Contingent Negative Variation

     and SAS
  a.1. Indications
         Oral Appliance
   a.3.1. AHI
   a.3.2. Sleep Quality 
   a.3.4. Snoring
   a.3.5. Blood Pressure
   a.3.6. Upper Airway 
             Resistance Syndrome
   a.3.7. Side Effects
   a.4.1. Anatomical Factors
   a.4.2. Functional Factors
   a.4.3. Sleep Position 
         Periodontal Disease
         Heart Disease
         Metabolic Syndrome

Contingent negative variation (CNV) is a long-latency event-related potential which develops in the interval between two stimuli, in which the first stimulus (S1) serves as a warning that prepares the subject to expect the second imperative stimulus (S2) which requires a decision or a motor response [162]. However, motor responses are not necessary to provoke the potentials, which occur during mere stimulus anticipation [163, 164]. It is assumed that CNV represents the neuronal activity necessary for sensorimotor integration or association and, therefore, is interpreted as an expression of the cognitive processes in preparation for a response directed to a purpose [162]. CNV consists of two components: a rapidly habituating early frontal component related to arousal and attention to the warning stimulus (S1), and a late component with a more central distribution [165]. The latter involves neural activities holding a motor response in readiness. The CNV is linked to different mental states and activities including arousal, stress, attention, expectation, level of vigilance, the will to elaborate a response, decisional performance, time estimation, and motor response preparation [166-168]. The frontal cortex has been considered as a likely generator for CNV in part because of its fronto-central scalp distribution, and the paradigm generating CNV is similar to memory tasks known to be dependent on the prefrontal cortex. Anterior cingulate gyrus, caudate nucleus, thalamus, and reticular formation may be crucial for the generation of the early CNV, and the dorsolateral prefrontal cortex may be involved in the generation of the late CNV [164]. Ikeda et al. [169], in a study with subdural recordings, stated that orbito-frontal and mesial frontal areas play an important role in regard to cognition and decision making. They reported that the basal ganglia are the most likely responsible for the generation of the late CNV [170]. Task-specific CNV amplitude loss was observed in patients with cervical dystonia [171] and writer's cramp [172]. The clinical symptom of dystonia may result from a deficient compensatory mechanism for abnormal motor programs in response to sensory stimuli [172, 173]. Reduced amplitude of movement-related cortical potentials has been reported in writer's cramp [173], torsion dystonia [174] and oromandibular dystonia [175]. These studies showed similar physiological abnormalities between other focal dystonia and oromandibular dystonia, suggesting a common pathophysiology [175]. It became of interest to clarify whether similar results in a cognitive paradigm would be found in oromandibular dystonia. So far, CNV has been studied mainly in association with the movements of the fingers, hands and feet. CNVs for various voluntary jaw and tongue movements and their cortical distribution have not been explored to our knowledge. In this study we examined CNVs associated with various jaw, tongue protrusion and hand extension tasks and compared the distribution and amplitude of the potentials among the movements to elucidate the motor control mechanism underlying mandibular and tongue movements.

We evaluated 10 healthy subjects (8 men and 2 women, ranging in age from 25 to 42 years with a mean of 34.4 years). They were all right-handed according to the Edinburgh inventory [148]. The warning signal (S1) was delivered using an apparatus, which was placed about 1 m in front of the subject and indicated 5 kinds of experimental movements by flashing 5 light emitting diodes. The upper light indicated mouth opening, lower light; mouth closing, left light; left lateral movement, right light; right lateral movement; middle light; tongue protrusion. These 5 signals were randomized but flashed with equal frequency on average. Two seconds after S1, a 1,000 Hz tone burst (S2) of 100 ms duration was delivered to both ears simultaneously through the earphones. The interval between S2 and the next S1 varied randomly between 4 and 7 s. During testing the subject sat in a comfortable reclining armchair with a head rest in a dimly lit, isolated and quiet room. The subject was asked to make the experimental movements from the jaw rest position as quickly as possible after hearing S2. To confirm whether the subjects performed the experimental movements correctly, three-dimensional incisal movements were recorded using a Mandibular Kinesiograph (MKG K6 Diagnostic System, Myo-tronics, Seattle, USA) [130, 176]. The right hand extension task was performed in the same fashion except for the S1. The subject was instructed to extend the right wrist after hearing the S2 irrespective of the 5 kinds of S1. The recording session for mandibular and tongue movements, and that for hand extension movements consisted of 150 and 30 trials respectively, and two sessions of each task were held interleaved by a rest period.

Electroencephalograms (EEGs) were recorded with a total of 19 surface tin cup electrodes (NeuroSoft, Inc. Virginia, USA) fixed on the scalp with gel according to the International 10-20 System (Fp1, Fp2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, O1 and O2). Each electrode was referred to linked ear lobe electrodes. Electrooculograms (EOGs) were recorded with a pair of the same surface electrodes, one placed 2 cm above and the other 2 cm below the left canthus. EEG and EOG were recorded with a band-pass filter of 0.05-100 Hz. Electrode impedance was kept below 3 kΩ. Surface electromyograms (EMGs) were obtained with a band-pass filter of 0.05-100 Hz from a pair of tin cup electrodes placed on the skin overlying the masseter muscle (for mouth closing) and the suprahyoid muscles (for mouth opening, lateral movement and tongue protrusion) or the extensor digitorum communis muscle (for hand extension) on the right side. Suprahyoid EMG was used for lateral movement and tongue protrusion, since the movements accompany mouth opening, and the anterior belly of the digastric is similarly activated [149, 150].

All data were digitized with a sampling rate of 200 Hz and stored in a computer for off-line analysis (Scan™ 4.1, Neuroscan Labs, Inc. Virginia, USA). EEG data were averaged from 3 s before to 1 s after S2. The baseline was determined by averaging the 1 s epochs before the onset of S1. EEGs, EOGs and rectified EMGs were displayed on the screen. Samples contaminated with artifacts such as eye movements or EMGs were completely rejected. The grand average of CNVs for each task was obtained across all subjects by averaging the EEGs. The onsets of early and late CNVs were obtained as the inflexion points of waveforms to a more negative phase. The early CNV amplitude was defined as the mean value 500 ms after the S1 onset and late CNV amplitude as that at 1,900 ms after the S1 [172].

Figure 35. Grand average waveforms of contingent negative variations (CNVs) across all subjects for six different kinds of voluntary movements. A gradually increasing, bilaterally widespread negativity is observed, with the maximum in the vertex region for all tasks. Open: mouth opening, Close: mouth closing, Left: left lateral movement, Right: right lateral movement, Tongue: tongue protrusion, Hand: hand extension, EMG: masseter EMG for mouth closing, extensor digitorum communis EMG for hand extension, suprahyoid EMG for the other tasks, EOG: electrooculogram, MKG: Mandibular Kinesiograph, C: close, O: open.

Statistical analysis was done by analysis of variance (ANOVA) with repeated measures design on the data using the following two factors: Task (six tasks) and Electrode (19 scalp sites). Task x Electrode interactions for early CNV and late CNV were examined with two-way ANOVA. To assess the significance in difference of amplitude at specific electrode positions, we used separate one-way ANOVA. The differences of these values were tested with a Tukey test.

The amplitudes of EMGs, EOGs and MKG showed no change after S1, indicating that the waveforms were not influenced by jaw and tongue muscle activities and their movement artifacts (Figure 35). After the S1, negative and positive peaks appeared with latencies of about 100 ms and 300 ms, respectively , and then a slowly increasing negativity was seen diffusely with the maximum in the vertex region (Figures 36, 37). The early CNV was identified about 400-500 ms after the S1, mainly at the midline-frontal (Fz) and midline-central (Cz) areas (Figure 37). The late CNV started approximately 1000-1200 ms after the S1 and gradually increased until the S2 (Figure 37). After the movement onset, the potentials were contaminated with artifacts arising from the EMGs of the masticatory muscles and movement artifacts just after the movement onset.

Figure 36. Early CNV and late CNV for mouth opening movement. The early CNV occurred 400-500 ms after S1. The late CNV was identified 1,000-1,200 ms after S1, and slowly increased until movement onset.

Figure 37. Grand average waveforms of CNVs at Cz (midline-central area) across all subjects for the six movements. The early CNV occurred 400-500 ms after S1. The late CNV was identified 1,000-1,200 ms after S1, and slowly increased until movement onset. After the onset, the potentials were contaminated with artifacts. Open: mouth opening, Close: mouth closing, Left: left lateral movement, Right: right lateral movement, Tongue: tongue protrusion, Hand: hand extension.

The two-way ANOVA for early CNV disclosed significant interactions in Task x Electrode (F=91.42, p<0.0001, Task: F=84,03, p<0.0001, Electrode: F=1084.2, p<0.0001). The results of two-way ANOVA for late CNV showed significant Task x Electrode interaction (F=66.71, p<0.0001, Task: F=95.0, p<0.0001, Electrode: F=1253.83, p<0.0001).

The early CNV amplitudes measured at 500 ms after S1 were maximum at Fz (mean ± standard error, mouth opening: -8.0 ± 0.9 µV, mouth closing: -2.6 ± 0.7 µV, left lateral excursion: -8.5 ± 1.0 µV, right lateral excursion: -3.7 ± 0.6 µV, tongue protrusion: -3.0 ± 0.8 µV, hand extension: -1.5 ± 0.3 µV) (Figure 38). The mean amplitude at Fz was significantly lower for the hand extension (p<0.001, ANOVA) than for other jaw and tongue movements except mouth closing. The amplitude for the left lateral movement was significantly higher (p<0.001) than for mouth closing, right lateral movement and tongue protrusion tasks. The right lateral movement showed significantly higher (p<0.001) amplitude than those for mouth closing and tongue protrusion.

The late CNV amplitude measured at 1,900 ms after S1 was maximal at Cz (opening: -15.7 ± 2.1 µV, closing: -8.2 ± 1.3 µV, left: -17.1 ± 2.2 µV, right: -15.3 ± 1.9 µV, tongue: -12.5 ± 1.6 µV, hand: -3.1 ± 0.4 µV), and it was almost symmetrical (Figure 38). The mean late CNV amplitude at Cz was significantly higher (p<0.001) for the left lateral movement than for right lateral movement, tongue protrusion, mouth closing and hand extension. The mean amplitude was significantly lower for the hand extension task (p<0.001) than for the other movements. The late CNV amplitudes for right lateral movement and mouth opening were significantly higher (p<0.001) than mouth closing and tongue protrusion.

Figure 38. Mean CNV amplitudes at each electrode for the six movements. In general the amplitudes are higher for the lateral movements than for mouth closing, and the hand extension shows the smallest amplitude. Each movement shows a symmetrical distribution of CNV with the maximum at the Cz electrode. Shaded bars are early CNV, and open bars are late CNV. Bars and lines indicate the means and standard errors. Open: mouth opening, Close: mouth closing, Left: left lateral movement, Right: right lateral movement, Tongue: tongue protrusion, Hand: hand extension.

In general, the amplitudes were higher for lateral movements and mouth opening than for mouth closing, and the hand extension showed the smallest amplitude (Figure 38). The early CNV amplitude was maximum at Fz and the late CNV was maximum at Cz for all tasks, but the amplitude for the hand extension at the C3 showed almost the same amplitude as that at Cz (Figure 39). Slowly increasing bilaterally widespread negativities were observed starting after the S1 until movement onset after the S2 in the representative topographical maps (Figure 39). The isopotential map showed symmetrical distribution over the head with the maximum in the vertex region. However, the peak of negativity for the hand extension was distributed between C3 and Cz (Figure 39).

Glossokinetic potentials as well as muscle activities may contaminate the CNV, therefore, we recorded EMGs from the masseter and suprahyoid muscles by surface electrodes, which should have picked up any artifacts from the masticatory, tongue or facial muscles. Since the averaged EMG traces showed no such activities before the movement (Figure 35), contamination is unlikely in this study.

Figure 39. Scalp topography of the CNV for the six movements. The isopotential map of CNVs showed a symmetrical distribution over the head with the maximum in the vertex region. Each voltage scale is coded in blue for negative and red for positive. Open: mouth opening, Close: mouth closing, Left: left lateral movement, Right: right lateral movement, Tongue: tongue protrusion, Hand: hand extension.

Although the exact generator of the CNV in humans is still unclear, multiple cortical and subcortical regions have been suggested to participate in the generation of the CNV [177]. The early CNV was believed to reflect an attention process in the frontal cortex [178]. The pre-frontal and parietal association cortices might be active in the judgment and decision-making process [179]. Cui et al. [180] suggested that there is an individual generator for the early CNV in the frontal lobe. A significant positive correlation between the amplitude of the early CNV and frontal cortex blood flow was reported [181]. The generators of the late CNV were found, using subdural recordings. They include supplementary, primary sensori-motor, pre-motor areas and pre-frontal cortex [177, 178, 182]. Further generator structures subdurally recorded in patients with epilepsy include the primary motor cortex, primary sensory area, supplementary motor area, basal pre-frontal, mesial pre-frontal, temporal and occipital regions [177, 182], and orbito-frontala and mesial frontal areas [169]. Cui et al. [180] suggested that the mesial wall motor areas as well as the primary motor cortexes and primary sensory areas participate in generating the late CNV, but these are not the only areas of generation for the late CNV.

The CNV amplitude may be affected by several factors including reaction times [162], the complexity of the task [183], the intensity of muscular effort [184], age [185] and mental status [186]. The amplitudes were -5-6 µV for hand movements [170, 187], -4.2 µV for finger extension [168] using a simple reaction time paradigm and -9 µV [172] using the S1 choice reaction paradigm. In a simple reaction time paradigm, the S1 and the task are always identical, while in the S1 choice paradigm, more than two kinds of S1 are delivered and the subject performs tasks according to the S1. The CNVs recorded with a simple reaction time paradigm suffered substantial effects of habituation in their amplitudes [171]. The CNV amplitudes for jaw and tongue movements in this study were higher than those already reported values for mouth opening and vocalization (-10-13 µV) [188, 189], which were performed using a simple reaction time paradigm. On the other hand, the amplitude for hand extension tended to be smaller than that reported previously [170, 187]. This discrepancy is possibly due to the different paradigm employed, because considerable variation in CNVs recorded with different paradigms has been reported [179]. In this study the subjects had to decide and prepare for one of the 5 tasks from the placement of the light, while the hand extension was executed similarly for all 5 kinds of S1. The subjects may have paid more attention to the jaw and tongue movements than the hand extension. The map of the movement-related cortical potentials, which begin about 1 s before a self-paced voluntary movement [141] measured at the movement onset for the lateral mandibular movements revealed predominance over the ipsilateral hemisphere to the direction of the movement [130]. In this study, however, all the jaw and tongue movements showed almost symmetrical distribution. The tendency of predominance over the hemisphere for lateral excursion may be overshadowed by the CNV potential, although the late CNV amplitude can involve a part of the movement-related cortical potentials. The map of CNVs for hand movement in this study showed a peak between C3 and Cz. CNV amplitudes were asymmetric, with left hemisphere dominance for right hand movement using the simple reaction time paradigm [171]. In a study using the S1 choice reaction paradigm, however, the difference was significant [172]. This discrepancy may be due to the different paradigms.

Bilateral CNV associated with jaw opening was recorded at several cortical locations, with a tendency for larger amplitudes at temporal placements [188]. In this study, however, as clearly shown in Figure 35, CNV for mouth opening revealed bilaterally spread negativity with the maximum at Cz. No remarkable amplitude was detected by the electrodes at temporal areas. The possibility that the potential at the temporal region was contaminated with temporal muscle EMG in the previous study could not be completely excluded. The maximal negativity recorded over the scalp from Cz in the present study is considered to reflect the activity of the supplementary motor area. The primary motor area for the jaw and tongue movements is thought to be located in the temporal areas, due to the orderly representation of the body within the precentral gyrus. Recently, a study using functional magnetic imaging with simultaneous CNV recording confirmed contributions from the supplementary area, cingulated, thalamus, and bilateral insula in anticipatory attention and motor preparatory processes engendered by the task that are indexed electrophysiologically by generation of the CNV [190]. The maximal negativity found presently is probably due to potentials from multiple cortical generators being summated at Cz through volume conduction. The scalp-recorded late CNV and Bereitschaftspotential (the main component of movement-related cortical potentials) have common cortical generators, but the underlying generating mechanism at the subcortical level may differ: the cortico-basal ganglia-thalamocortical circuit is responsible for the late CNV while the cortico-cerebello-cortical circuit produces the Bereitschaftspotential [170]. The more difficult the task, the larger the amplitude of the Bereitschaftspotential [191]. The amplitudes of movement-related cortical potentials are significantly higher for the lateral excursion of the jaw than for mouth opening and closing [130, 175]. Familiar or semiautomatically performed movements such as mouth opening and closing are associated with smaller amplitudes, which implies more contribution from subcortical mechanisms compared with those in unfamiliar tasks [130, 175]. The maximal CNV amplitude in this study was highest for the left lateral movement and lowest for mouth closing. The amplitude was higher for mouth opening than for right lateral movement, but not significantly. The reason for this difference remains uncertain. Further experiments with firm control of the direction, magnitude and velocity of the movements are needed to clarify the difference and the underlying control mechanism. Based on the similarity of the late CNV to Bereitschaftspotential related to the familiarity of tasks in this study, it may be that the two potentials share at least some cortical generators in common. The basal ganglia are most likely responsible for the generation of the late CNV and cerebellar efferent system for the generation of the Bereitschaftspotential [170].

Kaji et al. [171] recorded CNV in patients with cervical dystonia, by using neck and hand movements, as motor tasks in response to the imperative stimulus. They found a task-specific abnormality; the patients had significantly decreased late CNV amplitudes for head rotation, but not for finger extension. Dystonia is defined as a syndrome of sustained muscle contraction that frequently causes twisting and repetitive movements or abnormal postures. Hamano et al. [172] found significantly decreased CNV amplitudes for finger extension movement in patients with writer's cramp, a form of focal dystonia. Based on these findings, they proposed that cervical dystonia and writer's cramp are associated with defective retrieval or retaining of specific motor programs. Oromandibular dystonia is a focal dystonia manifested by involuntary masticatory and/or lingual muscle contraction [192, 193]. It is likely that similar results in a cognitive paradigm would be found in oromandibular dystonia. Reduced amplitude of movement-related cortical potentials was reported in oromandibular dystonia, suggesting the common pathophysiology and similar physiological abnormalities between other focal dystonia and oromandibular dystonia [175, 194]. CNV recording might be important for elucidating the neuronal activity necessary for the sensorimotor integration and pathophysiology of diseases with abnormal orofacial movements such as oromandibular dystonia or dyskinesia [131] and also SAS (Figure 40).

Figure 40. Scalp topography of the CNV for the six movements in a patient with SAS. The pattern of the isopotential map differs from that of normal control. The difference might be related to sleepiness due to SAS.