Investigation of Listening Effort in Tinnitus Patients by Providing Similar Peripheral Auditory Function With Control Group

The study addresses whether tinnitus independently increases listening effort when both hearing thresholds and peripheral auditory function (PAF) are matched to healthy controls. Tinnitus affects daily life and cognitive domains such as attention and memory, potentially impacting listening effort. Prior work suggested increased listening effort in tinnitus using subjective and physiological measures (e.g., pupillometry), but differences in autonomic nervous system activity and insufficient control of PAF may confound results. Given evidence that tinnitus patients can exhibit peripheral auditory dysfunction (PAD) despite normal audiograms, the authors aim to assess listening effort using EEG alpha-band power while ensuring similar hearing (0.125–20 kHz) and similar PAF (via ABR and speech perception in noise) between tinnitus and control groups.

Listening effort is defined as cognitive resource allocation to maintain a listening task and is evaluated via subjective, behavioral, and physiological methods (EEG, pupillometry). EEG alpha-band activity reflects central neural resource allocation and inhibitory mechanisms; pupillometry reflects autonomic responses and may be unreliable in tinnitus due to altered ANS activity. Earlier studies controlled hearing via pure tone thresholds but did not fully assess PAF; recent work shows PAD (e.g., cochlear synaptopathy) in tinnitus despite normal thresholds, affecting ABR and speech-in-noise performance. PAD may also relate to cognitive difficulties and listening problems even with normal audiograms. Therefore, controlling PAF using ABR and SPIN is important to isolate tinnitus effects on listening effort.

Ethics approval (GO22/1096) and written informed consent were obtained. Participants: 42 native Turkish speakers (undergraduate/graduate education) were recruited; after excluding three due to EEG artifacts, 16 tinnitus (8 males, 8 females; 19–34 years) and 23 controls (10 males, 13 females; 19–30 years). Inclusion: normal hearing thresholds 0.125–20 kHz (<20 dB HL) for both groups; tinnitus group had chronic tinnitus >6 months. Exclusion: pathologies related to tinnitus etiology (e.g., otitis, demyelinating diseases, cervical problems), prior tinnitus treatment, acute tinnitus (<6 months), prolonged noise exposure, antidepressant use. Assessments: - Otoscopy and tympanometry confirmed normal outer/middle ear function. - Pure-tone audiometry (Interacoustics AC-40; TDH-39P; Sennheiser HDA200 for 8–20 kHz; Radioear B-71 for bone). - MoCA (score ≥21 considered normal; all participants ≥21). - VAS for subjective listening effort frequency (0–10). - Tinnitus assessment: pitch/loudness matching (0.125–20 kHz; two-alternative forced choice at 30 dB SL; loudness in 5 dB steps), minimum masking level (MML) using narrowband noise centered at tinnitus frequency (binaural), residual inhibition (RI) assessed by presenting narrowband noise 10 dB above MML for 60 s; RI categorized as positive (+) if tinnitus perception decreased, negative (–) if not. - THI (25 items; Yes=4, Sometimes=2, No=0) for tinnitus handicap. - ABR: Vivosonic Integrity; EAR-TONE 3A inserts; two-channel montage (forehead non-inverting, low forehead ground, mastoids inverting); impedances ≤5 kΩ (electrodes) and ≤2 kΩ (inter-electrode). Clicks: alternating polarity, 80 dB nHL, 21.1 clicks/s, 2000 sweeps, 100 µs duration; bandpass 100–1500 Hz; analysis window 15 ms; both ears tested. Calculated amplitudes and absolute latencies for Waves I–V; interpeak amplitude ratios III/I, V/III, V/I. - Speech reception thresholds: Matrix Test (MT) in noise (multi-talker babble, 65 dB SPL noise; adaptive speech level in 1 dB steps; open-set; 30 lists × 20 sentences; syntactically correct, semantically anomalous; both ears tested). Outcome: SNR at 50% word correct (50% SRT). - EEG recording: 21-channel NuAmps II Neuroscan; 19 scalp electrodes (10–20), earlobe references; common average reference with A2 during analysis; impedances <5 kΩ; continuous data filtered 1–60 Hz, notch at 50 Hz; ocular artifact reduction via EEGlab v14.1.2; alpha band extracted with 8–12 Hz bandpass; Hilbert transform for envelopes. Stimuli: 6-channel noise-vocoded sentences presented at each participant’s 6ch 50% SRT level; total stimulus 9 min; participants repeated sentences after noise offset; attention monitored via camera and recorder. Noise-vocoding: sentences divided into 16 logarithmic bands; envelopes via Hilbert transform to modulate noise; multi-talker babble combined; RMS equalized; stimulus temporal structure: noise 0–1 s, vocoded speech 1–4.5 s, noise 4.5–6 s; binaural presentation. Alpha analyses focused on parietal electrodes (P3, P4, Pz). Encoding window: 200 ms before end of vocoded speech (3.3–4.3 s, 1 s interval); baseline: noise (300–800 ms after noise onset). Relative percent change: (mean alpha during encoding − mean alpha during baseline) / mean baseline × 100. - Statistics: G*Power estimated 14 per group for 95% power and 5% type I error based on pilot means/SDs. Normality confirmed. Independent samples t-tests for age, MoCA, EEG alpha, ABR; Levene’s test checked homogeneity; Pearson correlation between EEG alpha change and THI; Fisher exact test for gender. Significance threshold p=0.05.

- Hearing thresholds 0.125–20 kHz: no significant group differences (p>0.05). - MoCA: all >21; no group difference (Table 1: tinnitus 29±1, control 29±1; p=0.93). - EEG baseline alpha power: tinnitus 4.16±2.68 μV vs control 5.03±3.89 μV; p=0.41. - EEG alpha power change (percent relative to baseline, 6ch 50% SRT): tinnitus mean 105.88%±128.64 vs control 225.30%±66.88; p=0.003 (independent t-test), indicating smaller increase in tinnitus group. - THI correlation with EEG alpha change: not significant (p=0.71). - Residual inhibition subgroup comparison for EEG alpha change not performed due to uneven RI+ vs RI– distribution. - ABR (32 ears tinnitus; 46 ears control): No significant differences in amplitudes or absolute latencies across waves I–V or ratios III/I and V/I (all p>0.05). Example values: Wave I amplitude 0.15±0.03 μV (tinnitus) vs 0.17±0.08 μV (control), p=0.18; Wave V amplitude 0.53±0.10 vs 0.49±0.09 μV, p=0.23; Wave I latency 1.50±0.24 ms vs 1.56±0.36 ms, p=0.55; Wave V latency 5.66±0.83 ms vs 5.55±0.63 ms, p=0.65. - MT (50% SRT): tinnitus 1.63±0.37 dB SNR vs control 1.56±0.38 dB SNR; p=0.62. - VAS listening effort: tinnitus 4.93±2.25 (range 1–8) vs control 2.04±1.58 (range 0–5); p<0.01 (reported as p=0.003 in text). Overall, tinnitus participants reported higher subjective effort and showed reduced EEG alpha-band power increase despite similar hearing and PAF metrics.

Ensuring comparable PAF via ABR and MT allowed isolation of tinnitus effects on listening effort. The smaller increase in EEG alpha-band power during encoding in tinnitus suggests greater allocation of neural resources and reduced cortical inhibition associated with effortful listening. Despite normal speech-in-noise performance (MT) and ABR measures, tinnitus patients demonstrated higher subjective effort (VAS) and physiological indicators of increased effort (EEG), highlighting that SPIN tests may not capture effort differences. Attention and anxiety were considered as potential confounders; lack of correlation between THI and EEG alpha changes and comparable MoCA scores suggest these factors are less likely primary drivers in this cohort. The findings support a central mechanism for increased listening effort in tinnitus, potentially related to cortical inhibitory dynamics, rather than peripheral dysfunction, given matched PAF.

This study is the first to assess listening effort via EEG alpha-band power while controlling hearing thresholds (0.125–20 kHz) and ensuring PAF similarity between tinnitus and control groups. Tinnitus participants showed a lower increase in alpha power during speech encoding and higher subjective listening effort, indicating greater effort despite comparable PAF. Clinicians should consider cognitive and listening-related outcomes in tinnitus management. Future research should further explore central mechanisms underlying tinnitus-related listening effort and incorporate methods to directly assess causal factors.

- Single-center recruitment with limited timeframe may restrict participant diversity and introduce recruitment bias. - PAF similarity was inferred using ABR and MT; definitive demonstration of PAD requires histopathology, which is not feasible, so interpretations should be cautious. - VAS is not a standardized measure for listening effort. - Relatively small sample size (16 tinnitus, 23 controls). - Potential individual differences (age, education, culture, vocabulary) could influence effort or fatigue despite stimuli chosen to avoid fatigue. - RI subgroup analysis was not feasible due to unequal RI+ and RI– distribution. - Although hearing thresholds were matched across 0.125–20 kHz, subtle peripheral differences may persist beyond current measurement sensitivity.