ferri et al.,³ rather than opposing it. Moreover, whereas Ferri et al. analyzed the mean HRV only, we derived our pathophysiological hypotheses mostly from results of spectral HRV analyses. Extrapolating spectral HRV changes from the reported mean HRV in their study is similarly impossible as extrapolating spectral EEG changes from an evoked potential.

Ferri and Zucconi further expressed concerns that the >10 seconds inter-jerk time interval used for analyzing leg movements in our study might have been too short, a theoretical possibility which we had explored previously. In Figure 1 accompanying this response, we show that that the main finding of our study—the more prominent spectral HRV changes during PLMS than during other movements in sleep²—persists even if we analyze only PLMS with a inter-jerk interval of >30 seconds, which disproves the possibility, that these effects were due to a baseline bias. Furthermore, we have reported a stable baseline for EEG and mean HRV results, and similar findings as previous studies,^4,5 which further shows that the pre-movement interval had no major influence on the results. On the other hand, we are worried that the results reported in ³ cannot be used as a reference for other studies, because they appear to be contaminated by EMG and movement artifacts. In our experience, a careful exclusion of trials contaminated by artifacts is crucial for studies assessing movement related EEG changes.^2,6 Conversely, Ferri et al. state that they did not use any corresponding selection criteria,³ and the large error bars as well as the irregular average time courses presented in their figures are typical for EMG and movement artifacts. It might therefore be difficult to separate EEG changes from EMG changes in their study.

Unlike claimed by Ferri and Zucconi,¹ our study did control for differences in sleep stages, EMG activity, and arousals between the analyzed movement types. Only movements occurring during NREM stage 2 were included for comparisons between movement types. Figures 3 and 4 in our publication show that the combined EMG activity of both anterior tibial muscles was not significantly different between the movement types. Differences in arousals between the movement types were assessed with a spectral EEG analysis, as done by Ferri et al.³

We may thus conclude that none of the concerns raised by Ferri and Zucconi¹ constrains the validity of our results. Moreover, the proposed hypothesis that the sympathetic nervous system plays an important role in the pathophysiology of PLMS^2,7 has recently received additional support.^8,9