Twist and turn into chronic disorders of consciousness: Potential role of the auditory stapedial reflex

2.1 Subjects

Fifteen DOC patients (8 MCS and 7 UWS) and 10 HC participated in this study. All the patients met the criteria for vegetative state and MCS diagnosis (The Multi-Society Task Force on PVS, 1994; Giacino et al., 2004). The exclusion criteria were: pre-existing severe neurological or systemic diseases; actual critical conditions, such as inability to breathe independently and hemodynamic instability; administration of other modifying cortical-excitability drugs than L-Dopa and baclofen; presence of epileptic history, pace-maker, aneurysms clips, neurostimulator, brain/subdural electrodes or other electric/electromechanical devices (TMS safety guidelines) (Rossi et al., 2009); EEG burst-suppression pattern. DOC level was assessed through the CRS-R, which was performed daily for one month, independently by two DOC diagnosis-skilled neurologists. Detailed demographic and clinical characteristics are reported in Table 1. The Local Ethics Committee approved the present study and written informed consent was obtained from either HC or the legal guardian of each patient.

2.2 Experimental design

Patients and HC were lying in a bed in a mild-lighted room. At baseline (T_pre), we assessed the audio-motor domain score of the CRS-R (in DOC patients) and the ASR. Then, each participant underwent two different repetitive TMS (rTMS) protocols, administered in a random scheme at one week of interval: i) a real-rTMS, which consisted of 600 magnetic stimuli delivered at a frequency of 5 Hz over the primary auditory area; and ii) a sham-rTMS, which consisted of 600 sham-magnetic stimuli delivered at a frequency of 5 Hz. We repeated the baseline measures 10 and 45 min after the application of each conditioning protocol (T₁₀ and T₄₅). The experimenters who analyzed the data were blinded on the scheme procedure.

2.3 CRS-R

The CRS-R is a reliable and standardized tool, which integrates behavioral and clinical assessments, and includes the current diagnostic criteria for coma, UWS, and MCS, allowing the patient to be assigned to the most appropriate diagnostic category. It consists of 29 hierarchically organized items divided into six subscales addressing auditory, visual, motor, oro-motor, communication, and arousal processes. The total score ranges from zero to 23. A score of ≤2 on the auditory, motor, and oro-motor/verbal subscales and ≤1 on the visual subscale, and of zero on the communication subscale is consistent with the diagnosis of UWS. Thus, the CRS-R is an appropriate measure for characterizing the level of consciousness and for monitoring the neurobehavioral function recovery (Gerrard et al., 2015).

2.4 rTMS

Each rTMS protocol was carried out by published safety recommendations (Rossi et al., 2009). We delivered 600 stimuli over the left auditory cortex, at a frequency of 5 Hz (3 blocks of 200 pulses in 60 sec, with an inter-train interval of 10 sec), through a standard figure-of-eight coil connected to a Magstim Rapid stimulator (Magstim Company, Whitland, Dyfed, UK), which produced magnetic stimuli with a biphasic waveform and a pulse width of ∼300μs. The site of stimulation was identified in HC by using a standard procedure based on the 10–20-EEG (i.e. 2.5 cm upwards on the line T3–Cz, and then 1.5 cm in the posterior direction perpendicular to the line T3–Cz) (Langguth et al., 2006), whereas it was individually checked according to a recently performed brain MRI scan in DOC individuals. Ears were occluded during the stimulation with the ear tips of an earphone (minimum of 30dB-sound pressure level -SPL- of external noise exclusion). The intensity of magnetic stimulation was set at 90% of resting motor threshold (RMT) from the right first dorsal interosseous muscle (FDI), in analogy to previous TMS works on auditory cortex (Naro et al., 2015; Sowman et al., 2014; Schecklmann et al., 2011). For the sham_rTMS, we used the same TMS set-up, but with a sham coil.

To measure RMT, we positioned a standard figure-of-eight coil, wired to a high-power Magstim200 stimulator (Magstim, Whitland, Dyfed, UK), over the optimum position (hot-spot) to elicit the highest and steepest motor evoked potential (MEP) from the right FDI at rest. The magnetic stimuli had a monophasic pulse configuration, a rise time of ∼100μs, decaying back to zero over ∼800μs. The coil current during the rising phase of the magnetic field flowed toward the handle. Thus, the induced current in the cortex flowed in a posterior-to-anterior direction. The hot spot was identified by moving the coil in steps of 0.5 cm around the presumed hot spot. The coil was held tangentially to the scalp, with the handle pointing backward and laterally to 45° from the midline (approximately perpendicular to the line of the central sulcus). We thus estimated the RMT, which was defined as the minimum intensity able to evoke a peak-to-peak MEP amplitude of 50μV in at least five out of ten consecutive trials in the relaxed FDI muscle (Rossini et al., 2015).

2.5 Auditory Stapedial Reflex (ASR)

All the subjects underwent an accurate otoscopic examination. Then, a tympanogram was performed on each ear to estimate the tympanometric peak pressure (TPP) that served as the basis for the physical ear canal pressures evaluated during the ASR threshold (ASRt) testing. Tympanograms were obtained using a positive-to-negative pressure sweep covering a range of 200/–400 daPa at a pump speed of 400 daPa/sec, with a 226 Hz probe tone. Then, in all participants, ASRt measures were performed at TPP (TPP⁰, which was determined from the tympanogram) and at three positive and three negative pressure values relatively to the TPP⁰ (i.e. TPP^–150, TPP^–100, TPP^–50, TPP⁺⁵⁰, TPP⁺¹⁰⁰, and TPP⁺¹⁵⁰daPa). Ipsilateral (probe and reflex activator in the same ear) and contralateral ASRt were obtained using activator-tone frequencies of 0.5, 1, 2, and 4 kHz. Pulsed activator tones were used for both ipsilateral and contralateral ASRt measurements. ASRt evaluation began at least 10dB below the ASRt, and the activator tone was increased in 5dB steps until a response was obtained. Therefore, the ASRt level was reduced by 10dB and the process repeated. The ASRt was defined as the first ascending level at which two out of four reflex responses, which produced an admittance change of at least 0.02 mmH₂O, were obtained. If three consecutive presentations at 115dB-SPL produce no quantifiable response (being SPL a logarithmic measure of the effective pressure of a sound about a reference value), the ASRt would be labeled as “no response” (Hunter et al., 1999). Likewise, for statistical purpose, if a level of 120dB-SPL (i.e. 5 dB above the maximum safe output) is reached, the resulting response would be labeled as “no response”. The ASRt were obtained in random order within the ipsilateral and contralateral conditions, on each ear canal pressure and each frequency used. The ipsilateral and contralateral conditions were averaged together. Participants were alternated randomly between “ipsilateral first” and “contralateral first” testing.

Admittance measurements (tympanometry and ASRTs) were obtained using a clinical admittance-measuring device (Grason-Stadler Model GSI 33; Eden Prairie, MN, USA). The admittance-measuring device was calibrated yearly to conform to American National Standards Institute (ANSI) calibration standards for admittance measurement devices (ANSI S3.39-1987), and according to previously described procedures (Moller, 1962; Reker, 1977; Rawool, 1995, 1996). Calibration of the 226-Hz probe tone was verified before each measurement session using the cavities of known volumes of air (0.5, 2, and 5 cc) provided by the manufacturer. The ASR eliciting stimuli were 0.5, 1, 2, and 4 kHz pure tones. For 0.5 and 1 kHz conditions, the stimulus duration was 62 ms, with a 9 ms rise and fall time. In one presentation, the 62 ms pure tone was presented 12 times in succession with a silent period of 62 ms between each one. The 2 and 4 kHz conditions were identical to the 0.5 and 1 kHz ones, but the duration of the silent period was 53 ms. Air- and bone-conduction thresholds were obtained using a diagnostic audiometer (Grason-Stadler Model GSI 61 or GSI 16) calibrated yearly to conform to ANSI standards for audiometers (ANSI S3.6-2004). Tones were output to earphones (Telephonics Model TDH-49) mounted on supra-aural cushions (Telephonics Model MX-51/AR) and a bone vibrator (Radioear Model B-71).

2.6 Statistical analysis

We compared the baseline clinical and electrophysiological parameters among HC, MCS, and UWS, through unpaired t-tests. We thus evaluated the effects of the conditioning protocols on ASRt for each frequency and TPP through separated three-way repeated measure analyses of variance (rmANOVA), implying time (three levels: T_pre, T₁₀, and T₄₅) and protocol (two levels: real-rTMS and sham-rTMS) as within-subject factors, and group (three levels: MCS, UWS, and HC) as between-subject factor. Further rmANOVAs were performed to determine any differences in ASRt in the ipsilateral and contralateral stimulation modes. Moreover, we performed rmANOVAs employing also the factors TPP (seven levels: –150, –100, –50, 0, +50, +100, and +150 daPa) and frequency (four levels: 0.5, 1, 2, and 4 kHz) to assess whether rTMS could change ASRt distribution. The Greenhouse-Geisser method was used if necessary to correct for non-sphericity. Conditional on a significant F-value, we performed post-hoc t-tests (Bonferroni) to explore the strength of main effects and the patterns of interaction between the experimental factors. All statistical tests were applied two-tailed. A p-value <0.05 was considered significant. Data were reported as mean±variance. The latter is defined as the average squared difference of the scores from the mean, and expresses how close the scores in the distribution are to the middle of the distribution.

The effects of the conditioning protocols on audiomotor_CRS-R were measured through a Wilcoxon test. Finally, we calculated a Spearman correlation test to assess an eventual correlation among clinical and electrophysiological parameters. All the clinical data are given as means±SD.

3.1 Baseline findings

There were no significant MCS-UWS differences concerning the demographic features, whereas the CRS-R scores were significantly higher in MCS than UWS individuals (Table 1). The auditory CRS-R score at both the T_pre was superimposable to the monthly CRS-R score in each patient. Likewise, the electrophysiological parameters at both the T_pre were similar and stable during the two days of experimentation.

We reported the mean values of the ASRt values for each group in the Fig. 1. Individual data are reported in the supplementary Table 1. RMT and MEP amplitude did not differ between the three groups (p = 0.6). HC individuals showed a typical slope of ASRt (i.e. a concave curve pointing at TPP⁰) at each frequency, and at ipsilateral and contralateral testing. Extreme TPPs at each frequency gave higher ASRt values than mild TPPs did.

All the DOC subjects showed clear ipsilateral responses (i.e., we did not observe “no-response” data), although abnormal as compared to HC. In fact, the ASRt slope was severely destructured (Fig. 1), and had higher values than HC individuals did (p < 0.03), particularly in the extreme TPP values (p < 0.01) (thus suggesting a tonic ASR hyper-excitability). Finally, DOC patients had a marked impairment of contralateral reflex responses when stimulating either the left of right ear, i.e. the ASRt distribution was markedly irregular (Fig. 1).

3.2 Conditioning protocol’s effects on clinical assessment

The Wilcoxon test showed a statistically significant increase of the audio-motor CRS-R score in all the MCS patients (p = 0.04) and two UWS individuals only after the real-rTMS. In detail, five MCS (n. 4, 5, 6, 7, and 8) and two UWS patients (n. 1 and 4) presented a one-point increase at T₁₀ as compared to T_pre (Table 1).

3.3 Conditioning protocol electrophysiological effects

The real-rTMS protocol induced a significant ASRt decrease in all the HC and MCS individuals (i.e. a reduction of tonic ASR hyper-excitability), whereas the UWS subjects showed no evident change (Fig. 1), but two individuals (n. 1 and 4), who also got a transitory clinical amelioration (Table 1). Specifically, these two UWS subjects showed ipsilateral and contralateral electrophysiological after-effects that were similar to the MCS individuals (Fig. 1). On the other hand, the sham stimulation totally lacked significant effects in either HC or DOC sample (Supplementary Table 1).

In particular, we observed that ASRt decreased bilaterally in all the HC individuals when stimulating left or right ear (time×group F _(4,88)  = 24, p < 0.001), whereas all the MCS patients and the two abovementioned UWS individuals showed a clear ipsilateral and a mild contralateral ASRt decrease (time×group F _(4,88)  = 11, p < 0.001). Of note, such decreases were further mild when testing ASRt at extreme TPPs and 4 kHz. Indeed, the most affected frequencies were 1 kHz (time×group F _(4,88)  = 28, p < 0.001) and 0.5 kHz (time×group F _(4,88)  = 21, p < 0.001), whereas 2 kHz showed minor but still significant changes (time×group F _(4,88)  = 4.5, p = 0.02). Concerning TPP effects, ASRt changed when employing TPP⁺¹⁰⁰ (time×group F _(4,88)  = 25, p < 0.001), TPP⁺⁵⁰ (time×group F _(4,88)  = 8.9, p = 0.002), TPP⁰ (time×group F _(4,88)  = 20, p < 0.001), TPP^–50 (time×group F _(4,88)  = 21, p < 0.001), and TPP^–100 (time×group F _(4,88)  = 43, p < 0.001). Notably, all the frequency changes were more evident in the HC than the MCS and the two UWS individuals, but rTMS did not modify the frequency and TPP distribution, i.e. each frequency and TPP was able to elicit an ASRt (time×frequency×side×TPP×group interaction F _(72,1584)  = 1, p = 0.4). Finally, we found a correlation between high 1 kHz-ASRt values and low auditory_CRS-R scores at T_pre (r = –0.851, p = 0.04) and between 1 kHz-ASRt value decrease and auditory_CRS-R value increase at T₁₀ (r = –0.840, p = 0.04).

4.1 DOC differential diagnosis

Our data further confirm the role of thalamocortical system damage in determining the exaggerated and maladaptive reflex activity in DOC individuals, and propose ASR assessment as a potentially useful tool in differentiating MCS from UWS individuals. In fact, at baseline, DOC individuals showed a disorganized ASRt distribution within the four frequencies and all the TPP employed, and a significant deterioration of crossed ASR, which were both paralleled by a cortical-thalamocortical network impairment, as suggested by the low CRS-R scores. Hence, DOC patients may be poorly- or non-responsive to the auditory stimulation in reason of a wide connectivity breakdown (Boly et al., 2011, 2012), which determines the ASR over-strengthening that, in turn, contributes to limit motor responsiveness to acoustic stimuli.

Moreover, it is worthy to note that rTMS protocol specifically ameliorated auditory_CRS-R scoring. This finding integrates and further supports a previous work of our group demonstrating a re-modulation of audio-motor connectivity following rTMS conditioning protocol over the primary auditory area, thus outlining the possibility of identifying residual cortical networks sustaining audio-motor integration (Naro et al., 2015). Therefore, ASR modulation may be a marker of the entrainment by part of rTMS of a vast network probably encompassing other second-order auditory areas and subcortical structures (including thalamus), which account for higher audio-motor behaviors (e.g. sound localization) (Friederici, 2011). Thus, this network entrainment, paralleled by ASR re-modulation, could be considered as a marker of the preservation of the plastic properties of this network, and it could be used as a way to discriminate MCS patients (who showed residual cortico-subcortical networks), from UWS ones (who lacked such networks).

One could argue that the missing typical slope of ASRt may be a cause for concern. In fact, this lack may be owed to the low efficiency of our method in inducing ASRt changes at baseline in both patient groups. Nonetheless, MCS patients showed a deteriorated ASRt slope as compared to HC subjects, whereas only UWS individuals had a significantly abnormal ASRt slope. Further, electrophysiological data showed a low variance either at baseline (T_pre) or T₁₀, which was superimposable to HC individuals; this would support the robustness of ASRt testing even in DOC patients. Moreover, sham rTMS was ineffective in all the groups. Therefore, we believe that either ASRt testing or rTMS did not work in the UWS sample owing to the deterioration of cortico-brainstem connectivity overseeing the ASRt regulation, rather than the inefficiency of our method.

Even though rTMS after-effects on ASRt further confirmed the differences between MCS and UWS patients observed at baseline, we identified two clinically defined UWS patients showing an MCS-like electrophysiological and behavioral profile after real-rTMS application. Since we were able to include a limited number of DOC patients in this study, and we found MCS-like patterns in only two UWS individuals, we have to be cautious in reaching any conclusion concerning DOC diagnosis of these two subjects. Nonetheless, we may argue that in these patients a higher degree of cortico-subcortical connectivity, which should at least question their UWS diagnosis, could be supposed.

4.2 ASR physiology

Ear sounds activate the ipsilateral (uncrossed) ASR by stimulating a brainstem subpopulation of stapedius motoneurons (SMNs) that contract the stapedius muscle of the same ear. To this end, SMNs receive inputs from the ipsilateral cochlear nucleus (CN). It is not clear whether CN directly projects to the SMNs. The latter receive also inputs from the contralateral CN, thus mediating the contralateral ASR, but the neural pathway mediating such response is not well characterized. Noteworthy, SMNs probably receive other modulatory inputs coming from motor cortex (since some individuals can voluntarily contract stapedius muscle), superior olivary complex, locus coeruleus, and inferior colliculus (Rouiller et al., 1989; Mukerji et al., 2010). Our work electrophysiologically demonstrates for the first time a direct role of auditory cortex in modulating ASR. Since the simple procedure we employed, we can only hypothesize a putative model of the cortico-subcortical network, encompassing primary auditory area and the centers for ASR in the brainstem, which can bilaterally modulate ASR in a tonotopic manner. In fact, the 4 kHz band was not affected by rTMS. It is well known that the stapes transmits vibrations to the oval window on the outside of the cochlea, which vibrates the perilymph in the scala vestibuli. The ossicles are essential for efficient coupling and gaining of sound waves into the cochlea. This gain is a form of impedance matching (i.e. matching the sound waves traveling through the air into those through the fluid–membrane system). The Corti hair cells are tuned to certain sound frequencies because of their location in the cochlea, the degree of stiffness of the basilar membrane, and the frequency-specific ASRt (Borg, 1973; Ehret, 1978; Camhi, 1984). Thus, high frequencies do not propagate up to the helicotrema in reason of the stiffness-mediated tonotopy, whereas lower frequencies propagate along the complete route of the cochlea. We may hypothesize that rTMS-induced 4 kHz-ASRt modulation could not have involved high frequencies since these may be more dangerous than lower ones at a similar sound intensity (dB-SPL). In addition, we cannot exclude that: i) specific property of the magnetic field may have favored a stronger triggering of 0.5 up to 2 kHz; and ii) these frequencies may be more accessible to magnetic fields since their tonotopic displacement over primary auditory cortex surface (Sowman et al., 2014; Schecklmann et al., 2011). These issues would also explain why rTMS modulated only the ASRt values of the frequency tested at the different TPPs, and not the distribution of frequencies and TPPs.