Background Given the high prevalence and risk for outcomes associated with pediatric obstructive sleep apnea (OSA), there is a need for simplified diagnostic approaches. A prospective study in 140 children undergoing in-laboratory polysomnography (PSG) evaluates the accuracy of a recently developed system (Sunrise) to estimate respiratory efforts by monitoring sleep mandibular movements (MM) for the diagnosis of OSA (Sunrise (TM)). Methods Diagnostic and severity were defined by an obstructive apnea/hypopnea index (OAHI) >= 1 (mild), >= 5 (moderate), and >= 10 events/h (severe). Agreement between PSG and Sunrise (TM) was assessed by Bland-Altman method comparing respiratory disturbances hourly index (RDI) (obstructive apneas, hypopneas, and respiratory effort-related arousals) during PSG (PSG_RDI), and Sunrise RDI (Sr_RDI). Performance of Sr_RDI was determined via ROC curves evaluating the device sensitivity and specificity at PSG_OAHI >= 1, 5, and 15 events/h. Results A median difference of 1.57 events/h, 95% confidence interval: -2.49 to 8.11 was found between Sr_RDI and PSG_RDI. Areas under the ROC curves of Sr_RDI were 0.75 (interquartile range [IQR]: 0.72-0.78), 0.90 (IQR: 0.86-0.92) and 0.95 (IQR: 0.90-0.99) for detecting children with PSG_OAHI >= 1, PSG_OAHI >= 5, or PSG_ OAHI >= 10, respectively. Conclusion MM automated analysis shows significant promise to diagnose moderate-to-severe pediatric OSA.

Purpose:Differentiation between obstructive and central apneas and hypopneas requires quantitative measurement of respiratory effort (RE) using esophageal pressure (PES), which is rarely implemented. This study investigated whether the sleep mandibular movements (MM) signal recorded with a tri-axial gyroscopic chin sensor (Sunrise, Namur, Belgium) is a reliable surrogate of PES in patients with suspected obstructive sleep apnea (OSA).Patients and Methods:In-laboratory polysomnography (PSG) with PES and concurrent MM monitoring was performed. PSGs were scored manually using AASM 2012 rules. Data blocks (n=8042) were randomly sampled during normal breathing (NB), obstructive or central apnea/hypopnea (OA/OH/CA/CH), respiratory effort-related arousal (RERA), and mixed apnea (MxA). Analyses were evaluation of the similarity and linear correlation between PES and MM using the longest common subsequence (LCSS) algorithm and Pearson's coefficient; description of signal amplitudes; estimation of the marginal effect for crossing from NB to a respiratory disturbance for a given change in MM signal using a mixed linear-regression.Results:Participants (n=38) had mild to severe OSA (median AH index 28.9/h; median arousal index 23.2/h). MM showed a high level of synchronization with concurrent PES signals. Distribution of MM amplitude differed significantly between event types: median (95% confidence interval) values of 0.60 (0.16-2.43) for CA, 0.83 (0.23-4.71) for CH, 1.93 (0.46-12.43) for MxA, 3.23 (0.72-18.09) for OH, and 6.42 (0.88-26.81) for OA. Mixed regression indicated that crossing from NB to central events would decrease MM signal amplitude by -1.23 (CH) and -2.04 (CA) units, while obstructive events would increase MM amplitude by +3.27 (OH) and +6.79 (OA) units (all p<10-6).Conclusion:In OSA patients, MM signals facilitated the measurement of specific levels of RE associated with obstructive, central or mixed apneas and/or hypopneas. A high degree of similarity was observed with the PES gold-standard signal.

1. Institute of Medicine. Crisis standards of care: a systems framework for catastrophic disaster response. Vol. 1: Introduction and CSC framework. Washington, DC: The National Academies Press; 2012 [accessed 2021 Oct 19]. Available from: https://doi.org/10.17226/13351. 2. Piscitello GM, Kapania EM, Miller WD, Rojas JC, Siegler M, Parker WF. Variation in ventilator allocation guidelines by US state during the coronavirus disease 2019 pandemic: a systematic review. JAMA Netw Open 2020;3:e2012606. 3. Rosenbaum L. Facing Covid-19 in Italy—ethics, logistics, and therapeutics on the epidemic’s front line. N Engl J Med 2020;382: 1873–1875. 4. New York State Department of Health and New York State Task Force on Life and the Law Update Ventilator Allocation Guidelines; 2015 [accessed 2021 Oct 19]. Available from: https://health.ny.gov/regulations/task_force/ reports_publications/docs/ventilator_guidelines.pdf 5. Daugherty Biddison EL, Faden R, Gwon HS, Mareiniss DP, Regenberg AC, Schoch-Spana M, et al. Too many patients...a framework to guide statewide allocation of scarce mechanical ventilation during disasters. Chest 2019;155:848–854. 6. Wunsch H, Hill AD, Bosch N, Adhikari NKJ, Rubenfeld G, Walkey A, et al. Comparison of 2 triage scoring guidelines for allocation of mechanical ventilators. JAMA Netw Open 2020;3:e2029250. 7. Gershengorn HB, Holt GE, Rezk A, Delgado S, Shah N, Arora A, et al. Assessment of disparities associated with a crisis standards of care resource allocation algorithm for patients in 2 US hospitals during the COVID-19 Pandemic. JAMA Netw Open 2021;4: e214149. 8. Snow GL, Bledsoe JR, Butler A, Wilson EL, Rea S, Majercik S, et al. Comparative evaluation of the clinical laboratory-based Intermountain risk score with the Charlson and Elixhauser comorbidity indices for mortality prediction. PLoS One 2020;15:e0233495. 9. White DB, Lo B. Mitigating inequities and saving lives with ICU triage during the COVID-19 pandemic. Am J Respir Crit Care Med 2021;203: 287–295. 10. Daniels N. Just health: meeting health needs fairly. New York: Cambridge University Press; 2007. 11. Persad G, Joffe S. Allocating scarce life-saving resources: the proper role of age. J Med Ethics [online ahead of print] 22 Mar 2021; DOI: 10. 1136/medethics-2020-106792. 12. Antiel RM, Curlin FA, Persad G, White DB, Zhang C, Glickman A, et al. Should pediatric patients be prioritized when rationing life-saving treatments during COVID-19 pandemic. Pediatrics 2020;146: e2020012542.

Purpose: Sleep bruxism (SBx) activity is classically identified by capturing masseter and/or temporalis masticatory muscles electromyographic activity (EMG-MMA) during in laboratory polysomnography (PSG). We aimed to identify stereotypical mandibular jaw movements (MJM) in patients with SBx and to develop rhythmic masticatory muscles activities (RMMA) automatic detection using an artificial intelligence (AI) based approach. Patients and Methods: This was a prospective, observational study of 67 suspected obstructive sleep apnea (OSA) patients in whom PSG with masseter EMG was performed with simultaneous MJM recordings. The system used to collect MJM consisted of a small hardware device attached on the chin that communicates to a cloud-based infrastructure. An extreme gradient boosting (XGB) multiclass classifier was trained on 79,650 10-second epochs of MJM data from the 39 subjects with a history of SBx targeting 3 labels: RMMA episodes (n=1072), micro-arousals (n=1311), and MJM occurring at the breathing frequency (n=77,267). Results: Validated on unseen data from 28 patients, the model showed a very good epoch-by-epoch agreement (Kappa = 0.799) and balanced accuracy of 86.6% was found for the MJM events when using RMMA standards. The RMMA episodes were detected with a sensitivity of 84.3%. Class-wise receiver operating characteristic (ROC) curve analysis confirmed the well-balanced performance of the classifier for RMMA (ROC area under the curve: 0.98, 95% confidence interval [CI] 0.97-0.99). There was good agreement between the MJM analytic model and manual EMG signal scoring of RMMA (median bias -0.80 events/h, 95% CI -9.77 to 2.85). Conclusion: SBx can be reliably identified, quantified, and characterized with MJM when subjected to automated analysis supported by AI technology.

Le syndrome des apnées du sommeil (SAS) en pédiatrie est fréquent et co_morbide ; un diagnostic simplifié est nécessaire en pratique clinique. Nous évaluons ici la performance d’une machine learning basée sur l’analyse de l’effort respiratoire à travers le signal du seul mouvement mandibulaire (MM) pour le diagnostic du SAS pédiatrique (Sunrise, Namur, Belgium). L’étude prospective a été réalisée chez 140 enfants adressés pour une polysommographie au laboratoire de sommeil (PSG) avec enregistrement simultané des MMs. Le diagnostic et la sévérité sont définis en PSG par un index d’apnée/hypopnée (OAHI) ≥ 1/h (léger), ≥ 5/h (modéré) et ≥ 10/h (sévère). La concordance entre la PSG et Sunrise (Sr) est appréciée par un graphe de Bland-Altman comparant l’indice horaire des évènements obstructifs à type d’apnées/hypopnées et de micro-éveils liés à un effort respiratoire (ORDI). La performance de Sr-ORDI est déterminée par une courbe ROC évaluant la sensibilité et la spécificité aux 3 niveaux de sévérité de la PSG. Une différence médiane de 1,57 évènements/h (95 %CI : -2,49 ; 8,11) est mesurée entre Sr_ORDI et PSG_ORDI. Les aires (et leurs écarts interquartiles) sous les courbes de ROC sont respectivement de 0,75 (0,72 ; 0,78), 0,90 (0,86 ; 0,92) et 0,95 (0,90 ; 0,99) aux seuils de sévérité en PSG de ≥ 1/h, ≥ 5/h et ≥ 10/h. L’analyse automatisée de MM procure un diagnostic robuste du SAS pédiatrique modéré à sévère.

Les mouvements de la mandibule associés à l’effort respiratoire sont soumis à des variations conditionnées par les changements de stade de sommeil. Nous présentons les résultats d’une analyse de machine learning basée sur le seul signal mandibulaire pour le scoring des stades de sommeil. Le signal est acquis via un dispositif connecté Sunrise (Sunrise, Namur, Belgium) au cours d’une polysomnographie au laboratoire (PSG). Le modèle décrit 4 états : la veille (V), le sommeil lent léger (SLL), lent profond (SLP) et le sommeil paradoxal (SP). Il est entrainé sur 707 417 segments de 30 sec. collectés chez 800 sujets suspects de SAS correspondant à 1134 heures de V, 2806 h. de SLL, 1052 h. de SLP et 902 h. de SP. Il est validé sur les données ultérieures chez 226 autres sujets. Le modèle montre un niveau élevé de concordance segment par segment avec la PSG (Kappa = 0,71 (95 %CI : 0,59 to 0,80). Il fournit un niveau de concordance acceptable avec les variables de qualité de sommeil de la PSG : la médiane [95 %CI] de la mesure des biais est respectivement de −10,25 min. [−52,87 ; +19,00], de 0 min. [−23,37 ; +6,50], de −2,22 % [−11,06 ; +3,80] et de +7,20 % [−2,41 ; +21,26] pour le temps en SP, pour le temps total de sommeil (TST), pour le délai d’endormissement, pour l’efficience du sommeil et pour SP en % de TST. L’analyse par intelligence artificielle du signal mandibulaire se prête bien à la stadification automatique du sommeil. Elle offre une solution ambulatoire peu onéreuse pour apprécier sa qualité.

Abstract Introduction Sleep bruxism (BXM) is the result of rhythmic muscular masticatory activity (RMMA) and can be captured by masseters surface electromyography (sEMG). Despite the multiple adverse negative consequences of BXM, a simple reliable home diagnostic device is currently unavailable, with in laboratory audio-video polysomnography (type I PSG) remaining the gold standard diagnostic tool. Mandibular movements (MM) recordings during sleep can readily identify RMMA, are simple to set up and can be easily repeated from night to night. Here, we aimed to identify stereotypical MM in patients with BXM, and to develop RMMA automatic detection and BXM diagnosis using an artificial intelligence-based approach. Methods MM were recorded by a dedicated sensor (Sunrise, Namur, Belgium) in 12 patients with BXM during type I PSG. The Sunrise system consists of a coin-sized hardware that is comfortably placed on the subject’s chin. Its embedded inertial measurement unit communicates via Bluetooth with a smartphone and automatically transfers MM signals to a cloud-based infrastructure at the end of the night. Data processing and analysis are then performed in Python programming language. A time series cluster analysis was applied to sequences of masseters sEMG and MM signals during BXM episodes (n=300) and during spontaneous micro-arousals (n=300). Then, a convolutional neuronal network (CNN) was developed to identify BXM and distinguish it from spontaneous micro-arousals while exclusively relying on MM signal. Results Based on the cluster analysis, BXM periods were characterized by a specific pattern of MM signals (higher frequency and amplitude), which was closely associated with the sEMG signals but clearly differed from the MM signal patterns during micro-arousals. CNN-based classifier distinguished the BXM events from other RMMAs during micro-arousals and respiratory efforts with an overall accuracy of 91%. Conclusion Sleep bruxism can be automatically identified, quantified, and characterized with mandibular movements analysis supported by artificial intelligence technology. Support This work was supported by the French National Research Agency (ANR-12-TECS-0010), in the framework of the “Investissements d’avenir” program (ANR-15-IDEX-02). https://life.univ-grenoble-alpes.fr.

This diagnostic study compares the performance of mandibular movement monitoring during sleep coupled with an automated analysis by machine learning vs polysomnography for the diagnosis of obstructive sleep apnea in adults.

Context: Automated sleep stages scoring from surrogate signals is challenging. Mandibular movements (MM) are reliable markers of the neural respiratory drive, with activity patterns reflecting sleep stages. We introduce a new automated sleep stage scoring method using MM signal. Method: Study was conducted on 96 adults (18 to 58 yrs, AHI<10 events/h) who underwent in-lab polysomnography coupled with MM recording by a wearable device (Sunrise, Namur, BE). MM signals were split into 30 seconds epochs, from which 217 features were extracted to classify 3 labels: wake, nonREM and REM sleep. The model was trained on PSG data from 68 subjects consisting of 169 hrs of nonREM, 26 hrs of REM and 21 hrs of wake. The model was validated on unseen data from 28 subjects composed of 147 hrs of nonREM, 38 hrs of REM sleep and 33 hrs of wake. Results: Subjects had a mean total sleep time of 7.14 hrs (CI: 5.53 to 8.49), a mean AHI of 3.57 events/h (CI: 1.00 to 6.92) and a mean arousal index of 9.22 events/h (CI: 4.36 to 16.69). The model showed good global performances, with a balanced accuracy of 0.79 (CI: 0.78 to 0.80) and an area under the receiver operating curve of 0.94, 0.91 and 0.93 for wake, nonREM and REM. Conclusion: This study demonstrated for the first time that MM signals are suitable for automated sleep stage scoring, providing a promising solution for home monitoring of sleep architecture.

Rationale: Mandibular position and motion during sleep rely on the balance between mandibular elevators and depressors. We hypothesized that vertical mandibular position (VMP) modulates airflow amplitude during sleep. Methods: VMP, tidal nasal flow pressure (NFP) and concurrent surface electromyographic activity of the masseters (sEMG-m) were recorded and processed by a customized algorithm from 100 polysomnographic fragments including a micro-arousal (25 obstructive sleep apnea patients). The relationship between mandibular position and changes in airflow was analysed. Result: Concurrent VMP and sEMG-m activity changes routinely occurred before a new steady state of airflow documented by NFP. Vertical mandible depression was associated with a median (95% CI) reduction in NFP of 40.9% (14.6%-71.3%, p = 0.007) while vertical mandible elevation and mouth closure were associated with a median (95% CI) relative increase in NFP after arousal of 52.6% (17.9%-56.2%, p = 0.001). Conclusion: Elevation and lowering of the mandible were associated with changes in masseteric EMG activity modulating airflow amplitude during sleep.

Le bruxisme pendant le sommeil (BXM) est produit par une activité rythmique musculaire massétérine (RMMA) qui peut être capturée par l’EMG de surface (sEMG). Cet évènement moteur est difficile à bien caractériser en polysomnographie et une simplification pourrait être apportée par l’analyse des mouvements de la mandibule (MM) comme signal caractéristique contemporain du BXM. L’objectif était d’identifier les profils de MM par comparaison à sEMG au cours des RMMAs et de développer un algorithme de détection automatique de RMMA par intelligence artificielle (IA). Les MMs (Sunrise®) ont été enregistrés chez 12 patients bruxeurs. Une analyse typologique a été appliquée au cours des accès de BXM (n = 300) et des micro-éveils (n = 300). Le BXM est caractérisé par un motif spécifique de mouvements mandibulaires à haute fréquence et haute amplitude concordant avec le sEMG mais distincts dans leurs phénotypes d’autres MM survenant à l’occasion de micro-éveils. Une approche par IA de type réseau neuronal convolutif appliqué aux seuls MM est capable de détecter les RMMAs et de les différencier des micro-éveils d’autres origines avec une précision de 91 %. Le bruxisme peut être identifié pendant le sommeil et caractérisé de manière automatique par l’analyse des mouvements mandibulaires exercée par un algorithme d’intelligence artificielle.

Objectif La position de la mandibule et ses mouvements verticaux pendant le sommeil (VMM) sont le resultat d’une interaction specifique de ses muscles abaisseurs ou elevateurs. Nous avons etudie l’effet causal de VMM sur les changements de pression nasale captures par les lunettes (FPres) a l’occasion d’un micro-eveil et d’un changement d’activite EMG de surface des muscles antagonistes de la mandibule (sEMG). Methodes Les signaux bruts (VMM, FPres et sEMG) ont ete extraits a partir de 100 fragments PSG de duree moyenne de 2 minutes puis exposes a un algorithme de segmentation pour localiser les evenements d’interet. L’effet causal de VMM sur FPres est verifie par les 4 criteres de Bradford Hill : temporalite, force de l’association coherence et une relation dose–effet. Resultats Un effet causal significatif de VMM sur FPres a ete observe dans 80/100 fragments. Le deplacement de mandibule s’est toujours developpe avant la modification de flux. L’ouverture buccale est associee a une perte de flux (− 24,1 μv (− 17,6 a − 30,6) ; p Conclusion Le flux respiratoire pendant le sommeil est conditionne par la transposition mandibulaire, qui a son tour est determinee par l’activite EMG des muscles antagonistes. Controler la dimension verticale de ce deplacement apparait critique pour le succes therapeutique de l’orthese d’avancee mandibulaire.

Background: Sleep mandibular movements (MM) patterns inform about sleep/wake phases, micro-arousals and respiratory efforts; providing a robust bio-signal for the detection of obstructive apnea/hypopneas and respiratory efforts-related arousals (RERAs). Objective: To validate a new MM recording sensor and machine learning algorithm (Sunrise, Belgium) in computing total sleep time (TST), arousal index (ArI) and respiratory disturbances index (RDI). The usefulness of the indices derived by the algorithm will be validated against reference polysomnography. Methods: 400 subjects with suspected OSA underwent in-lab polysomnography (PSG) supplemented with concomitant MM recording with the sensor. Scoring was conducted independently for the two methods. PSG were scored by experienced investigators. PSG_TST, PSG_ArI and PSG_RDI were derived to guide OSA diagnosis using ICSD-3 rules (5, 15 events/h). MM data were processed in a cloud-based platform without human intervention. Machine learning identification of sleep/wake, micro-arousals and respiratory efforts allowed the computation of Sunrise Sr_TST, Sr_ArI and Sr_RDI. Bland-Altman agreement analysis and ROC curves analysis were conducted to compare the two methods. Results: Mean difference between the two methods was -2.6 min (95%CI: ± 35.2) and -1.7 e/h (95%CI: ± 7.3) for TST and ArI, respectively. OSA detection reached a sensitivity - specificity of 0.90 - 0.94 at RDI cutoffs of 5 e/h and 0.92 - 0.85 at 15 e/h. Conclusion: Automatic machine learning diagnosis using MM patterns showed reliable performances in detecting sleep/wake, arousals and RDI, demonstrating a promising value as a standalone OSA diagnosis tool.

Context: Accurate discrimination between obstructive and central hypopneas requires quantitative assessments of respiratory effort by esophageal pressure (OeP) measurements, which preclude widespread implementation in sleep medicine practice. Mandibular Movement (MM) signals are closely associated with diaphragmatic effort during sleep. Objective: We aimed at reliably detecting obstructive off central hypopneas events using MM statistical characteristics. Methods: A bio-signal learning approach was implemented whereby raw MM fragments corresponding to normal breathing (NPB; n = 501), central (n = 263), and obstructive hypopneas (n = 1861) were collected from 28 consecutive patients (mean age = 54 years, mean AHI = 34.7 n/h) undergoing in-lab polysomnography (PSG) coupled with a MM magnetometer, and OeP recordings. Twenty three input features were extracted from raw data fragments to explore distinctive changes in MM signals. A Random Forest model was built upon those input features to classify the central and obstructive hypopnea events. External validation and interpretive analysis were performed to evaluate the model's performance and the contribution of each feature to the model's output. Results: Obstructive hypopneas were characterized by a longer duration (21.9 vs. 17.8 s, p < 10-6), more extreme low values (p < 10-6), a more negative trend reflecting mouth opening amplitude, wider variation, and the asymmetrical distribution of MM amplitude. External validation showed a reliable performance of the MM features-based classification rule (Kappa coefficient = 0.879 and a balanced accuracy of 0.872). The interpretive analysis revealed that event duration, lower percentiles, central tendency, and the trend of MM amplitude were the most important determinants of events. Conclusions: MM signals can be used as surrogate markers of OeP to differentiate obstructive from central hypopneas during sleep.

Objective: We explored the causal relationship between Head-Body positions (HBp) and Sleep obstructive apnea/hypopnea (OAH) to fill the evidence gap in this topic. Methods: Rotations (pitch/yaw) and directions (prone, supine, up/left, supine, right) of HBp were captured by two miniaturized sensors under 45° segments during type1 PSG in 20 consecutive OSA patients. Sequential categorical data of Arousals (A, n=559) and OAH events (n=630) (accordingly to AASM2012) associated with HBp status (n=1633) were built. A timestamps-based algorithm was applied to extract every possible combinations between changes in HBp and events. The association between HBp and events risk was evaluated with Lasso-logistic and RandomForest models Results: Head and body prone pitching (PP) and right yawing (RY) were found significantly associated to either A or OAH events. Head position played the most important role (OR=2.2 for PP and 3.9 for RY), compared to that of body postural changes (OR=1.2 for RY and OR=1.1 for PP). Chronological order analysis showed that HBp changes occurred after A (73.7%) or OAH (72.6%) episodes. A and OAH events were triggered by HBp in only 23% observations. Rarely, A and OAH events occurred during a long and stable HBp (2.5%). Conclusion: Prone pitching and right yawing were the most important positions contributing to the classification of A and OAH events. HBp seemed here more likely consequences than causes of A or OAH risk.

Objective: We developed and validated an artificial intelligence based diagnostic platform for OSAS diagnosis by combining mandibular movement indices with co-morbidities and self-reported symptoms. Methods: 400 OSAS patients were diagnosed by PSG scored using AASM 2012 rules. A neural network model was constructed by unbiasedly combining information of a single channel recording sleep mandibular movements (MM) with anthropometry, self-reported symptoms, Epworth score (ESS) and co-morbidities [https://keras.io]. Indices derived from MM were total sleep time (TST), arousal index (ArI) and respiratory effort disturbances (ORDI) (Sunrise, Erpent, Belgium). Results: Validation on unseen data (n=120) confirmed a robust diagnostic performance, with a NPV=0.94, PPV=0.80, accuracy=91.9% and AUC=0.97. The model can also estimate the impact of each input feature on the overall model outputs. For the overall group, model interpretation Fig.1 revealed that ORDI and ArI were the most important contributors to OSAS diagnosis. Individualized interpretation allowed for confirmation of the plausibility of the diagnosis and improved understanding of clinical status. Conclusion: Sleep MM recordings interfaced with individual clinical symptoms and co-morbidities may provide optimization approaches and enable high accuracy of home-based simplified OSAS diagnostic tools.

L’orthèse d’avancée mandibulaire (OAM) réduit AHI et ODI avec une amplitude d’effet qui doit être mesuré. Son effet de stabilisation du mouvement mandibulaire (MM) lors de l’avancée est peu connu. Nous l’analysons pour le comparer aux indices ODI et AHI lors de la titration. Vingt-cinq SAHOS consécutifs explorés en PSG puis lors de la titration en PG ont été équipés d’un magnétomètre (Brizzy®) pour capturer MM et délivrer un indice horaire d’événements respiratoires scorés manuellement (MM-RDI). L’effet d’OAM sur les 3 variables (MM-RDI, AHI et ODI) est évalué en cours de titration, lorsque le ronflement a disparu, par un modèle mixte à mesures répétées et par inférence Bayésienne. L’âge moyen du groupe est : 44,3 ; l’IMC : 25,4. Avant OAM, les médianes (n/h) de MM-RDI, AHI et ODI sont respectivement : 18,2, 14,8 et 7,6. Le changement moyen (IC95 %) sous OAM est estimé à -17,3 (-11,8 à -2,1) pour MM-RDI, -11,1 (-15,2 à -7,5) pour AHI et -6,5 (-11,8 à -2,1) pour ODI. Pour un seuil de titration relatif des indices de -50 % de la valeur de base, le rapport de vraisemblance (facteur de Bayes, BF) pour MM-RDI est très crédible (seuil = -9, BF = 213), moins crédible pour AHI (seuil = -5, BF = 149) et très faible pour ODI (seuil = -7, BF = 0,79). L’inférence basée sur les ratios de vraisemblance (BF) montre que MM-RDI annonce de manière sensible et crédible, la disparition du ronflement sous OAM. MM-RDI est une mesure très sensible de l’amplitude de l’effet de l’OAM. Ce résultat suggère également qu’un effet principal de l’OAM est la stabilisation des mouvements mandibulaires.