Effects of oropharyngeal exercises on orofacial function and corticomotor excitability in healthy individuals Authors: Dongxiang Mi, Eduardo Castrillon, Mohit Kothari, Taro Arima, Peter Svensson Published: 2026 ABSTRACT AbstractOropharyngeal exercises/myofunctional therapy (OE/MT) offer complementary benefits for obstructive sleep apnea. However, the complex nature of the intervention impedes the understanding of the underlying mechanisms related to sensorimotor function and corticomotor control of involved muscles. This study explored the sensorimotor rehabilitation effects of OE/MT tasks in twenty-two healthy adults (24.1 ± 3.1 years) using a cross-over design to better understand the normal physiology of the tongue before and after these different tasks. Sessions involved breathing training (BT), tongue training (TT), or a no-training control (CT). Transcranial magnetic stimulation (TMS) evaluated motor-evoked potentials (MEPs) of the tongue and first dorsal interosseous (FDI, internal control) muscles before and after each session. Secondary outcomes included orofacial muscular pressure/force, peak expiratory flow rate (PEFR), oral stereognosis ability, and subjective ratings. Results showed TT significantly increased tongue MEP amplitude (P < 0.001) and volume (P < 0.001) compared to BT and CT. Conversely, FDI MEPs were unaffected by any task (P > 0.050). No significant differences were observed in orofacial pressure, PEFR, stereognosis, or subjective ratings (P > 0.050). These findings demonstrate that TT elicits unique corticomotor excitability in tongue motor control, which may represent early-phase neurophysiological adaptation rather than established neuroplastic change, providing mechanistic insight into the neurophysiological effects of OE/MT. BT may need further development and be tested longer before functional effects can be observed.Keywords: Myofunctional therapy, Oropharyngeal exercise, Obstructive sleep apnea, Neuroplasticity, Transcranial magnetic stimulationSubject terms: Health care, Medical research, Neurology, Neuroscience, Physiology Introduction In the past two decades, dental sleep medicine has witnessed empirical evidence indicating that oral healthcare professionals can contribute to evaluating and managing obstructive sleep apnea (OSA) through multidisciplinary collaboration1. OSA is characterized by the recurring occurrence of complete or partial obstructions in the upper airway, commonly accompanied by snoring, compromised sleep quality, and cardio-metabolic diseases2. Globally, 425 million adults have moderate to severe OSA3. Multiple community-based studies revealed that the prevalence of OSA in individuals aged 60 years and older varies between 27% and 80%4. Moreover, as individuals age, there is a gradual rise in the collapse of the oropharyngeal muscles5. The correlation between muscle strength and OSA suggested the potential involvement of sarcopenia in the pathophysiology of OSA in elderly adults6. There is an increasing acceptance of the presence of neurologic factors: inadequate upper-airway sensory responsiveness or motor nerve function may contribute to OSA’s pathogenesis7. Recently, beyond the oral appliances such as mandibular advancement devices (MAD) and continuous positive airway pressure (CPAP), several new treatment strategies have emerged targeting oropharyngeal muscle responsiveness and control: oropharyngeal exercise/myofunctional therapy (OE/MT), hypoglossal nerve stimulation, and daytime transoral electric stimulation8. OE/MT comprises a series of exercises designed to improve tongue position and function, the lip seal, and nasal breathing9. Although evidence has confirmed that OE/MT may improve upper airway stability and reduce daytime sleepiness10,11, the underlying mechanisms of OE/MT are not yet well understood. It is generally theorized via increased muscle tone or improved control of oropharyngeal or upper airway muscles9. Rodríguez-Alcalá et al. proposed a new name for OE/MT when treating OSA as ‘sensorimotor muscle rehabilitation (SMR)’, as some OSA patients had mainly hypotony and gnosis-praxis deficit12, and the training appears to provide a therapeutic tool for improving motor tongue function and oral stereognosis ability13. Recent animal research has provided evidence suggesting that OE/MT (tongue training) may enhance upper airway stability by potentially augmenting the corticomotor excitability of the genioglossus muscle measured by transcranial magnetic stimulation (TMS)14. Human studies have reported abnormalities in cortical excitability in OSA patients, characterized by elevated resting motor threshold (RMT) and cortical silent period15, as well as diminished neural plasticity16. Importantly, changes in corticomotor excitability measured using TMS do not necessarily reflect long-term neuroplasticity or functional improvement. Instead, such changes may represent transient modulation associated with task execution or early phases of motor learning17. The functional significance of training-induced plasticity in corticomotor control of the orofacial muscles is extrapolated to be associated with oral motor function enhancement17. The limitations of OE/MT as an integrated training modality are that the non-standardized training regimens make it challenging to draw solid conclusions from the available evidence and determine the effects of each specific exercise on the overall result8,9. To our knowledge, no studies have investigated the sensorimotor effect of the different training subsets of OE/MT on corticomotor excitability in tongue muscles by TMS. Therefore, this study aimed to explore the sensorimotor rehabilitation effects of OE/MT tasks in healthy humans to better understand the normal physiology of the tongue before and after these different tasks. We hypothesized that the OE/MT intervention influences corticomotor excitability and oral stereognosis, as well as other measures of oropharyngeal function, compared to a control task (no training). Results Of the 27 participants screened for the study, 22 (11 men, 11 women; mean age 23.9 ± 3.2 years; range 20 to 31 years) completed all sessions. Five participants were excluded due to discomfort with the tongue electrode or TMS. Self-reported NRS OE/MT tasks did not evoke significant effects on motivation, fun, pain, and difficulty NRS (P > 0.050). However, all groups showed significantly higher fatigue score at post-training than that of pre-training (F1,63 = 28.536, P < 0.001, Table 1). Table 1Mean and standard error of the mean (SE) of self-reported 0–100 NRS motivation, fun, pain, fatigue, and difficulty scores pre- and post-training in BT, TT, and CT groups. Different lowercase letters in the same row indicate significant within-group differences. *Repeated-measure ANOVA, following pairwise post hoc comparisons (P < 0.050).OutcomePre-trainingPost-training Breathing training (n = 22) Motivation63.2 ± 4.362.0 ± 4.0Fun53.9 ± 4.153.4 ± 3.9Pain0.00.9 ± 0.9Fatigue4.5 ± 2.714.1 ± 3.7*Difficulty18.0 ± 5.118.4 ± 5.2 Tongue training (n = 22) Motivation66.4 ± 3.966.0 ± 4.5Fun61.1 ± 5.159.1 ± 5.5Pain0.00.5 ± 0.5Fatigue4.5 ± 2.616.6 ± 3.9*Difficulty14.1 ± 3.518.0 ± 3.5 Control training (n = 22) Motivation66.1 ± 4.563.4 ± 4.6Fun60.7 ± 4.258.9 ± 4.2Pain0.00.5 ± 0.5Fatigue5.0 ± 2.617.5 ± 4.8*Difficulty18.2 ± 5.718.6 ± 5.3 Mean and standard error of the mean (SE) of self-reported 0–100 NRS motivation, fun, pain, fatigue, and difficulty scores pre- and post-training in BT, TT, and CT groups. Different lowercase letters in the same row indicate significant within-group differences. *Repeated-measure ANOVA, following pairwise post hoc comparisons (P < 0.050). Corticomotor excitability There was a significant interaction between group and time (F2,63 = 6.559, P = 0.003), where the participants in the TT group increased the tongue MEP amplitude after training (Sidak: P < 0.001); the MEPs at 120% and 160% MT were significantly higher after training (Sidak: P < 0.001, Fig. 1b). However, the BT and CT groups showed no significant difference after training (P > 0.050, Fig. 1a, c). In addition, there was a significant main effect of time and intensity of stimulation on the MEP amplitude (F1,63 = 7.565, P = 0.008 and F3.189 = 71.191, P < 0.001, respectively) but no main effect of training was observed across the groups (F2,63 = 1.068, P = 0.350). There was a significant main effect of stimulus intensity (F 3,189 = 77.914, P < 0.001); however, no effect of time and group (F1,63 = 0.557, P = 0.458 and F2,63 = 0.004, P = 0.996, respectively) and no interaction between these factors for FDI MEPs (Figs. 1d-f). There were no significant effects of training on MT for any of the tested muscles (tongue: F2,63 =1.885; P = 0.160 and FDI: F2,63 = 1.336; P = 0.270, Tables 2 and 3). Stimulus–response MEP obtained by TMS of the tongue and FDI area of the motor cortex pre and post BT (a, d), TT (b, e), and CT (c, f) *Significantly higher MEP post- training than pre- training (Sidak, P < 0.001). BT, breathing training; TT, tongue training; CT, control training. MT. motor threshold. MEP, motor evoked potential. Table 2Mean and standard error of the mean (SE) of the tongue corticomotor map area, volume, RMT, and center of gravity (CoG) measures pre- and post-training in BT, TT, and CT groups. Different lowercase letters in the same row indicate significant within-group differences. *Repeated-measure ANOVA, following pairwise post hoc comparisons (P < 0.050).OutcomePre-trainingPost-training Breathing training (n = 22) Area > 10 µV (cm2)6.6 ± 0.56.6 ± 0.5Volume (µV × cm2)66.6 ± 15.773.1 ± 17.3RMT39.8 ± 1.539.5 ± 1.5CoG ant-post (cm)2.9 ± 0.12.9 ± 0.1CoG med-lat (cm)10.0 ± 0.110.0 ± 0.1 Tongue training (n = 22) Area > 10 µV (cm2)6.6 ± 0.47.5 ± 0.4Volume (µV × cm2)52.7 ± 8.298.6 ± 12.7*RMT39.5 ± 1.638.9 ± 1.6CoG ant-post (cm)3.0 ± 0.12.9 ± 0.1CoG med-lat (cm)10.1 ± 0.0210.0 ± 0.1 Control training (n = 22) Area > 10 µV (cm2)6.5 ± 0.56.8 ± 0.4Volume (µV × cm2)53.6 ± 6.652.0 ± 7.2RMT38.9 ± 1.738.9 ± 1.7CoG ant-post (cm)3.0 ± 0.12.9 ± 0.1CoG med-lat (cm)10.1 ± 0.110.1 ± 0.1 Mean and standard error of the mean (SE) of the tongue corticomotor map area, volume, RMT, and center of gravity (CoG) measures pre- and post-training in BT, TT, and CT groups. Different lowercase letters in the same row indicate significant within-group differences. *Repeated-measure ANOVA, following pairwise post hoc comparisons (P < 0.050). Table 3Mean and standard error of the mean (SE) of the FDI corticomotor map area, volume, RMT, and center of gravity (CoG) measures pre- and post-training in BT, TT, and CT groups.OutcomePre-trainingPost-training Breathing training (n = 22) Area > 50 µV (cm2)6.8 ± 0.47.2 ± 0.4Volume (µV × cm2)346.4 ± 66.1367.7 ± 73.4RMT36.6 ± 1.936.6 ± 1.9CoG ant-post (cm)1.0 ± 0.10.9 ± 0.1CoG med-lat (cm)6.0 ± 0.16.0 ± 0.1 Tongue training (n = 22) Area > 50 µV (cm2)6.5 ± 0.56.9 ± 0.5Volume (µV × cm2)356.5 ± 77.2369.0 ± 82.0RMT37.7 ± 1.736.6 ± 1.7CoG ant-post (cm)0.9 ± 0.10.8 ± 0.1CoG med-lat (cm)6.0 ± 0.16.0. ± 0.1 Control training (n = 22) Area > 50 µV (cm2)6.4 ± 0.67.0 ± 0.4Volume (µV × cm2)335.8 ± 64.1338.1 ± 64.2RMT36.6 ± 1.936.6 ± 1.9CoG ant-post (cm)0.9 ± 0.10.9 ± 0.1CoG med-lat (cm)6.0 ± 0.16.0 ± 0.1 Mean and standard error of the mean (SE) of the FDI corticomotor map area, volume, RMT, and center of gravity (CoG) measures pre- and post-training in BT, TT, and CT groups. Figure 2 shows the training effects on the tongue muscle corticomotor mapping. For the mapping area, there was no significant interaction between group and time (F2,63 = 1.623, P = 0.205), and no main effect of training was observed (F2,63 = 0.401, P = 0.671); nevertheless, the mapping volume presented a significant interaction between group and time (F2,63 = 5.164, P = 0.008), where the mapping volume was increased after the training when compared to the pre-training values only in the TT group (Sidak: P < 0.001, Table 2; Fig. 1). In addition, a main effect of time was presented for the mapping volume (F1,63 = 14.625, P < 0.001), where the values after training were larger than pre-training (Sidak: P < 0.001). The mapping area and volume for the FDI muscle were not significantly dependent on the interaction between group and time and training (area: F2,63 = 0.183, P = 0.833; volume: F2,63 = 0.065, P = 0.937, Fig. 3). There were no significant changes among sessions for any COG outcomes (P > 0.050, Tables 2 and 3). Pre- and post- training motor cortex representation of the tongue area in BT (a), TT (b), and CT (c) generated by TMS of multiple scalp sites arranged in a 1 × 1cm2 grid. Arrows indicate directions (A anterior, L lateral, M medial, P posterior). BT, breathing training; TT, tongue training; CT, control training. Pre- and post- training motor cortex representation of the FDI area in BT (a), TT (b), and CT (c) generated by TMS of multiple scalp sites arranged in a 1 × 1cm2 grid. Arrows indicate directions (A anterior, L lateral, M medial, P posterior). BT, breathing training; TT, tongue training; CT, control training. Functional assessment of orofacial muscular pressure and force, pefr and oral stereognosis ability Table 4 presents the secondary outcomes of the assessment pre- and post-training. RM-ANOVA showed no significant interaction effects of group and time on any of the secondary outcomes: tongue lifting pressure (F2,63 = 1.56, P = 0.218); tongue protrusion pressure (F2.63 = 1.308, P = 0.277); left cheek pressure (F2,63 = 0.785, P = 0.461); right cheek pressure (F2,63 = 0.663, P = 0.534); lip closure pressure (F2,63 = 0.525, P = 0.594); PEFR (F2,63 = 0.999, P = 0.374); and oral stereognosis ability (F2,63 = 0.170, P = 0.844). In addition, there was also no main effect of time and group on secondary outcomes (P > 0.007). Table 4Assessment of secondary outcomes pre- and post-training in BT, TT and CT groups.OutcomePre-trainingPost-training Breathing training (n = 22) Tongue lifting pressure (kPa)33.2 ± 2.233.7 ± 2.3Tongue protrusion pressure (kPa)27.7 ± 1.928.0 ± 1.9Cheek pressure left (kPa)14.6 ± 0.815.1 ± 0.7Cheek pressure right (kPa)15.5 ± 0.716.0 ± 0.6Lip closure pressure (kPa)6.4 ± 0.66.7 ± 0.6Peak expiratory flow rate (L/min)458.0 ± 24.6464.5 ± 24.2Oral stereognosis test score20.5 ± 0.521.0 ± 0.5 Tongue training (n = 22) Tongue lifting pressure (kPa)32.8 ± 2.234.2 ± 2.6Tongue protrusion pressure (kPa)28.0 ± 2.129.2 ± 2.4Cheek pressure left (kPa)15.1 ± 0.815.1 ± 0.9Cheek pressure right (kPa)15.7 ± 0.815.8 ± 0.8Lip closure pressure (kPa)6.3 ± 0.56.4 ± 0.6Peak expiratory flow rate (L/min)459.2 ± 25.6460.9 ± 25.4Oral stereognosis test score20.5 ± 0.720.9 ± 0.5 Control training (n = 22) Tongue lifting pressure (kPa)31.9 ± 2.332.1 ± 2.3Tongue protrusion pressure (kPa)27.7 ± 2.027.9 ± 2.0Cheek pressure left (kPa)14.7 ± 0.815.0 ± 0.8Cheek pressure right (kPa)15.8 ± 0.716.3 ± 0.7Lip closure pressure (kPa)6.4 ± 0.66.5 ± 0.6Peak expiratory flow rate (L/min)464.1 ± 23.2464.0 ± 23.8Oral stereognosis test score20.7 ± 0.621.0 ± 0.5 Assessment of secondary outcomes pre- and post-training in BT, TT and CT groups. Discussion This mechanistic study provides novel evidence that a single session of tongue training (TT), but not breathing training (BT), can elicit acute modulation of corticomotor excitability in the corticomotor regulation of the tongue musculature in healthy individuals. The MEP amplitude and cortical motor map volume substantially increased following a single 40-minute session of multidirectional TT. Furthermore, the observed alterations seemed to be limited to training-specific muscles, as there was no indication, based on cortical and MEP characteristics, of any excitability in the FDI muscle linked with the learning of the different training movements. Our previous research has shown that standardized tongue protrusion training (TPT) can induce corticomotor neuroplasticity18. To investigate the underlying mechanisms of the therapeutic benefits of TT as part of an OE/MT for patients with OSA, we deviated from our previously established TPT paradigm. Instead, we provided participants with a clinical and commonly employed TT task for OSA. The observed increase in the stimulus-response MEP induced by TMS immediately after the training session aligns with findings from other investigations on short-term TPT ranging between 15 min and 1 hour19,20. Furthermore, one study reported that 41 min of tongue lifting training (TLT) increased the tongue’s corticomotor excitability and, in addition, also the excitability of the masseter muscle21. In a subsequent study from our group, an interesting but rational finding was that jaw protrusion training in turn promoted neuroplasticity of corticomotor control of tongue muscles22, leading to more attention on the possible effects of mandibular advancement devices (MAD) in patients with OSA. It suggested that the efficacy of these devices may not solely be attributed to anatomical factors, such as the opening of the upper airway through the forward repositioning of the genioglossus muscle23. Matsuzaki and Costa et al. demonstrated that the use of MAD resulted in the induction of neuroplasticity within the corticomotor circuit associated with the muscles of the tongue and masseter in both healthy individuals and in patients with OSA, respectively24,25 The observed outcomes might be ascribed to the convergence of the motor cortical region responsible for controlling the tongue and jaw muscles. This convergence is likely due to the close physical proximity of these muscles and the need for their motor activities to be coordinated22. In addition, a notable augmentation in the amplitude of MEPs was detected, accompanied by a substantial rise in cortical map volume after TT. In contrast, the mapping outcomes in the BT and CT groups exhibited no changes. The RMT, mapping area (> 10 µV), and CoG of tongue muscle before and after BT, TT, and CT did not change. The findings presented in this research align with the outcomes of a prior survey on short-term TT20. In that earlier study, no significant expansion of the tongue motor map region was seen, and there were no discernible changes in the CoG coordinates for tongue skill, strength, and passive training paradigm20. Compared to the expansion of the mapping area in the relative long-term training studies21,26, it seems that changes in MEP and the extension of the map area do not occur at a parallel pace with improvement in motor performance. This finding is consistent with a previous study that reported a lack of cortical expansions but did find significant associations between improvements in success rate and increased facilitation of the MEP amplitudes after a one-hour TPT session27. It is possible that the expansion of the map area may not exhibit a linear correlation with enhancements in motor performance due to its manifestation in the later, gradual stage of training. Compared with mapping area expansion and CoG shift, the calculation formula for mapping volume shows that it reflects changes in muscle-cortical total excitability rather than cortical representation28. Peng et al. reported that somatosensory stimulation of the forearm led to a considerable increase in the map volume of the cortical motor representation, without significant changes in map area or the CoG29. This suggests that electroacupuncture primarily modulates cortical excitability rather than altering the spatial distribution of the motor map. In contrast, changes in map area and CoG are typically associated with long-term somatosensory stimulation or prolonged training. In healthy individuals, long-term daily application of transcutaneous electrical nerve stimulation (TENS) to hand muscles, or daily training of coordinated movements, has been shown to result in enlargement of the cortical representations of the targeted muscles30. In the present study, the increase in mapping volume without a corresponding expansion of map area suggests an amplification of corticomotor excitability within already active cortical regions, rather than recruitment of previously inactive or subthreshold sites. This pattern is consistent with short-term modulation of synaptic efficacy rather than structural reorganization of cortical representation31. Although effect sizes were not explicitly reported, the magnitude of the observed changes is consistent with previous studies demonstrating acute increases in corticomotor excitability following short-term interventions32,33. Furthermore, distinct stages of training may be facilitated by various mechanisms of neuroplasticity34. Costa et al. proposed the inclusion of additional map features in the assessment of cortical representations, specifically in areas that are typically determined by a threshold of 5 µV or 10 µV. These additional features, such as volume and areas ≥ 50% max, allow for a more comprehensive evaluation of the topography of cortical representations35. Incorporating these features takes into account the natural variation in the individuals’ excitability at baseline and enhances the sensitivity to detect even minor changes in cortical maps resulting from short-term interventions. According to systematic reviews, the OE/MT training cycle typically averages 4–8 weeks, with 30–40 min of daily training10. The training duration in this study was selected based on the average daily training load. Taken together with the MEP and volume results, these findings suggest that even a single session of OE/MT can induce early-phase neuroplastic changes. This result is consistent with the findings of Gobin et al., and the fact that H-reflexes did not change after 4 sessions of knee training suggests that the neuroplasticity occurred at the supraspinal level36. Our prior research demonstrated that repetitive tongue and masseter training (5–7 days) correlates with enhanced performance (task accuracy) and elevated MEPs of the corresponding muscle in the primary motor cortex26,37. One limitation of OE/MT is that the training protocol is often broad in scope and requires a long and repetitive training. Additionally, some OSA patients exhibit reduced neuroplasticity in the genioglossus and diaphragm muscles38. Therefore, using an acute measurement/training protocol with TMS to assess patients’ phenotypic variability in cortical excitability and their capacity for neuroplastic changes may help identify those who are likely to benefit more from the training. This is supported by previous findings showing that cyclists exhibit greater motor cortex plasticity after acute training compared to untrained individuals39. Similarly, Baad-Hansen et al. found a significant correlation between training-related functional outcomes and changes in MEP amplitude, indicating that individuals who showed greater performance improvements also exhibited MEP facilitation40. Besides MEPs, it is important to note that MEP amplitude reflects not only cortical excitability but also spinal/trigeminal and peripheral contributions. Recently, Ranieir et al. suggested that the ratio of MEP amplitude to compound muscle action potential (CMAP) amplitude could serve as an accessible marker of corticospinal excitability, enabling amyotrophic lateral sclerosis patient stratification by phenotype, disease stage, and survival41. This biomarker allows for the distinction between cortical and peripheral contributions to MEP modulation. Moreover, it may encourage future studies investigating patterns of cortical excitability in OSA patients to integrate the MEP: CMAP ratio, thereby enabling a more objective biomarker of cortical excitability that is independent of the hypoglossal nucleus in rehabilitation settings. The dissociation between neurophysiological and functional outcomes observed in the present study suggests that the increase in corticomotor excitability reflects early-phase neuroplasticity or transient cortical facilitation, rather than immediate functional improvement. In motor learning frameworks, such changes are associated with the acquisition phase, whereas behavioral improvements typically require repeated training and consolidation42,43. These acute changes may also reflect transient post-activation facilitation, and should not be interpreted as evidence of long-term cortical reorganization. Repeated training over days or weeks would be expected to lead to progressive changes in corticomotor organization, including potential map expansion, stabilization of excitability, and stronger associations with behavioral outcomes17. In line with this interpretation, the absence of functional improvements in the present study can be understood within the temporal dynamics of motor learning and neuroplastic adaptation. In the current study, we decided to utilize orofacial musculature pressure as the primary measure of performance data rather than task accuracy. This decision was based on previous research findings, which have demonstrated that various training subtypes of OE/MT can improve respiratory muscle tone and strength, thereby enhancing upper airway stability44. Additionally, OE/MT has been shown to increase tongue strength, facilitate forward tongue repositioning13, and augment labial closure force to ameliorate mouth breathing45. No significant changes were seen in any of the orofacial musculature pressure outcomes after the interventions of BT and TT in the present study. This lack of significant change may be attributed to the relatively short period of the training session, which lasted for 40 min, compared to the studies conducted on patients with OSA, which typically spanned at least two weeks. The findings of this study are consistent with previous research, which also found no statistically significant change in electromyographic (EMG) root-mean-square and tongue protrusion pressure following either 40 min or 5 days of TLT21. Evidence from a systematic review and meta-analysis indicated that a single bout of strength training (averaging 4–12 movements per set, 6–8 repetitions) shifts M1 intracortical circuitry towards facilitation (increased MEP) and improves corticospinal transmission31. Another multivariate meta-analysis reported that strength improvement requires 2–4 weeks of strength training, with the effect being particularly evident in metronome-paced and resistance training46. Besides, they regressed the true effect sizes of strength change on MEP amplitude change using a SEM-based meta-analysis, indicating that changes in MEP amplitude strongly predict changes in muscle strength after resistance training, and the increase in strength is attributable not solely to muscle hypertrophy, with neural adaptations also playing a role in the resultant force gains46. One of the notable discoveries in recent animal studies is that an 8-week tongue strength and skill training regimen has been found to have several significant effects. These include increased EMG activity, maximum voluntary force, and corticomotor excitability of the tongue muscle. Additionally, this training has been shown to enhance the stability of the upper airway, as measured by critical pressure14. The findings from prior research investigating the influence of various training regimens on neuroplasticity have shown inconclusive results36,43. The TT paradigm used in this research was a combination of skill training, which consisted of random sequenced and multidirectional exercises and endurance training. Previous studies on TT in humans utilizing an oral typing keyboard have shown that multidirectional training leads to a greater functional engagement of motor cortical neurons than bidirectional training. Multidirectional training has been found to induce neuroplasticity, characterized by both positive and negative change sites on the motor map47. A study conducted on nonhuman primates has demonstrated that training in three different directions of tongue protrusion increases the firing rate of direction-sensitive neurons. This increase in firing rate is observed not only in the tongue-MI but also in the tongue-SI. In the tongue-SI, individual neurons exhibit significant changes in spiking activity that correspond to the direction of tongue protrusion48. Improvements in the monkeys’ behavioral performance were accompanied by neuronal changes occurring in parallel in face-MI and face-SI over short (within minutes) and long (over days) time scales49. Sessle et al. also reported an increased number of face-SI and face-MI neurons showing tongue protrusion-related activity and lingual mechanosensory receptive fields after weeks of tongue-task training in monkeys42, and increased sensory inputs caused by training or stimulation can also induce face MI neuroplasticity in humans50. In addition to its involvement in orofacial motor control, the face-SI also plays a significant role in touch localization and stereognosis. One retrospective study reported that 11–18% of pediatric and adult patients with OSA exhibited atypical tongue stereognosis and praxis abilities12. That study revealed that impairments in the cortical imprinting processing of tongue stereognosis and praxis abilities during early development may persist as dysfunctions into adulthood12. Likewise, Wallace et al. suggested that individuals with OSA have diminished upper airway sensation, which may play a role in OSA’s pathogenesis and might serve as a therapeutic intervention target51. A pilot study discovered that people with OSA showed improvements in tongue strength and oral stereognosis ability after smartphone-guided OE/MT treatment, and the authors suggest that the sensory muscle rehabilitation may be attributed to brain stimulation through proprioceptive training13. The present study found that the oral stereognosis score exhibited no significant changes following a short-term TT and BT session. This outcome can be attributed to the fact that the participants were in good health and predominantly in their early twenties. The average accuracy in the current research, around 85%, aligns with the findings of a previous study examining young, healthy individuals, which reported accuracy rates ranging around 87%52. All training groups, including the CT group, were required to concentrate and perform training following the PowerPoint guides. After 40 min of training, all participants subjectively reported fatigue. This study did not objectively measure peripheral fatigue using biological indicators such as lactate. The study by Klein et al. showed that moderate-intensity exercise-induced changes in lactate do not predict changes in corticospinal excitability53. The current study requires acknowledgment of certain limitations. This study aimed to assess the effects of different OE/MT subtypes on the central nervous system in healthy participants. Therefore, the findings cannot be directly generalized to patients with OSA. It should be considered a proof-of-concept neurophysiological study in healthy individuals rather than evidence supporting therapeutic efficacy in OSA. Importantly, the present findings do not demonstrate clinical benefit, but rather reflect acute neurophysiological responses to the training tasks. Further research is warranted to determine whether these changes translate into meaningful functional or clinical outcomes in patient populations, and to further investigate the neurophysiological mechanisms of OE/MT in OSA. Second, it should be noted that the duration of the training was relatively brief. Further studies are needed to evaluate the effect of TT in an OE/MT on motor cortex neuroplasticity from a long-term perspective. Furthermore, monitoring the specific training effects (accuracy, precision) was not conducted. Several studies have conducted additional observations to monitor TT outcomes20,33. Considering the inherent features of the present cross-over study approach, the corresponding wash-out period and overall study duration will be multiplied if the training duration is prolonged and the subsequent changes are monitored. Nevertheless, the present study represents a feasible model to examine such parameters before clinical studies are designed. It cannot be denied that the cross-over design did not include formal statistical testing for carry-over effects, which may influence neurophysiological outcomes. Thirdly, only tongue muscle TMS-related outcomes were measured. Other oropharyngeal muscle TMS measurements were not conducted due to technological limitations. Specifically, participants would have required a trans-nasally intraluminal catheter with electrodes to record MEP over the pharyngeal motor cortex. Fourth, the distinction between the impact of cortical and hypoglossal neurons on MEP alterations was not made in our study. Fifth, this study applied easy-use landmark-guided navigation. To improve the accuracy of TMS targeting, robotic MRI-guided navigation is needed. Additionally, combining TMS with structural and functional MRI may help account for interindividual variability in evoked corticomotor responses54. In addition, the absence of peripheral measures (for example, CMAP or M-wave) limits the ability to isolate cortical contributions to MEP changes. In conclusion, this study provides mechanistic insight into the neurophysiological effects of OE/MT by demonstrating that performing a clinically feasible TT can elicit a unique plasticity of corticomotor excitability associated with tongue motor control. Further investigation is warranted to determine the neurophysiological and therapeutic effects of OE/MT in a long-term perspective in patients diagnosed with OSA. The present study has helped to explore the feasibility and necessity for multi-dimensional assessment of OE/MT in order to better understand mechanisms associated with OSA and its management. Methods Participants Twenty-seven healthy individuals (40.7% women; mean age 24.1 ± 3.1 years) were recruited for preliminary screening. The volunteers were recruited by advertising on a webpage of the Section for Orofacial Pain and Jaw Function, Aarhus University, Denmark (http://odont.au.dk/om-odontologi/sektioner/kof/) and other social media platforms. Inclusion criteria were defined as age > 18 years, right-handedness and no ongoing orofacial pain or other types of chronic pain in the last 6 months (pain-related TMD symptoms were excluded using the TMD pain screener)55. Exclusion criteria included the presence of possible OSA issues (STOP-BANG score ≥ 3)56, neuromuscular diseases, significant craniofacial abnormalities, a short tongue frenum, extensive nasal obstruction, or a history of oropharyngeal surgery, the presence of contraindications to TMS, such as metal implants in the cranium, history of epilepsy, and pregnancy57. The study was conducted in accordance with the second version of the Helsinki Declaration and with regional ethics committee approval (1–10-72-9-23, Aarhus University). All participants were allowed to be accompanied by family or acquaintances while explaining the procedure, and were blinded to the study hypotheses. Study design This mechanistic study was conducted using a cross-over design comprising three sessions, with a minimum of one week’s interval between sessions to minimize potential carry-over effects. This interval was based on previous studies26,32, demonstrating that corticomotor assessments, including MEP and cortical mapping outcomes returned to pre‑training levels, after one week of tongue muscle training, followed by a two-week washout period, Each session entailed one of the three different training tasks: breathing training (BT), tongue training (TT), or control training (CT: no training). The allocation of tasks to participants was randomized (Fig. 4a). Before and immediately after each training session, corticomotor excitability was assessed using TMS. At the same time, the assessments of orofacial muscular pressure and force, peak expiratory flow rate (PEFR), oral stereognosis ability, and self-reported numerical rating scales (NRS) scores were conducted to evaluate the impact of the interventions. (a) Overview of the study design. Each participant was randomly assigned to one of six counterbalanced sequences (#1–6). A washout period of at least one week was allocated between sessions (BT: Breathing Training; TT: Tongue Training; CT: Control Training) to prevent carry-over effects. (b). Timeline and protocols for the BT (Breathing Training) and TT (Tongue Training) sessions. Top panel (BT Session): Three different breathing exercises were performed in cycles: BT1; breathe in maximally through the nose and exhale maximally through the mouth using a straw placed in a glass of water and make bubbles whenever possible; BT2; breathe in maximally through the nostril and exhale through the mouth with sufficient force to inflate a balloon and repeat it without removing the balloon from the mouth, and BT3; breath in maximally through one nostril and exhale maximally through the other while closing the other nostril with the subject’s thumb. These three tasks were repeated in 9 cycles (3 tasks [BT1–3] × 9 cycles = 27 total tasks). Bottom panel (TT Session): A set of six multidirectional tongue exercises (TT1–6) constituted one task unit, which was repeated 27 times (1 unit [TT1–6] × 27 repetitions = 27 total tasks). The total duration for each session (BT, TT, and CT) was 40 min. Before and immediately after each training session, the assessments of corticomotor excitability, orofacial muscular pressure and force, peak expiratory flow rate (PEFR), oral stereognosis ability, and self-reported NRS (numerical rating scales) questionnaire were conducted (↑). Training tasks BT and TT were derived from the evidence-based systematic and literature review on myofunctional therapy for OSA9,10. Detailed information on training task modalities and examples is shown in Figs. 4b and 5. Each task was accompanied by a PowerPoint file that counted the time and instructed the participant to perform the corresponding training maneuver. (a-c). Examples of breathing training exercises BT1-BT3; (d-i): Examples of tongue training exercises TT1-TT6. BT consisted of three exercises (Fig. 5-a, b, and c): BT1. Breathe in maximally through the nose and exhale maximally through the mouth using a straw placed in a glass of water and make bubbles whenever possible (Fig. 5a); BT2. Inhale maximally through the nostril and exhale through the mouth with sufficient force to inflate a balloon. Repeat without removing the balloon from the mouth (Fig. 5b); BT3. Breathe in maximally through one nostril and exhale maximally through the other while closing the other nostril with the subject’s thumb (Fig. 5c). A 30-second interval was permitted between each task. The total duration of the BT session was 40 min. TT series consisted of six exercises (Fig. 5d, e, f, g, h, and i): TT1. Press the entire tongue against the front of the palate as firmly as possible for four seconds and slide the tongue backward (Fig. 5d); TT2. Protrude the tongue tip forward as far as possible and maintain this position for four seconds (Fig. 5e); TT3 and TT4. Move the tongue to the left and right corners of the mouth as far as possible and maintain this position for four seconds (Fig. 5f and j). TT5. Stick out the tongue, reach the chin with the tongue tip as far as possible, and maintain this position for four seconds (Fig. 5h); TT6. Stick out the tongue to reach the nose with the tip as far as possible and maintain this position for four seconds (Fig. 5i). There was a 30-second break between each series (= TT1–TT6). The total duration of the TT was 40 min. CT (sham therapy) consisted of 40 min of exercises without therapeutic function (relaxation and stretching of the neck muscles). Corticomotor excitability assessed by transcranial magnetic stimulation (TMS) The TMS technique was used to trigger motor-evoked potentials (MEP) in order to assess the corticomotor excitability and the possible effects of the training tasks. The participants were seated on a dental chair in a reclined position supported by a headrest. An elastic silicone cap was placed over the head, standardized based on anatomical markers and following the International 10–20 Electrode Placement System guidelines58. The electromyographic (EMG) activity was recorded from the right side of the tongue dorsum and FDI (First Dorsal Interosseous) muscle by placing disposable self-adhesive silver chloride electrodes (ALMEVAN, Myotrace, Spain) on the right dorsal surface of the tongue (2–3 mm from midline, 10 mm from tongue tip) with an inter-electrode distance of 20 mm. After using a clean tissue to gently dry the surface of the tongue, the surface electrodes were applied. Throughout the measurement, saliva was intermittently aspirated using the dental chair’s suction device. When poor electrode contact was detected or signal-to-noise ratio varied, the electrodes were replaced promptly20, while disposable surface electrodes (Ambu, Neuroline 720, Denmark) were placed over the right FDI19. The sampling rate was 4 kHz and the EMG signals were amplified, filtered (10–3,000 kHz) and stored in an electrodiagnostic system (Sierra Summit, CADWELL, USA). TMS (Magstim 200, The Magstim Co. Ltd., UK) pulses were delivered with a 5 cm diameter focal figure–of–eight stimulating coil to the left side of the scalp. The coil of the stimulator was oriented 45° obliquely to the sagittal midline so that the induced current flowed perpendicular to the estimated alignment of the central sulcus33. Following previous studies, MEPs in the right tongue musculature could be evoked by stimulation of discrete areas of the left scalp, approximately 2–3 cm anterior to the vertex (Cz) and 7–10 cm lateral to the midsagittal plane. The MEPs from the FDI (control) were evoked by stimulation of the scalp about 1 cm anterior to the Cz line and about 6 cm lateral to the midsagittal plane20,33. Markings on the coil and reference lines on the cap helped identify the position related to the scalp sites. The cranium locations (‘hot spots’) at which EMG responses were elicited in the tongue or FDI muscles at the weakest stimulus amplitude were identified and used for subsequent stimulations. The motor threshold (MT) was determined in the relaxed muscles using a descending and ascending method and was defined as the minimum stimulus intensity that elicited 5 out of 10 discrete MEPs that were distinguishable from the background EMG activity (tongue MEP > 5 µV and FDI MEP > 50 µV)33. The MEPs were recorded using a stimulus–response (S-R) curve, followed by a corticomotor mapping evaluation. Peak-to-peak amplitudes were measured for 0.9, 1.0, 1.2, and 1.6 times the MT, where MT was measured each time the S–R curve was created32. The MEP amplitude (µV) was the mean of 12 stimuli administered at each stimulus level with a 10–15 s interval between stimuli. For corticomotor mapping, each grid site received 8 stimuli at 1.2 times the MT. The grid was stimulated in a fixed pattern, beginning in the center of the hotspot and moving anteriorly and posteriorly at increasing and decreasing latitudes (the sites typically covered 3 cm from Cz and 3 cm anterior and posterior to the interaural line, which corresponded to a total of 9 grids with 1 cm steps)25. When the amplitude of MEPs exceeded 10 µV (tongue) and 50 µV (FDI), the grid point was counted as positive and analyzed as the mapping area. In addition, Volume (uVxcm2: Averaged peak-to-peak MEP amplitude at each grid point35. Following Ridding et al., the center of gravity (CoG) was calculated to identify potential alterations in each muscle’s representation on the motor cortex maps59. Orofacial muscular pressure and force Measurements of tongue lifting and protrusion pressure, lip and cheek closure pressure were recorded in kilopascals (kPa) with the use of a pressure measurement device (TPM-01, JMS Co.Ltd, Japan), by pressing a disposable standardized air-filled bulb connected to a probe. To measure anterior tongue lifting pressure, the bulb was positioned on the anterior part of the palate with the lips closed. The participants then raised their tongues and compressed the bulb onto the palate. Tongue protrusion pressure was measured with the bulb positioned behind the upper and lower incisors; the probe was placed horizontally in the cuspid and premolar region, lightly bitten and held in position. Participants were instructed to protrude the tongue as hard as possible against the lingual surface of the anterior teeth60. For lip pressure, the bulb was centered in front of the upper central incisors. The participants were instructed to purse their lips and compress the bulb between the lips and the labial surface of the anterior teeth while maintaining the teeth occluded. To measure cheek pressure, the bulb was positioned in the space between the upper and lower first molars and the buccal mucosa on both sides; the participants then closed their lips and compressed the bulb against the buccal surface of the molars61. Three measurements were taken, with the maximum value being recorded. Peak expiratory flow rate (PEFR) PEFR measurements were recommended over respiratory muscle strength (RMS) measurements to assess respiratory function without lung disorders. Direct evidence of a correlation between PEFR and RMS has been reported in a previous study62. PEFR was measured by a digital peak flow meter (asma-1, Vitalograph, UK) with a disposable mouthpiece. Volunteers were instructed to take a maximum inhalation, followed by a maximum rapid and intense exhalation. Three measurements were taken, with the maximum value being recorded. Oral stereognosis ability Six small test pieces with the most extended side length or diameters equal to 10 mm were used and combined into three pairings of related shapes, such as (a) circle and ellipse, (b) square and rectangle, and (c) triangle and semicircle52. The test pieces were constructed from 3D-printed and sterilizable materials to allow free oral manipulation (VOVO V-Print SG, VOCO GmbH, Germany), and dental floss was affixed to prevent aspiration. A chart depicting the enlarged forms of the test pieces was used as an identification aid to illustrate the procedure. After the explanation, participants had to wear a blindfold and were not informed of the test results. Each test piece was placed on the tongue of the participants in a random order, which was the same for all participants. The participants were free to manipulate the test piece in their mouths without using their teeth to help. The participants raised their hands when they could provide a distinguishable result, and after the examiner removed the test piece, the participant provided their responses. Using a three-point scale, each response was graded. The correct, partially correct, and incorrect responses were awarded 2, 1, and 0 points, respectively. If the response was not precise but fell within the group shape, it is partially correct. The test was carried out twice, and the cumulative score ranged from zero to twenty-four points. Self-reported numeric rating scales Participants reported perceived levels of motivation, fun, pain, fatigue, and difficulty before and after the training tasks. These were reported using separate 0–100 numerical rating scales (NRS). On the scale, ‘0’ indicated no motivation, fun, pain, fatigue, or difficulty, and ‘100’ indicated the highest level of motivation, fun, pain, fatigue, or difficulty25. Statistical analysis The following outcome variables were evaluated: (a) MEP and corticomotor mapping of the tongue and FDI muscles (primary outcomes); (b) functional assessment of Orofacial Muscular Pressure and Force, Peak Expiratory Flow Rate (PEFR), Oral Stereognosis Ability; and c) Self-Reported NRS (secondary outcomes). All outcome variables were reported as means and standard error of means (SE) unless otherwise noted. Normal distribution was assessed with Q-Q plots, and Ln transformations were applied to the continuous variables when needed. Based on previous studies25,32, It was expected that a medium effect size of 0.2 for the differences in corticomotor excitability would be possible to quantify the magnitude of observed effects,, considering the within-between interactions from ANOVA, with a power of 80% and a significance level of 5%. Therefore, the sample size estimation was considered to include at least 22 participants. One between-subject factor, i.e., group—3 levels (BT, TT and CT), and two within-subject factors, i.e., time—2 levels (pre- and post-training) and stimulus intensity—4 levels (0.9, 1.0, 1.2 and 1.6 times the MT), were considered in a mixed ANOVA to assess differences in the tongue and FDI MEP amplitude. In addition, repeated-measure ANOVA was performed to assess differences in the MT, map area, volume (Ln transformed values), and CoG, orofacial muscular pressure, PEFR, oral stereognosis ability and NRS scores between the groups (between-subject factor—3 levels), with time—2 levels (pre and post training) serving as the within-subject factor. When appropriate, post hoc analyses using the Sidak test were conducted. The significance level was established at 5% (P = 0.050). The adjustment was made to account for these outcomes (N = 7) due to the multiple comparisons in the secondary outcome analyses. Therefore, we implemented the Bonferroni correction and established the significance level at 0.05/7 (P = 0.007). Author contributions D.M. : Methodology; Investigation; data curation; formal analysis; writing – original draft. E.C. : Conceptualization; methodology; project administration; formal analysis; writing – review and editing. M.K. : Conceptualization; methodology; project administration; formal analysis; writing – review and editing. T.A. : Conceptualization; methodology; writing – review and editing; funding acquisition; supervision. P.S. : Conceptualization; methodology; project administration; formal analysis; writing – review and editing; funding acquisition; supervision. Funding Open access funding was provided by Suzhou Vocational Health College (SZWZYQDJ2511). The study was funded by the Danish Dental Association and Japan Science and Technology Agency SPRING, Grant Number JPMJSP2119. Data availability The data supporting the findings of this study are available from the corresponding author upon reasonable request. Due to ethical considerations, access to the data is subject to restrictions. Declarations Competing interests The authors declare no competing interests. Informed consent An informed and written consent form agreeing to participate was obtained from each patient. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data supporting the findings of this study are available from the corresponding author upon reasonable request. Due to ethical considerations, access to the data is subject to restrictions.