Raffaele Ferri, MD, Sleep Research Centre, Department of Neurology I.C., Oasi Institute for Research on Mental Retardation and Brain Aging (IRCCS), Via C. Ruggero 73, 94018 Troina, Italy. Tel.: +30-0935-936111; fax: +39-0935-936694; e-mail: firstname.lastname@example.org
The aim of this study was to evaluate quantitatively the amplitude of the submentalis muscle EMG activity during sleep in controls and in patients with idiopathic REM sleep behavior disorder (RBD) or with RBD and multiple system atrophy (MSA). We recruited 21 patients with idiopathic RBD, 10 with MSA, 10 age-matched and 24 young normal controls. The average amplitude of the rectified submentalis muscle EMG signal was used for the assessment of atonia and a Sleep Atonia Index was developed; moreover, also chin muscle activations were detected and their duration and interval analyzed. The Sleep Atonia Index was able to distinguish clearly REM from NREM sleep in normal controls with values very close to 1 in young normal subjects and only slightly (but significantly) lower in old controls. Idiopathic RBD patients showed a further significant decrease of this index; MSA patients showed the lowest values of REM Sleep Atonia Index, which were very well distinguishable from those of normal controls and of idiopathic RBD patients. The distribution of the duration of chin activations was monomodal in all groups, with idiopathic RBD patients showing the highest levels. This study is a really quantitative attempt to provide practical indices for the objective evaluation of EMG atonia during REM sleep and of EMG activations. Our proposed Sleep Atonia Index can have a practical application in the clinical evaluations of patients and represents an additional useful parameters to be used in conjunction with the other criteria for the diagnosis of this sleep motor disorder.
The current sleep scoring rules require the recording of surface electromyography (EMG) from the submentalis muscle in order to detect the presence of skeletal muscle atonia during rapid eye movement (REM) sleep (Rechtschaffen and Kales, 1968). The presence of REM sleep without atonia with the EMG finding of excessive amounts of sustained or intermittent elevations of submentalis muscle EMG tone or excessive phasic submentalis muscle twitching (or upper/lower limbs) is the main polygraphic feature of the REM sleep behavior disorder (RBD) (American Academy of Sleep Medicine 2005), a condition characterized by abnormal behaviors emerging during REM sleep that can appear in an idiopathic form or linked to several pathologic conditions, causing injury or sometimes sleep disruption (Schenck and Mahowald, 2002; Schenck et al., 1986). RBD is frequently associated with some neurodegenerative diseases such as Parkinson’s disease, multiple-system atrophy, dementia with Lewy bodies, and in most of the cases it precedes the major symptoms of these diseases of many years. RBD has been postulated to be a possible early marker and an heraldic symptom for synucleinopathies (Boeve and Saper, 2006), even though it has been recently described in progressive supranuclear palsy (Arnulf et al., 2005), Machado-Joseph disease (Iranzo et al., 2003), spinocerebellar ataxia type 2 (Boesch et al., 2006), and in probable Alzheimer disease (Gagnon et al., 2006); for this reason, RBD might be considered to be related more to the localization affecting the tonic and/or phasic REM sleep motor circuitry rather than to a specific type of neuronal degeneration. Moreover, idiopathic RBD (iRBD) does exist with symptoms and neurophysiologic features similar to the secondary forms (Iranzo et al., 2005) and the term ‘cryptogenic’ instead of idiopathic has been proposed (Fantini et al., 2005).
Despite the crucial relevance of the increased chin EMG activity during REM sleep for the diagnosis of RBD, the reliability of polysomnographic criteria for REM without atonia is largely unexplored (Walters et al., 2007) and only few systematic attempts have been carried out in order to analyze quantitatively submentalis muscle EMG activity during sleep (Brunner et al., 1990), probably because of the use, for decades, of paper recordings in sleep research. Basically, only three papers have appeared in the literature from 1992 to 2006, which have attempted to quantify submentalis muscle EMG activity in RBD patients by means of a visual approach, with the aim of counting the amount of epochs without atonia (elevated background EMG activity) and the amount of mini-epochs (2- to 3-s long) containing phasic EMG activity (Bliwise et al., 2006; Consens et al., 2005; Lapierre and Montplaisir, 1992). In these studies, beside their simple and apparently quantitative parameters, no solid mathematical and quantitative definitions, in terms of amplitude and duration, have been provided for the elements taken into account: atonia, phasic and tonic EMG activations.
Surface EMG is an uncalibrated signal, which is believed to show large inter- and intra-individual variability, even during the same recording; for this reason, it is very difficult to consider its absolute amplitude as a parameter for quantitative analysis. Nonetheless, the definition of the apparently quantitative parameters introduced by the papers cited above appears to be weak and difficult to implement within a computer-based analysis of a digital EMG recording, which is now the standard method for storing biologic data.
For these reasons, the aim of this study was that of evaluating, from a strictly quantitative point of view, the amplitude of the submentalis muscle EMG activity during sleep in normal controls, in patients with iRBD, and in patients with RBD and multiple system atrophy (MSA), a condition almost invariably associated to RBD (Plazzi et al., 1997; Vetrugno et al., 2004). For this purpose we analyzed the results statistically, in order to extract information able to help the arrangement of a quantitative definition of parameters, such as atonia and EMG activations of different durations during sleep.
Subjects and methods
A total of 31 patients with RBD were recruited for this study [21 affected by iRBD and 10 fulfilling the criteria for clinically probable MSA (Gilman et al., 1999)] together with 10 age-matched normal controls and 24 young normal controls. Age and gender composition of the groups are shown in Table 1; only the iRBD group is clearly male-predominant, with the other three groups having almost equal gender composition. This was because of the consecutive character of the recruitment, which did not allow us to arrange balanced groups from the gender point of view. The diagnosis of iRBD was based on the International Classification of Sleep Disorders (ICSD-2) criteria (American Academy of Sleep Medicine 2005) for RBD, including presence of REM sleep without atonia, sleep related injurious-disruptive behaviors by history or abnormal sleep behaviors documented during polysomnographic monitoring, absence of EEG epileptiform activity during REM sleep, and sleep disturbance not better explained by another sleep disorder, medical or neurologic disorder, mental disorder, medication use, or substance use disorder. Secondary forms of RBD were excluded on the basis of historical data, neurologic examination, and encephalic MRI findings. All RBD patients with at least one subtentorial vascular lesion or at least two vascular sopratentorial lesions greater than 0.5 cm were excluded. All iRBD patients were recorded during the hospitalization during which their condition was first diagnosed; this occurred on average within 3.1 years (S.D. 2.81, range 0.5–12 years) from the beginning of the clinical symptomatology and none of them had been treated with medications specific for this disease. Moreover, MSA patients were recorded within 2 years from the beginning of the clinical symptomatology and all presented a clinical complaint of RBD at the time of the recording. All patients were drug-free for at least 3 weeks at the time of the polysomnographic recording.
Table 1. Subjects included in this study
Age, years ± SD (range)
iRBD, idiopathic REM sleep behavior disorder; MSA, multiple system atrophy.
29.9 ± 3.54 (25–39)
65.0 ± 4.97 (60–76)
70.1 ± 4.55 (61–80)
63.6 ± 8.48 (46–73)
None of the 34 normal subjects had any physical, neurologic or psychiatric disorder or history of sleep problems and none was taking medication at the time of recording. Exclusion criteria included (a) a sleep disorder diagnosis (including sleep apnea); (b) a major mental illness; (c) a significant history of cognitive difficulties (d) prior (within one year) or current use of a neuroleptic agent or SSRIs, venlafaxine, (e) history of alcohol or other substance abuse.
Nocturnal polysomnography was carried out after a night of adaptation in a standard sound-attenuated (noise level to a maximum of 30 dB nHL) sleep laboratory room. Subjects were not allowed caffeinated beverages the afternoon preceding the recording and were allowed to sleep in until their spontaneous awakening in the morning. Light-out time was based on individual habitual bed time and ranged between 21.30 and 23.30 h. The following signals were recorded: EEG (at least two channels, one central and one occipital, referred to the contralateral earlobe); electrooculogram (electrodes placed 1 cm above the right outer cantus and 1 cm below the left outer cantus and referred to A1), electromyogram (EMG) of the submentalis muscle (bipolar derivations with two electrodes placed 3 cm apart and affixed using a collodion-soaked gauze pad), impedance was kept less than 10 KΩ (typically <5 KΩ), EMG of the right and left tibialis anterior muscles, and ECG (one derivation). Sleep signals were sampled at 200 or 256 Hz and stored on hard disk in European data format (EDF) (Kemp et al., 1992), for further analysis. The sleep respiratory pattern of each patient was monitored using oral and nasal airflow thermistors and/or nasal pressure cannula, thoracic and abdominal respiratory effort strain gauge and by monitoring oxygen saturation (pulse-oxymetry). This was performed in all subjects in a previous recording (within 1 week) or during the study recording; patients with an apnea/hypopnea index ≥5 were not included. Sleep stages were scored following standard criteria (Rechtschaffen and Kales, 1968) on 30-s epochs; epochs containing technical artifacts or extremely elevated muscle activity causing saturation of amplifiers were carefully detected and marked for exclusion from the subsequent quantitative EMG analysis.
Quantification of the submentalis muscle EMG amplitude
The submentalis muscle EMG signal was digitally band-pass filtered at 10–100 Hz, with a notch filter at 50 Hz and rectified; subsequently, each sleep epoch included in the analysis was subdivided into 30 1-s mini-epochs. The average amplitude of the rectified submentalis muscle EMG signal was then obtained for each mini-epoch. Figure 1 reports an example of the rectification and quantification of the submentalis muscle EMG signal amplitude in an REM sleep epoch of a patient with iRBD, subdivided into 30 mini-epochs. This type of representation of the submentalis muscle EMG signal amplitude allowed us to establish, heuristically, some thresholds for further analysis. In fact, we noticed that in correspondence of periods of submentalis muscle EMG atonia, the mini-epoch average amplitude was usually ≤1 μV (even in the presence of contamination from ECG) while, when phasic or tonic muscle activations were detectable in the unrectified signals, the average mini-epoch amplitude was usually >2 μV.
The values of the submentalis muscle EMG signal amplitude in each mini-epoch were used to draw normalized distribution histograms, for each sleep stage (REM, S1, S2, and SWS) of the percentage of values in the following 20 amplitude (amp) classes (expressed in μV): amp ≤ 1, 1 < amp ≤ 2, …, 18 < amp ≤ 19, amp > 19. On the basis of the considerations made above, in these graphs, muscle atonia is expected to be reflected by high values of the first left column while phasic and tonic activations are expected to increase the value of the other columns. For this reason, we also arranged an index in order to summarize in a single value the degree of preponderance of the first column in these graphs by normalizing its value with respect to the cumulative value of all columns from 2 < amp≤ 3 to amp > 19 (the column 1 < amp≤ 2 was not taken into account because it might reflect both atonia and EMG activation), thus:
Mathematically, this index can vary from 0 (absence of mini-epochs with amp ≤ 1) which means complete absence of EMG atonia, to 1 (all mini-epochs with amp ≤ 1) or stable EMG atonia in the epoch, and was calculated for each sleep stage.
We also counted, for REM sleep, all sequences of consecutive mini-epochs exceeding the value of 2 μV and calculated their number per hour; also these data were used to draw normalized distribution histograms, within the following 20 duration (dur) classes (expressed in s): dur = 1, dur = 2, …, dur = 19, dur > 19. These data served for the analysis of differences between the different subject groups.
Also the intervals between the EMG events detected as described above were measured and the data were used to draw normalized distribution histograms, within the following 49 interval (int) classes (expressed in s): int = 2–4, int = 4–6, …, int = 98–100.
Finally, in order to make a correlation between the results obtained with our approach with those expected from the previously proposed visual analysis, we quantified only in REM sleep of all groups also the two parameters proposed by Lapierre and Montplaisir (1992), adapted to our recordings. First, each 30-s epoch was scored as tonic or atonic depending on whether tonic chin EMG activity was present for more or less than 50% of the epoch; then, we calculated the proportion of the total REM epochs scored as atonic in order to obtain a number varying between 0 and 1. Second, we evaluated phasic EMG density as the number of 2-s mini-epochs per hour of REM sleep containing phasic EMG events (defined as any burst of EMG activity lasting 0.1–5 s, with an amplitude exceeding four times the background EMG activity).
Statistical data analysis
All data were coded and analyzed blind to the subject group upon completion of recruitment. Comparisons between groups were carried out using the non-parametric Kruskal–Wallis anova, followed by the Mann–Whitney test for independent data sets, for post hoc comparisons; for between stages comparisons in the same group, the non-parametric Friedman anova was used. The correlation between the results obtained with our approach with those obtained by visual analysis was evaluated by means of the nonparametric Spearman rank correlation coefficient. Differences and correlations were considered as significant when they reached a P < 0.05 level. The data analysis software system STATISTICA (StatSoft, Inc. 2004, version 6. http://www.statsoft.com) was used for statistical analysis.
Submentalis EMG analysis in young normal controls
Figure 2 shows the normalized distribution histograms of mini-epoch amplitude, for each sleep stage (REM, S1, S2, and SWS), in young normal controls. Clear differences can be seen between REM and non-REM (NREM) sleep; the first is characterized by a very high and prominent first column indicating that about 80% of mini-epochs during this sleep stage have amp ≤ 1. The amplitude of the same column is much lower in NREM sleep with evident differences between stage 1 and the other NREM sleep stages; in particular, during sleep stage 1 the first column shows the smallest value with a shift of the graph towards the other columns representing higher values of amp.
Figure 3 shows the values of the Sleep Atonia Index in the different sleep stages of young normal controls; a clear and significant difference is evident between the different stages with the maximum value for REM sleep (0.933) and the minimum for NREM sleep stage 1 (0.333).
Figure 4 contains the normalized distribution histograms of consecutive mini-epoch sequences exceeding 2 μV, for each sleep stage (REM, S1, S2, and SWS). In these graphs, a similar distribution is evident for all sleep stages, with a tendency for sleep stage 2 and slow-wave sleep to show higher values mostly for columns in the left part of the graphs (shorter activations).
Figure 5 reports the normalized distribution histograms of intervals (onset-to-onset) between consecutive mini-epoch sequences exceeding 2 μV, for each sleep stage (REM, S1, S2, and SWS). Beside the obvious difference in amplitude of the graphs because of the different number of EMG activations contained in each sleep stage (see above), these distributions are all monomodal with a maximum for the shortest intervals (2–6 s) and values rapidly decreasing for longer intervals. This type of distribution is very similar to that expected for 1/f or 1/f2 (Brownian) noise.
Comparison between young normal controls and old controls
Figure 6 shows the comparison between chin EMG parameters found during REM sleep in young and old normal controls. The two graphs on the top report the comparison between the normalized distribution histograms of mini-epoch amplitude; old controls show a significantly smaller first column (amp ≤ 1 μV) while all the other columns show higher values than young controls. The middle graphs depict the comparison between the normalized distribution histograms of consecutive mini-epoch sequences exceeding 2 μV; a clear and significant difference is evident because the first three columns of old controls, representing phasic muscle activations, show markedly higher values than those of young controls; also the last column on the right, representing long tonic activations is significantly higher in old controls. Finally, the two graphs on the bottom report the normalized distribution histograms of intervals between consecutive mini-epoch sequences exceeding 2 μV; most of the columns in the left part of these graphs differ significantly being much higher in the old control group; on the contrary, the shape of the two distribution is clearly similar and monomodal, with a maximum for the shortest intervals (2–6 s) and values rapidly decreasing for longer intervals.
Comparison between old normal controls and iRBD or MSA patients
Figure 7 reports the comparison between chin EMG parameters found during REM sleep in old normal controls and in iRBD or MSA patients. The three graphs on the top report the comparison between the normalized distribution histograms of mini-epoch amplitude; iRBD patients show a significantly smaller first column (amp ≤ 1 μV) while all the other columns show higher values than old controls with most of them reaching statistical significance. Similarly, MSA patients show a significantly smaller first column (amp ≤ 1 μV) while practically all the columns with amp ≥ 3 μV show significantly higher values than old controls. Important differences are also evident between the graphs of iRBD and MSA patients.
The middle graphs show the comparison between the normalized distribution histograms of consecutive mini-epoch sequences exceeding 2 μV; all columns obtained from iRBD patients show higher values than those pertaining to old controls, with several of them reaching statistical significance; on the contrary, the same graph obtained in MSA patients is very similar to that of old controls with very few significant differences.
Finally, the three graphs on the bottom report the normalized distribution histograms of intervals between consecutive mini-epoch sequences exceeding 2 μV; almost all columns in the left part of the graph of iRBD patients are significantly higher than those of normal controls; on the contrary, no significant differences were found between MSA patients and normal controls for this type of graphs. However, also in this case, these three graphs show a similar monomodal distribution of intervals with a maximum for the shortest intervals (2–6 s) and values rapidly decreasing for longer intervals.
REM Sleep Atonia Index in young and old normal controls and in iRBD or MSA patients
Figure 8 shows, in details, the REM Sleep Atonia Index in young and normal controls and in patients with iRBD or MSA. In this figure, it is possible to notice that normal controls show less variable values of this parameter than iRBD patients; however, an important overlap exists between iRBD patient and old controls. In any case, two-thirds of iRBD patients show a REM Sleep Atonia Index lower that 0.7 while this happens only in one young and two old controls. All MSA patients show a REM Sleep Atonia Index lower than 0.5, with a relatively small variability, and an average value significantly lower than that of all the other groups of subjects.
Comparison between quantitative statistical and visual analysis
Figure 9 shows the correlation between the results obtained with our approach with those expected from the previously proposed visual analysis, quantified only in REM sleep of the four groups of subjects. The top graphs show the correlation between our REM Sleep Atonia Index and proportion of the total REM epochs scored as atonic, following the visual analysis method proposed by Lapierre and Montplaisir (1992). The non-parametrics Spearman correlation coefficient is high and statistically significant in all cases. Similarly, the bottom graphs of this figure show the correlation between the number of EMG activations (lasting ≤5 s)/h of REM sleep, detected with our approach and the number of 2-s mini-epochs per hour of REM sleep containing phasic EMG events, as defined by Lapierre and Montplaisir (1992). Also in this case the correlation coefficients are high and statistically significant.
To our knowledge, there are only few papers in the literature which tried to evaluate in a true quantitative way the submentalis muscle EMG background activity (among other muscles) and with different purposes. The mean activity level of mentalis muscle has been reported to be significantly decreased during REM sleep in comparison with NREM sleep and such EMG activity was always higher than zero (Bliwise et al., 1974); this was confirmed many years later by means of a computerized quantitative analysis (Brunner et al., 1990; Kato et al., 2004). A recent study (Okura et al., 2006) has shown that computer quantitative analysis of the background suprahyoid muscle EMG activity can be performed reliably in normal controls and is able to show its significant progressive reduction from wakefulness to sleep and a further decrease from NREM to REM sleep.
Thus, even if in the past it was believed that the evaluation of the background muscle tone recorded by surface electrodes is difficult because they pick up a heterogeneous mixture of multiple motor units firing at varying periods (Basmajian, 1967), new computerized quantitative analysis is probably able to provide statistically reliable results. A very preliminary study by means of an automatic analysis of the chin EMG has been described in a short abstract (Mayer et al., 2006) reporting that very short phasic activity in iRBD patients is embedded in long-lasting activity (1–12 s) and that the automatic quantification of the chin EMG activity by criteria of duration and amplitude might allow to establish thresholds for clinical and subclinical forms of RBD.
Differently from the studies listed above, we quantified practically the whole sleep period (with the exclusion of very few artifact epochs), at least in normal controls, and used mini-epochs long enough to allow a reliable estimation of the background EMG mean amplitude and short enough to pick up short-lasting phasic activities.
After a preliminary visual inspection of the mean amplitude of these mini-epochs, we approached the definition of atonia without pre-established rules and tried to let the data to guide the subsequent analysis. In this process, we could observe that, in young controls, approximately 80% of REM sleep mini-epochs had a mean amplitude of no more than 1 μV; this was very different from NREM sleep and was used as the nominator of our Sleep Atonia Index, a single measure able to distinguish clearly REM from NREM sleep in normal controls. We also observed that the REM Sleep Atonia Index had values very close to 1 in young normal subjects and only slightly (but significantly) lower in old controls; iRBD patients showed a further significant decrease of the mean value of this index and an arbitrary threshold of 0.7 might be used as a supporting criterion for the diagnosis of iRBD (if confirmed in future studies) because normal controls only rarely show values of REM Sleep Atonia Index lower than this threshold. Moreover, the use of this index in MSA patients showed that they are very well distinguishable from normal controls and from iRBD patients too; in fact, MSA patients showed the lowest values of REM Sleep Atonia Index in this study, corresponding to the highest levels or REM sleep without atonia.
These results, showing that approximately one-third of iRBD patients have values not distinguishable from the range expected from normal controls, confirm the need to consider also the presence of an abnormal level of chin activations during REM sleep for the diagnosis. We can hypothesize, in our approach, that it is possible to draw normal ranges for the graph representing submentalis muscle EMG activations of different durations for REM sleep and subsequently comparing the patient data with such normative graphs. It is important to underline that with this approach we do not need to establish ‘a priori‘ definitions for phasic and tonic activations because we do take into account almost all of them. Moreover, our results do not support the concept of the existence of these categories because, in such a case, we should have observed a bimodal distribution of these graphs. This was not the case and the small peak visible in the last category is only because of the fact that it represents the cumulative count of all activations lasting >19 s.
In this respect, it is interesting to note that our iRBD subjects showed normalized values of EMG activations higher than those of old controls for a wide range of durations (between 2 and 14 s) but not for duration 1 s, probably representing very short phasic activations; these activations are also present in old controls with values significantly higher than young controls. This might represent an important factor to take into account for the future delineation of this disorder, which typically occurs at about 50–60 years of age (Schenck and Mahowald, 2002), when also normal controls are expected to have significant levels of very short phasic submental muscle EMG activity, as seen in our data.
Interestingly, MSA patients only showed few differences for this type of graph from normal controls; this finding, together with the evident differences shown by the REM Sleep Atonia Index between iRBD and MSA patients allows to affirm that this new quantitative approach is not only able to show differences between normal controls and patients, but also between nosologically different groups of patients who might present a different type of REM sleep-related motor disturbance which is today called RBD in all cases. These differences are potentially able to indicate different neurochemical and neurophysiologic mechanisms for a clinically apparently similar sleep disturbance occurring in different conditions such as iRBD and MSA (Vetrugno et al., 2004). RBD has been reported to occur in a variety of neurologic conditions, such as narcolepsy (Nightingale et al., 2005), Parkinson disease (Boeve et al., 2004; Onofrj et al., 2002), spinocerebellar ataxia (Friedman, 2002), supranuclear palsy (Arnulf et al., 2005), Levy bodies disease (Boeve et al., 2004), Alzheimer disease (Gagnon et al., 2006), amyotrophic lateral sclerosis (Thomas et al., 2007), multiple sclerosis (Plazzi and Montagna, 2002), Tourette syndrome (Trajanovic et al., 2004) and, finally, RBD can be induced by medications, especially the tricyclic antidepressants and serotonin-specific reuptake inhibitors (Mahowald et al., 2007; Thomas et al., 2007); we can speculate that our new approach might be able to provide new insights on the specific patterns of chin EMG activation during REM sleep also in these disorders.
Finally, another important finding of our study is the lack of any apparent time structure (periodicity) of the intervals separating chin EMG activations in normal controls and patients, as already introduced in the description of the results, in all cases the interval distribution histograms were monomodal with a maximum for the shortest intervals (2–6 s) and values rapidly decreasing for longer intervals. This type of distribution is very similar to that expected for 1/f or 1/f2 (Brownian) noise. In order to understand the meaning of this finding, it is important to take into account the different spectral content of different types of noise. For white noise the power is unrelated to the frequency, so we say this noise has a 1/f0 distribution. Thus, White Gaussian noise has a spectrum of the same power content at all frequencies (flat spectrum). As, f0 is a constant, this is just another way of saying the power is independent of the frequency. On the other hand, Brownian noise exhibits a 1/f2 power spectrum; a third type of scaling noise is called 1/f noise, and its power spectrum has a 1/f shape. Both these two last types of noise show a periodogram characterized by a maximum for the shortest intervals and values rapidly decreasing for longer intervals. We did not analyze in more details this point because it was not crucial for our purposes; we think that it was only important to establish the type of time structure of the chin EMG activations during sleep and we were able to exclude the presence of clear periodicity, in contrast with other sleep motor disorders such as periodic leg movements, which do have different degrees of periodicity in different conditions (Ferri et al., 2006a,b).
In this study, we did not evaluate other muscles than the submentalis muscle but we can suggest that an approach similar to that described in this paper might be able to evaluate activations in other muscular groups and to disclose significant differences between patients and normal controls. Among the other muscles, anterior tibialis muscles are often recorded for the analysis of periodic limb movements; we believe that the analysis proposed in the present study might be performed also on these muscles, provided that a careful detection of the eventual periodic movements is carried out in advance. As stated before, the consecutive character of the recruitment did not allow us to arrange balanced groups from the point of view of gender; in this study, we made the assumption that gender plays little or no role in explaining the findings but future research should take gender into account in order to evaluate an eventual effect of this factor.
Before concluding, it is important to underline that our quantitative approach to the analysis of the chin EMG activity during sleep is very sensitive to the presence of noise or artifacts; for this reason, it requires great attention in recording the EMG signal correctly and in the eventual careful rejection of artifacts. We cannot exclude that some contribution to the number of EMG activations detected with our method might be due to ‘random’ fluctuations of the EMG amplitude, because we used an absolute amplitude value as a threshold for their detection (2 μV). However, we have demonstrated that there is a high and significant correlation between the results obtained with our method and those obtained by using a visual approach proposed in the past (Lapierre and Montplaisir, 1992); for this reason, we are reasonably certain that these eventual spurious detections do not affect our results in a significant way.
In conclusion, we believe that this approach can be considered to be a really quantitative attempt to provide useful and practical indices for the quantitative and non-subjective evaluation of EMG atonia during REM sleep and of EMG activations. The combination of atonia and activation indices promises to be useful in the scientific study of different clinical conditions characterized by abnormal chin EMG activation during sleep and our proposed REM Sleep Atonia Index can have a practical application in the clinical evaluations of patients and represents an additional useful parameters to be used in conjunction with the other criteria for the diagnosis of this sleep motor disorder.