Abstract
Importance
Evidence regarding the mechanism of tinnitus in patients with normal full-frequency hearing remains limited. This study investigated the auditory brainstem response (ABR) waveform patterns in these patients to address this unmet need, which is of considerable assistance in the clinical management of tinnitus with normal hearing.
Objective
This study aims to investigate the potential underlying mechanisms of tinnitus with normal full-frequency (125 Hz-16 kHz) hearing threshold by analyzing waveform alterations in ABR among tinnitus patients.
Design
Cross-sectional study.
Setting
Otolaryngology outpatient of a tertiary referral hospital.
Participants
Patients aged 20 to 40 years with unilateral subjective tinnitus and normal hearing (pure-tone thresholds ≤25 dB HL across 0.125-16 kHz) were recruited. Age-matched healthy volunteers with normal full-frequency hearing (0.125-16 kHz) and without tinnitus served as controls.
Exposures
Participants were categorized according to the presence of unilateral subjective tinnitus.
Main Outcome Measures
The primary outcomes included pure-tone audiometry testing (0.125-16 kHz) and ABR testing using 80 dB nHL click stimuli with alternating polarity (19.9 Hz, 1024 scans).
Results
The study included 136 adult subjects with normal hearing at all frequencies, including 60 unilateral tinnitus patients and 76 normal volunteers. Results showed no significant differences in ABR waveform characteristics between affected and unaffected ears in unilateral tinnitus patients ((P > .05). However, compared with normal volunteers, unilateral tinnitus patients exhibited significantly prolonged latencies of waves I, III, and V on the tinnitus side (all P < .05), with median differences (95% CI) of 0.08 (0.04-0.11), 0.07 (0.00-0.12), and 0.15 (0.09-0.20), respectively. In addition, the interpeak latency between waves III and V was also prolonged, with a median difference of 0.08 (95% CI: 0.03-0.13).
Conclusion
Tinnitus patients with normal full-frequency hearing exhibit differences in the ABR waves, which may be associated with abnormal middle ear muscle activity. Emotional responses may trigger abnormal middle ear muscle activity via the serotonin system, which in turn affects auditory signal transmission through the trigeminal system, leading to tinnitus and abnormal ABR wave patterns.
Relevance
Future research will focus on the middle ear muscles as a potential breakthrough point, exploring the relationship between abnormal electrical activity in these muscles and tinnitus, as well as the electrical activity of the trigeminal nerve nuclei.
Level of Evidence
3.
Keywords
Key Message
Tinnitus can occur in individuals with normal full-frequency hearing and may be linked to abnormal activity of the middle ear muscles.
Emotional states can modulate middle ear muscle activity through serotonergic pathways, thereby influencing auditory signal transmission via the trigeminal nerve system.
Introduction
Tinnitus is defined as the perception of meaningless sounds in the absence of external sound sources, and its pathogenesis remains incompletely understood. 1 Based on current research findings, tinnitus perception appears to represent an abnormal plasticity of the cortical auditory system compensating for reduced peripheral auditory input. Less commonly, in approximately 10% to 30% of tinnitus cases, audiometric thresholds fall within the normal range (audiometric thresholds ≤25 dB hearing level [HL] from 125 Hz to 8 kHz).2-4 Our study also found that when expanding the hearing test range to 125 Hz to 16 kHz, 96.6% of tinnitus patients exhibited hearing loss. 5 However, a very small number of tinnitus patients with normal hearing across all frequencies remain unexplained by existing hypotheses. Recent studies indicate that tinnitus in individuals with normal hearing may involve brain regions outside the auditory system, including the limbic system, parietal lobe, and prefrontal cortex, which participate in tinnitus maintenance processes.6,7 Given the clinical existence of middle ear myogenic tinnitus, literature reviews indicating projections between the middle ear muscle nuclei and auditory pathways, we propose that tinnitus in individuals with normal hearing may be associated with abnormal middle ear muscle activity. The trigeminal nucleus innervating the tensor tympani muscle projects to auditory brainstem regions including the inferior colliculus (IC) and cochlear nucleus (CN). Stimulation of the trigeminal nerve inhibits the spontaneous discharge rate (SR) of IC and CN. 8 Moreover, the serotonin (5-hydroxytryptamine, 5-HT) is associated with mood disorders such as depression and anxiety,9,10 and middle ear muscle function is affected by the serotoninergic system, which is consistent with the link between emotional state and middle ear muscle contraction. 11 Emotional factors, such as anxiety and fear, can activate the trigeminal nucleus by affecting the tension in the middle ear muscles,12,13 which may in turn affect signaling in the auditory pathway. Therefore, we hypothesize that in tinnitus patients with normal full-frequency hearing thresholds (125 Hz-16 kHz), abnormal activity in trigeminal nuclei related to middle ear muscles may interfere with auditory pathway conduction, thereby inducing tinnitus.
Given the unclear mechanisms of tinnitus and the complexity of research, our clinical examinations (audiometry, acoustic immittance, and acoustic reflex testing) cannot detect tinnitus related to middle ear muscles. Furthermore, endoscopic examination rarely reveals abnormal tympanic membrane activity. Therefore, developing new diagnostic techniques is our primary objective. Auditory brainstem response (ABR) serves as an objective electrophysiological indicator for assessing brainstem auditory pathway function, 14 in which wave I originates from the distal auditory nerve, wave II originates from the proximal part of this nerve, 15 while wave III reflects activity in the cochlear nucleus and superior olivary complex. Waves IV and V reflect activity in the lateral colliculus and inferior colliculus.15-18 Abnormal electrical activity in trigeminal nuclei may disrupt signal transmission along the auditory pathway and influence ABR waveform generation. This study therefore aims to explore the pathophysiology of tinnitus with normal hearing by analyzing ABR waveform differences in patients with tinnitus who exhibit normal hearing across the full-frequency range (125 Hz-16 kHz).
Materials and Methods
Participants and Grouping
Tinnitus patients were enrolled from December 2023 to October 2025 at the Otolaryngology outpatient of Huadong Hospital affiliated with Fudan University. Group A included 60 subjects aged 20 to 40 years (mean 30.93 ± 5.59 years) with unilateral chronic tinnitus (≥6 months) and normal full-frequency thresholds (125 Hz-16 kHz, ≤25 dB HL). Among them, 32 (53.3%) had right ear tinnitus, 28 (46.7%) had left ear tinnitus, with 28 females and 32 males. We designated the tinnitus ears (TE) as Group A1 and the nontinnitus ears (NTE) as Group A2. Group B, the control group, consisted of 76 healthy nontinnitus individuals, aged 20 to 40 years (mean 30.18 ± 5.95 years), with 41 females and 35 males. Their hearing thresholds (125 Hz-16 kHz) were normal (≤25 dB HL). Inclusion criteria were as follows: (1) Unilateral persistent tinnitus greater than 6 months; (2) Hearing level: air conduction thresholds (125-16,000 Hz) ≤25 dB HL; (3) Distortion Product Otoacoustic Emissions (DPOAE) results from 500 to 8000 Hz are normal and can be recorded (with a signal-to-noise ratio of each frequency (s/n-ratio) >6 dB); (4) The tympanometric curve of acoustic immittance is type A; (5) The ABR test: an 80 dB nHL stimulus sound can clearly elicit I, III, and V waves, and the ABR elicitation threshold for both left and right ears is ≤35 dB nHL. Exclusion criteria were as follows: (1) Otogenic tinnitus, which refers to tinnitus caused by organic lesions in the outer, middle, or inner ears; (2) Objective tinnitus; (3) Tinnitus caused by acute or chronic diseases; (4) Patients with incomplete medical records and examinations.
Auditory Brainstem Response
The 580-NAVPR2 ABR device and the 580-SINSER insert earphones (Biologic, Natus Medical Incorporated, Middleton, WI, USA), both with calibration, were utilized. Referring to the 10 to 20 standard system of the International Federation of EEG Societies, the active electrode was placed on the forehead (Fz), the reference electrode on the ipsilateral mastoid (Mi) of the tested ear, and the ground electrode on the contralateral mastoid (Mc). The response was recorded by the potential difference between the Fz and Mi electrodes for ipsilateral acquisition. The ABR was elicited with an alternating polarity. The stimuli included 100 μs clicks and 3 tone bursts (1000, 4000, and 8000 Hz). The clicks were presented at 80 dB nHL level, with an repetition rate of 19.9 Hz and a superposition of 1024 times. Signal acquisition was conducted in a monaural mode using a single channel, with an amplification of 100,000. The filter bandwidth used for recording was 100 to 3000 Hz. Any waveform with amplitude greater than 23.80 μV was rejected as artifact. In case of excessive artifact occurrences, which were usually caused by the subject’s movement, the subject was repositioned or given a break.
Subjects were asked to sleep to minimize EEG variability. The recording sites were cleaned using alcohol and conductive gel. Two waveforms were obtained from each ear and then assessed through averaging. The recording parameters were marked manually by visual inspection by an experienced audiologist. The ABR examiner was blinded to the participants’ clinical status (ie, with or without tinnitus) throughout the data acquisition and analysis process to prevent potential bias.
Pure-Tone Hearing Threshold
Pure tone (PT) audiometry and extended high-frequency (EHF) audiometry tests were performed using the calibrated audiometer Astera 1066 Clinical Audiometer device (Otometrics) with ME70 and DD450 supra-aural earphones. In addition, the bone conduction thresholds were measured by a Radioear B-71 bone conduction vibrator (RadioEar Co., Middelfart, Denmark). The PT hearing thresholds were measured between 125 to 8000 Hz, and the EHF audiometry test was recorded at 10, 12.5, and 16 kHz. The tests were conducted by the standard clinical method (modified Hughson–Westlake 19 ), manually increasing the signal level by 5 dB until the response was given by the subject and then reducing it by 10 dB and increasing again by 5 dB until the response of the subject. All patients underwent examination and matching by a qualified and trained audiologist within a soundproof room.
Tinnitus Matching
Tinnitus matching was conducted using Tinnifit SFTest 330 (BOZYTM Medical Technology, Foshan, Guangdong Province, China). Participants performed a frequency matching based on a recursive 2AFC test.20,21 The main tone frequency was matched with pure tone between 0.125 and 16 kHz. Play a test tone at 5 to 10 dBSL above the hearing threshold with a duration of 1.5 s and ask the patient to compare whether the tinnitus sound is consistent with the test tone. The frequency of tinnitus was refined to 1/24 octave. Loudness matching was performed using tones of the same frequency as the tinnitus tone, with results accurate to 1 dB. Also, tinnitus sound type (pure tone, narrow-band noise, white noise) are involved. In this study, we defined 125 to 1000 Hz as low frequency (LF), 2000 to 4000 Hz as middle frequency (MF), 4000 to 8000 Hz as high frequency (HF), and above 8000 Hz as EHF. 5 Acoustic immittance detection and Distortion product otoacoustic emissions
E: Frequency ratio FL/F2 = 1.22, pure tone intensities: L1 = 65 dB SPL, L2 = 55 dB SPL. Each frequency point is superimposed 16 times with a 90 s recording time. The amplitude and SNR of F0 (2f1−f2) at 0.5, 1, 2, 4, 6, and 8 kHz are recorded, and the computer draws the DPOAE diagram. The leading-out standard of DPOAE is the judgment standard: S/N - ratio of each frequency >6 dB.
Acoustic impedance audiometer (AT235h, Danish International Hearing/Acoustics).
Statistical Methods
Statistical analysis was performed using IBM SPSS Statistics for Windows, version 22.0 (IBM Corp., Armonk, NY, USA).with analysis of variance. Results were expressed as mean ± standard error. Data were analyzed using chi-square test, paired sample t–test, Wilcoxon paired rank sum test, independent samples t-test, Mann–Whitney U test. P < .05 was considered statistically significant.
Results
Basic Information
Patient characteristics
Group A consisted of patients with unilateral tinnitus and normal hearing. The TE was classified as Group A1, and the contralateral NTE was classified as Group A2. Group B comprised normal-hearing individuals without tinnitus. As shown in Table 1, Group A consisted of 60 cases (32 male, 28 female, mean age 30.93 ± 5.59), while Group B consisted of 76 cases (35 male, 42 female, mean age 30.18 ± 5.95). In order to reduce the differences in ABR results caused by age differences, the age of the enrolled patients in both groups ranges from 20 to 40. There was no significant difference in gender and age between the 2 groups.
Basic Information of Each Group.
Sex was analyzed using the chi-square test, and age was analyzed using the Mann–Whitney U test.
Characteristics of tinnitus pitch distribution and pure-tone audiometry threshold
In our study (Table 2), LF accounted for 26.7% of the total tinnitus cases, MF 20%, HF 25%, and EHF 28.3%. There was no significant difference in the distribution of tinnitus cases in each frequency band (Figure 1).
Distribution Characteristics of Tinnitus Pitch (n = 60).
Chi-square goodness-of-fit test (expected equal proportions across LF/MF/HF/EHF) showed no significant deviation from the expected distribution (P = .817).
Abbreviations: EHF, extended high frequency; HF, high frequency; LF, low frequency; MF, middle frequency.
Tinnitus pitch distribution in Group A1. LF, low frequency; MF, middle frequency; HF, high frequency; EHF, extended high frequency.
Figure 2 shows that the average pure-tone audiometry (PTA) values of the 3 groups of patients at each frequency are all ≤25 dB, and there were no statistically significant differences among them (Figure 3) (Group A1 vs Group A2: P = .502; Group A1 vs Group B: P = .458).
PTA threshold for each frequency of Group A1, Group B, and Group A2. PTA, pure-tone audiometry.
Average PTA thresholds for each patient in Group A1, Group B, and Group A2. PTA, pure-tone audiometry.
Comparison of ABR Waveforms Between Two Ears in Patients With Unilateral Tinnitus
As shown in Table 3, we compared the TE of unilateral tinnitus patients (Group A1) with that of NTE (Group A2) and found that there was no statistical difference between them in the amplitude and latencies of Ⅰ, Ⅲ, Ⅴ waves and the interpeak latency (IPL) of Ⅰ-Ⅲ, Ⅰ-Ⅴ, Ⅲ-Ⅴ waves (P > .05).
ABR Waveforms Between Group A1 and Group A2.
Data are presented as mean ± SD. Comparisons between groups were performed using the Mann–Whitney U test.
Comparison of ABR Waveforms Between Tinnitus Patients and Nontinnitus Patients
As shown in Table 4, there was no statistical difference between TE of unilateral tinnitus patients (Group A1) and Group B in the amplitude of Ⅰ, Ⅲ, and Ⅴ waves, IPL of Ⅰ-Ⅲ and Ⅰ-Ⅴ waves (P > .05), but there was a significant difference between the two groups in the latencies of Ⅰ, Ⅲ, Ⅴ waves and the IPL of Ⅲ-Ⅴ waves (P < .05).
ABR Waveforms Between Group A1 and Group B.
Data are presented as mean ± SD. Comparisons between groups were performed using the Mann–Whitney U test. Bold values indicate statistically significant results. P < 0.05 was considered statistically significant.
P values are denoted as follows: **P < .05, ***P < .001.
Discussion
The mechanisms underlying tinnitus remain unclear. Eggermont proposed 3 major mechanisms: increased spontaneous firing rate of neurons, neural synchronization, and central remodeling following reduced peripheral auditory input. 22 However, the mechanism of tinnitus in patients with normal full-frequency hearing thresholds remains incompletely elucidated. By reviewing the projection between the trigeminal nucleus and auditory pathway nuclei, we propose the hypothesis that “Tinnitus in individuals with normal hearing may be associated with abnormal electrical discharges of the trigeminal nerve.” Trigeminal nerve nuclei include the sensory, mesencephalic, and main motor nuclei. The sensory nucleus receives afferent axons from the face, oral and nasal cavities, transmitting tactile and pressure sensations. The mesencephalic nucleus receives proprioceptive responses from extraocular muscles, mastication muscles, joints, teeth, and periodontal tissues. 23 The motor nucleus is responsible for innervating muscles including palatal sail tensor muscle, palatal tensor muscle, tensor tympani muscle, etc.24,25 The trigeminal nucleus can project to the auditory brainstem region, such as the inferior colliculus and cochlear nucleus. Haenggeli’s retrograde and anterograde fluorescent tracing experiment revealed that anterograde projections from the trigeminal nuclei primarily target the granular cell domain of the cochlear nucleus and the deep layers of the dorsal cochlear nucleus, with minor projections in the inferior colliculus. 26 This confirms the synaptic connections between the trigeminal nucleus and both the cochlear nucleus and the inferior colliculus. Through electrical stimulation of the trigeminal ganglion in guinea pigs, Shore found that approximately 30% of cochlear dorsal nucleus units responded to ganglion stimulation with excitatory or inhibitory reactions even in the absence of auditory input. 27 The projection of the trigeminal nerve’s nucleus establishes a structural basis for the middle ear myogenic tinnitus. For instance, the tympanic membrane tension reflex is a startle reflex that intensifies under high stress in which the serotonin system related to emotional problems plays an important role. 12 Researchers employing viral tracing techniques have discovered a direct anatomical connection between the terminals of serotonergic neurons and the motor neurons controlling the muscles of the middle ear. This finding suggests that serotonergic neurons may influence the activity of middle ear muscles via this pathway. 11 In addition to this theory, there is evidence that serotonin may also modulate nonreflex contractions of the middle-ear muscles. 28 This tension reflex might affect signal transmission within the trigeminal nucleus, thereby influencing the auditory pathway. Previous studies have also found that somatic tinnitus is closely related to the trigeminal nucleus. Tinnitus patients can modulate the loudness or pitch of their tinnitus by performing certain head and neck somatic movements or altering their posture. For instance, when patients clench their teeth, move their chin or neck, the symptoms of tinnitus may change, potentially due to the regulatory function of the trigeminal nerve on the dorsal cochlear nucleus. 29 Given the clinical existence of middle ear myogenic tinnitus, we propose that abnormal trigeminal nerve discharges arise from middle ear muscle activity may prolong nerve conduction in the inferior colliculus or cochlear nucleus, thereby inducing tinnitus.
Previous studies have mostly suggested that middle ear myogenic tinnitus is generally low-frequency and often presents as a rhythmic click. 30 However, as shown in Figure 1, the frequency distribution of tinnitus in patients with normal hearing is relatively scattered, dominated by EHF tinnitus, followed by LF tinnitus. This may be due to irregularities in the middle ear muscles triggering irregular abnormal discharges in the trigeminal nerve, which in turn cause varying degrees of interference to the auditory pathway. Therefore, we propose that if there is a relationship between normal hearing tinnitus and abnormal middle ear muscle activity, then the frequency of tinnitus caused by abnormal middle ear muscle activity is not only at low frequency but also at various frequencies.
Clinically, ABR is primarily used to assess signal conduction in the auditory pathway. The sound signals measured by ABR represent activation of the neural pathway between the cochlear nerve and the inferior colliculus. 31 Although there remains some controversy regarding the correspondence between the generation of waves III and V and their precise neuroanatomical origins, the prevailing view holds that IPL of waves III-V is typically attributed to prolonged neural conduction time from the cochlear nucleus or superior olivary complex to the inferior colliculus. 32 Waves IV and V reflect activity in the lateral lemniscus and inferior colliculus.16-18 Previous studies on tinnitus and ABR waveforms have primarily focused on hearing thresholds within 250 Hz to 8 kHz range. Little attention has been paid to EHF pure tones. EHF hearing loss refers to one or more thresholds at extended high frequencies (10-16 kHz) >25 dB HL, 33 which is potentially associated with cochlear synaptic lesions. 34 However, the mechanism by which individuals with normal hearing tinnitus across all frequencies, including EHF, remains unclear. This study controlled for normal hearing across the entire frequency range (125 Hz-16 kHz, as shown in Figure 2) to exclude the potential mechanism of “hidden hearing loss,” ensuring normal conduction and processing functions of the auditory pathway for high-frequency stimuli. This approach enabled exploration of possible mechanisms for tinnitus in patients with normal full-frequency hearing. Results revealed statistically significant prolongations in the latency of ABR waves I, III, and V, as well as in the IPL between waves III-V in tinnitus patients compared to nontinnitus subjects. This suggests impaired neural conduction in the cochlear nucleus or inferior colliculus, leading to altered ABR waveforms. In clinical practice, we observed that tinnitus patients with completely normal hearing thresholds often exhibit emotional abnormalities. We hypothesize that emotions may influence middle ear muscle activity through the serotonergic system, leading to abnormal trigeminal nerve discharges. This, in turn, affects neural transmission in the cochlear nucleus or inferior colliculus, ultimately resulting in abnormal ABR waveforms. Therefore, these tinnitus patients with normal hearing will exhibit alterations in their ABR waveforms, particularly the prolongation of the latency of ABR waves III and V, as well as the IPL between waves III-V.
This study also found that in patients with unilateral tinnitus and normal hearing, there were no significant differences in ABR waveforms between the affected and unaffected ears. Theoretical evidence supports this finding. Previous research indicated that 3 primary relay nuclei exist between the auditory nerve and the primary auditory cortex—the cochlear nucleus, inferior colliculus, and medial geniculate body. Fibers along this primary ascending pathway can branch to contralateral neural nuclei, with fiber bundles conducting bilaterally at multiple levels. This allows auditory signals from each ear to project to the contralateral temporal cortex. 35 Furthermore, studies indicate that the medial olivocochlear system (MOC), which is distributed in the medial olivary nucleus of the brainstem, regulates spontaneous activity coupling in bilateral auditory pathways through cholinergic modulation. MOC efferent fibers synapse directly onto outer hair cells, modulating their activity through acetylcholine release. Furthermore, their axons project bilaterally to both cochleae, which coordinates the auditory information input from the 2 ears. 36 Therefore, if abnormal activity of middle ear muscles can influence the electrical activity of the cochlear nucleus or inferior colliculus via the trigeminal nucleus, such abnormal activity may also be transmitted bilaterally, affecting neural conduction in both auditory pathways and consequently influencing bilateral ABR waveforms.
In our study, the latency of wave I in the tinnitus group was significantly prolonged compared to the nontinnitus group, consistent with the findings of Lemaire and Beutter. 37 They observed significantly prolonged latency of ABR waves I and III in the tinnitus group. In addition, they believed that this correlates with functional states in the cochlea and lower brainstem structures (such as the cochlear nucleus), which aligns with our inferences. Oliveira 38 postulated a fluid-homeostasis change, which led to tinnitus. He believed that latency prolongation reflects a slowing of synaptic processes in the organ of Corti or decreased neural conduction velocity in the first auditory neuron. Marchetta et al 39 investigated the effects of central mineralocorticoid and glucocorticoid receptor (MR and GR) deficiency on auditory nerve processing. By knocking out MR and GR in limbic structures such as the hippocampus, cortex, and amygdala, they discovered that the latency and amplitude of ABR I waves and compound action potentials are shortened. Thus, they concluded that the limbic system may impair auditory processing capabilities, responses, and synchronization through MR and GR, without affecting IHC ribbon numbers. The mechanism of tinnitus generation is still unclear. The influence of the trigeminal nerve related to the middle ear muscles might be one aspect, and various central networks may all be involved. Under conditions of emotional stress or psychological strain, the limbic system may influence the processing capacity and responsiveness of the auditory nerve, potentially triggering abnormal discharges or slowed conduction within the auditory nerve, thereby precipitating tinnitus.
Potential limitations of the present study include the fact that single-center design and limited sample size may reduce the external validity of our findings. Multicenter, larger-scale trials are needed to confirm the results. Moreover, despite multivariable adjustment, residual confounding (e.g., lifestyle, medication usage) cannot be ruled out.
The strength of our research lies in presenting an innovative hypothesis—abnormal activity in the middle ear muscles triggers abnormal electrical activity in the auditory cortex via the trigeminal nerve system, thereby causing tinnitus. In addition, we offered a novel hypothesis regarding the potential mechanisms underlying tinnitus in individuals with full-frequency normal hearing. In addition, we matched the age of the 2 groups (both aged 20-40), thereby eliminating the influence of age on ABR.
Conclusion
Upon expanding the normal auditory frequency range from 125 Hz to 16 kHz, we observed significant differences in the latency of ABR waves I, III, and V in the tinnitus group compared to subjects without tinnitus. Additionally, the IPL between waves III-V exhibited significant difference. We suspect that tinnitus in individuals with normal full-frequency hearing may be associated with abnormal middle ear muscle activity. Emotional responses may trigger abnormal middle ear muscle activity via the serotonin system, which consequently affects auditory signal transmission through the trigeminal system, leading to tinnitus and abnormal ABR wave patterns.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Key Disciplines of Huadong Hospital (Award number: ZDXK2214).
Ethical Approval
The study was approved by the Ethics Committee of Fudan University (2020K125).
Data Availability Statement
The datasets generated and analyzed during the current study are not publicly available due to ethical and privacy restrictions, but are available from the corresponding authors on reasonable request.