Invited Reply to: “Instrumental Analysis of Gait Abnormalities in Idiopathic Rapid Eye Movement Sleep Behavior Disorder”

Quantitative gait analysis may be able to predict future phenoconversion in individuals with idiopathic rapid eye movement sleep behavior disorder (iRBD). However, further investigation and replication are needed, especially in large and heterogeneous cohorts. Notable differences between our original study, and the study by Viteckova and colleagues makes it difficult to compare results between the studies and as such does not challenge our original findings. In our original study our sample was small, but very well controlled. Rigorous neuropsychological testing on both groups was performed, ruling out mild cognitive impairments (eg, montreal cognitive assessment (MoCA): healthy controls (HC) = 28, iRBD = 27.4), and successfully correlating cognitive performance to aspects of gait known to be associated. In contrast, although having a larger sample, detailed neuropsychological testing was not reported, and both groups had below-average cognition (MoCA <26; HC = 25.3, iRBD = 23.6). This suggests that neither group was truly “healthy” but, rather, a significant portion of them likely had mild cognitive impairments (MoCA range 19–25 indicates mild cognitive impairments according to www.mocatest.org/faq/). Therefore, latent neurodegeneration may have been present in both groups, which could explain why no differences were found. Importantly, the difference in the samples between the two studies limits our ability to compare findings. Methodological differences also explain the lack of coherence between the findings (ie, a more difficult cognitive dual task [DT] was used previously—subtracting 7s vs. 3s, fast-paced walking is not equivalent to a tray-carrying task, and collection using GAITRite compared to Zeno Walkway allows for the visual cues along the side to reduce the DT interference). These factors aside, we found it interesting that after plotting the means and standard errors of both groups, similar patterns to our previous results emerged (ie, an interaction showing that iRBD increased their stepwidth variabilitywith a cognitive DTwhen comparedwith normal,whereasHCdonot; but theHCwiden their step width during the cognitive DT, although the iRBDdo not). One reason the authors did not find this interaction is because they used t tests rather than amixed analysis of variance. Future consideration should be taken as to whether 1-tailed rather than 2-tailed t tests may be appropriate given that there is a directional hypothesis justified by previous studies. Indeed, if thiswere applied (albeitwithout multiple comparison correction) gait cycle time and step width coefficient of variation (CV) during cognitive DT are greater in iRBD than HC. Differences in normal gait also emerge (ie, iRBD have greater step lengthCV),which alignwith previouswork. Finally, an outstanding problem relating to all studies in iRBD is the need to characterize the heterogeneity in our samples. This would mean standardizing the reporting of markers that can quantify the severity and estimate iRBD patients’ proximity for conversion (eg, report duration of iRBD, smell loss, parkinsonism, color vision, or a combination of these factors as a prodromal severity index). This is essential to interpret the differences in findings that may be the result of one sample being closer to conversion than another. Nonetheless, this remains an exciting field of studywith a large potential for impact.