Enhancing Speech Intelligibility Using Transcranial Alternating Current Stimulation (tACS)

Background and Rationale In Europe, 71 million people - 16% of the population - suffer from hearing difficulties. One of the major consequences of hearing impairment is reduced speech-comprehension ability, especially in acoustically challenging conditions such as in reverberant or noisy environments.

The proposed project is intended to develop our understanding of the mechanisms that can be exploited to support speech comprehension under challenging listening conditions, and how they can be enhanced in hearing-impaired and older listeners.

To this end, the compensatory neural mechanisms of articulatory-motor integration will be tested by applying transcranial alternating current stimulation (tACS) over the vMC.

Objective(s) The overall objective of the study is to shed light on the compensatory mechanisms supporting speech comprehension under perceptually-challenging conditions.

Experiment 1 to 3 The primary objective is to test the utility of auditory-motor integration for degraded speech (DS) comprehension.

One secondary objective is to assess the interaction between degradation level of DS and tACS-induced improvement of performance. Another secondary objective is to explore potential differences of tACS-induced improvements in younger and older hearing-impaired listeners.

Experiment 4 The primary objective is to test the utility of cerebro-acoustic coherence for SIN comprehension. The secondary objective of Experiment 4 is to test whether cerebro-acoustic coherence reflects syllable-level processing.

Statistical Considerations:

For all experiments, mixed effect repeated measures ANOVAs will be conducted. Paired t-tests corrected for multiple comparisons will be applied for posthoc comparisons.

Number of Participants with Rationale

Power analyses were conducted separately for every experiment assuming an intermediate effect size. The result of the power analysis suggests the minimal total sample size for Experiments 1 to be 30, for Experiments 2+3 to be 50, and 60 for experiment 4.

Experiment 1 The result of the power analysis suggests a minimal total sample size of 30 younger participants.

Experiment 2-3 The result of the power analysis suggests a minimal total sample size 50, including 25 younger and 25 older participants; Experiment 4 The result of the power analysis suggests a minimal total sample size of 60, including 30 younger and 30 older participants in the experiment.

Handling of Missing Data and Dropouts Drop-outs as well as participants who have to be excluded due to sub-normal hearing thresholds as determined by pretest screening will be replaced by newly recruited participants. In case of missing data, the analyses are made with the data available. Data of drop-out will be analysed and anonymized upon termination of data analysis.

Study Intervention

The main study intervention is tACS stimulation. Two stimulation protocols will be used.

Stimulation Protocol 1 (SP1) used in Experiments 1-3 involves continuous high-frequency (2 kHz) stimulation. In experiment 1, SP1 will be applied online (during aMEP recording). In experiments 2-3, SP1 will be applied offline (before the start of the task) during 20 minutes.

Stimulation Protocol 2 (SP2) used in Experiment 4, involves online (during task execution) low-frequency speech-envelope (env) tACS stimulation (1-8Hz).

In all experiments, the stimulation intensity will be 2 mA (peak-to-peak value, corresponding to a sine wave of ±1 mA amplitude).

Stimulation will be delivered through conductive rubber electrodes placed over the ventral motor cortex (vMC). The electrodes will be placed according to the 10-20 International EEG System to identify the target location.

In addition, TMS stimulation will be applied in experiment 1 to measure articulatory motor-evoked potentials (aMEPs).

Stimulation intensity will be 120% of the individual motor threshold of the participants [3].

Monophasic single pulse TMS will be applied over the lip motor cortex in the vMC to induce aMEPs.

Study Procedures

Pre-screening

• assessment of the inclusion/exclusion criteria
• collection of demographic information

Screening of audition and speech perception

• Screening of the hearing threshold
• Speech reception threshold
• Text reception threshold

Subjective intelligibility level Prior to the experimental session of every experiment, the subjective intelligibility level will be assessed, in order to estimate the degradation levels that yield recognition performance of 75% (easy), 50% (medium) and 25% (difficult). Degradation will take the form of noise vocoding, which allows parametric manipulation of speech intelligibility. This is a manipulation that reduces the spectral content of speech, without affecting its temporal content. The outcome resembles a hoarse robotic whisper.

Experiment 1:

Each participant undergoes two tACS sessions: an experimental session with online tACS at 2kHz, and a control session, in which sham stimulation is delivered, each lasting 20 minutes (session order is counterbalanced). During the experiment, TMS-induced aMEPs will be recorded from the participants' lip muscles.

Experiment 2:

Each participant undergoes two tACS sessions: an experimental session in which 2kHz tACS is delivered and a control session in which sham stimulation is delivered, each lasting 10 minutes (session order is counterbalanced). In the 20-minute period following each tACS session, participants will hear speech stimuli degraded at various levels of intelligibility. They will be asked to transcribe the words they hear (using a computer keyboard) as quickly as possible, and with a maximum of 10s per item. They will be presented with 30 words at each of low, medium and high levels of degradation as well as 30 clear stimuli, in a random order.

Experiment 3:

Participants' baseline noise-vocoded speech comprehension (for a description of the procedure see section 3.2) will be assessed using the same testing procedure as in Experiment 1. In an experimental group, tACS with the same settings as described above (SP1) will be used to induce increased articulatory-motor excitability; in a control group, sham tACS will be delivered. In the following 20-minute period, participants will receive training on noise-vocoded speech. Training will consist of 250 words presented in pairs, first a clear instance of a word, then its degraded counterpart, 200ms after the offset of the clear word. Word pairs will be presented with a jittered inter-trial interval (ITI) of 1500±500ms. It has been shown that this type of presentation can be highly effective for rapidly improving noise-vocoded word comprehension: with the presentation of 120 items (taking around 10 minutes to present) participants' accuracy in word identification improved from approximately 20% to over 30% [4]. Noise-vocoded word comprehension will be assessed again after training.

Experiment 4:

Participants undergo two tACS sessions, in which they will be tested in DS comprehension whilst they receive either env-tACS or time-reversed env-tACS trials, interleaved by sham stimulation trials. As in Experiment 1, participants will be presented with 30 stimuli degraded to each of three different extents, and 30 clear stimuli and will be asked to transcribe the words they hear within 10 seconds.

Further Aspects Overall Ethical Considerations The study has the potential to carry both a scientific value, associated with improving the understanding of the compensatory mechanisms supporting degraded speech comprehension, and a social value, as its results might inform the development of new intervention strategies for overcoming hearing impairment or inform the amelioration of existing ones. The results of the study are generalizable to many (if not most of all) situations of listening under perceptually-challenging conditions, such as hearing impairment. Given that the greatest societal burdens of hearing impairment are associated with normal age-related hearing decline (i.e., presbycusis), we aim at extending the generalizability of the results to the third age by including, a sample of participants from an elderly population.

Good Clinical Practice (GCP) Statement This study will be conducted in compliance with the protocol, the current version of the Declaration of Helsinki, the International Conference on Harmonization-GCP, the Swiss Human Research Act as well as other locally relevant legal and regulatory requirements

Risk-Benefit Assessment There are no risks associated with the study interventions themselves, as long as exclusion criteria are respected, with the exception of rare mild and transient effects of tACS stimulation. Neither are risks of unauthorised data access through third parties or unwanted identification of participants expected.

The studies neither expects immediate benefits to the participant. However, participants (and anyone else) may benefit from the potentially strong scientific and social values of the study resulting from a better understanding of the (compensatory) neural mechanisms involved in challenging speech perception and medical devices (e.g., hearing aids) that may follow.

Quality Control and Data Protections Quality Measures In order to assure a good data quality, all the project members are trained on all the important study-related aspects. Personnel will be provided with theoretical and practical knowledge about the specific non-invasive electric brain stimulation device as well as experimental protocol used in the particular experiment in which he/she is involved, through careful consideration of the experimental protocol and performance of practice sessions. In order to ensure that the data are not lost or damaged, they will be stored dually. The investigator will perform regular data quality checks.

For quality assurance the sponsor, the Ethics Committee or an independent trial monitor may visit the research sites. Direct access to the source data and all study related files is granted on such occasions. All involved parties keep the participant data strictly confidential.

Data Recording and Source Data

An electronic case report form (e-CRF) will be held for:

• participants who were included* and successfully concluded the study
• participants who were included* in the study but had to be excluded due to sub-normal hearing as determined by pre-screening (participants are aware of this possibility, as it will be stated both in the participant information sheet and informed consent) *inclusion is determined by signing the consent form. It is the responsibility of the investigator to assure that all data in the course of the study will be entered completely and correctly in the respective database. Essential documents must be retained for at least 10 years after the regular end or a premature termination of the respective study (Art. 25 der KlinV). e-CRFs will not identify participants by their name or birth date, but will provide appropriate dummy-coded identification IDs. e-CRFs will be stored into a password-protected Excel sheet and will only be accessible to authorised personnel who require the data to fulfil their duties within the scope of the study. The principal investigator will control the correct data entry in the database.

Source documents in this study are demographic data, informed consent forms, randomization numbers, (S)AEs, and results of relevant examinations. Data recorded in the e-CRFs are the anonymized participant ID, age, date of participation, inclusion and exclusion criteria, hearing thresholds, assigned conditions and subjective treatment evaluation.

Confidentiality and Coding Trial and participant data will be handled with uttermost discretion and will only be accessible to authorised personnel who require the data to fulfil their duties within the scope of the study. On the CRFs and other study specific documents, participants are only identified by a unique participant number, stored in the participant identification list.

In order to protect it from unauthorised or accidental disclosure, alteration, deletion, copying and theft, but at the same time to ensure traceability (audit trail; ClinO, Art. 18), the investigator will store the participant identification list in two formats. An electronic version will be stored on a secure network folder of the responsible research centre (the University of Zurich (UZH) or Inselspital Bern) accessible only by the investigators. In addition, a paper version will be locked in the Psychologisches Institut building of the UZH and accessible only to the investigator.

Retention and Destruction of Study Data All study data are archived for 10 years after study termination or premature termination of the study. All the study data (including digital copies of physical documents) will be stored in two archives, in order to guarantee availability of a backup copy in case files in one of the archives become corrupted, incorrectly modified or lost. Thus, data will be stored on a secure network folder of the UZH accessible only by the project members and only within the University of Zurich network as well as on a folder of a mass storage hard disk drive (HDD) protected by password. If a UZH central backup storage unit will become available prior to termination of the study, data from the hard disk drive will be moved there. The paper data (i.e., informed consent and questionnaires) will be locked in the UZH building.

Monitoring and Registration Due to the low risks associated with this study, an internal study monitoring will be applied: the principal investigator will check the protocol, data, and informed consents obtained by the project members regularly for accuracy and will ensure that no data is missing and data storage in compliance with the protocol. The source data/documents will be accessible to monitors and questions will be answered during monitoring.

The study will be also registered in a national language (German) in the Swiss National Clinical Trial Portal (SNCTP) via BASEC.

Funding/Publication/Declaration of Interest The study is fully funded by the Swiss National Science Foundation (SNSF). After statistical analyses of each experiment, the project manager will make the unpublished not peer-reviewed pre-prints accessible through an open-access portal (e.g., BioRχive®) and will make every endeavour to publish the data in a scientific, peer-reviewed journal. Only non-clinical data will be made available in anonymized form. Trial results will be communicated to participants if they indicate that they wish to be informed of the study results.

All the project members declare no conflict of interest.