Technical Evaluation of Earswitch Phase B

The Earswitch is a demonstrated proof-of-concept that detects the voluntary contraction of a small muscle in the ear, called the tensor tympani (TT) muscle, which can be effectively used as an input switch. The voluntary activation of the TT makes the eardrum move, and such movement can be detected using a small camera inserted into the ear canal. A previously funded NIHR 'i4i Connect' research project showed how the Earswitch may be advantageous to populations with severe neuro-disabilities, where other communication methods are limited. This project aims to realise the EarSwitch's potential as an assistive communication device and provide supporting evidence towards regulatory approval of the medical device. The robustness and usability of the device will be tested on participants with mild-to-moderate neuro-disabilities and healthy participants. This complements data collected from assistive technology users following the same protocol but will provide additional data to train and understand the underlying detection algorithms for the Earswitch. The ease of which an assistive technology device can be installed and calibrated independently (without guidance from the researcher) so that it is ready to use is also an important consideration for its potential adoption. Participants will have the opportunity to interactively use the Earswitch daily at home, over a prolonged period of 4 weeks. Interviews and questionnaires will be used to gather information on usability and comfort of the device, whilst data from the interactive tasks will provide feedback of engagement and performance. Overall, this will provide crucial insights into how viable the Earswitch is as an assistive technology device and how accurately and reliably the current detection algorithm can detect contraction of the TT muscle. This data will be analysed to inform the final design of the Earswitch ready for commercial production.

Following recruitment and screening there are a number of different stages for the trial: Initial face-to-face appointment (Clinical groups 2,3, and 4) At-home setup (Clinical group 1) At-home testing (Clinical groups 1, 2, 3, and 4) Final face-to-face appointment (Clinical groups 2, 3, and 4) 1. Initial Face-to-Face Appointment (Clinical groups 2, 3, and 4) Participants will book an in-person appointment with PHU at Queen Alexandra Hospital. The initial face-to-face appointment will involve six main parts: Demographics Self assessment of skin colour by self identification on Fitzapatrick visual scale EarSwitch insertion Vital sign measurements, with synchronous EarSwitch in-ear video recording EarSwitch Click Distribution Analysis tasks Questionnaires on insertion and data collection Training for home use This should take 60-90 minutes. Part 1: Demographics (5 minutes) This will gather basic information about the participant including age, gender, ethnicity, self assessment of skin colour by Fitzpatrick visual scale and how long they have been aware of their ability to ear rumble (where applicable). Part 2: EarSwitch insertion (5 minutes) The participant will select an earbud size which will be inserted into the participant's ear on the left or right ear dependent on side of personal preference. The participant will then be asked if it is comfortable or whether they would like to try a different size according to the comfort or fit of the earbud. Once the participant is happy with the size of the earbud, the EarSwitch camera module will be inserted in the earbud and the EarSwitch control box connected to the research laptop by USB-C, on which the face to face trial software is loaded and running. The experimenter will ensure that the EarSwitch camera is outputting video on the experimenters trial software on the laptop, Once this is confirmed the experimenter willadvise the user how to install the Earswitch device in their ear. The experimenter will advise the subject how to position the EarSwitch camera.so that the ear drum is visible to the device, as viewed by the experimenter and user on the trial laptop. If the subject wishes the experImenter may help with initial insertion and positioning. During this procedure, the participant will be asked to let the experimenter know if it is uncomfortable or if they wish to stop. This should be no different to putting a regular ear bud into the participant's ear. The depth, direction, and curvature of the in-ear component of the EarSwitch is adjustable and the participant will be informed how to make these adjustments to ensure best comfort while ensuring the eardrum is visible. If earwax is obscuring the view of the video feed from the EarSwitch the subject will be offered microsuction of wax - either by research nurse trained to micro suction or audiologist available to microsuction, so that the protocol can be continued. Part 3: Vital sign measurements (15 minutes) As part of the exploratory objectives, the participant will have their vital signs measured while the EarSwitch records video of their ear drum and ear canal. The participant will be asked to sit at rest, avoiding unnecessary talking, during this period. The following vital signs will be measured: pulse, temperature, blood pressure, respiratory rate, ECG rhythm and oxygen saturations. Part 4: EarSwitch Click Distribution Analysis tests (25 minutes) The test software will then be run including recording in-ear videos of the participant contracting the tensor tympani (TT) muscle voluntarily in response to some tasks. The tasks will ask the participant to voluntarily contract the TT muscle in response to different stimuli when requested. These stimuli will be presented to the participant visually on a screen,. The purpose of this is to see how well the participant can voluntarily contract their TT muscle on demand, and to see how well the Earswitch can detect this. Tasks the participants will be asked to do include: Reactive task: contracting the TT muscle when a target appears. This will be repeated THREE times in succession Predictive task: contracting the TT muscle at the end of a countdown. This will be repeated THREE times in succession Double rumble task: contracting the TT muscle twice in quick succession (like a double click on a mouse). This will be repeated THREE times in succession Consecutive rumbling: contracting the TT muscle multiple times in a row in a short period of time (for example, 3 seconds). This will be repeated THREE times in succession Hold rumble: holding the TT contraction for a pre-defined amount of time (for example, 1 second). This will be repeated THREE times in succession In addition, video of the participant's ear drum will be recorded when they perform everyday tasks to understand if the EarSwitch detects other IEMs. The everyday tasks will be selected based on the capability of the participant but will include doing whatever they wish for two minutes, the participant will be asked to raise their eyebrow, yawn, move their head, and drink some water. During all these tasks the participant will not get any immediate feedback from the system, instead it will be recording the data which will be analysed after they have finished the study. Part 5: Questionnaires on insertion and data collection (5 minutes) Participants will remove the EarSwitch before completing questionnaires to gain feedback on the insertion of the device and the ability to ear rumble. The researcher will also measure the physical configuration of the EarSwitch by taking three photos of the device in a jig at 90 degree angles (x,y,z axes). The researcher will complete a study observation form to detail the researchers experience with the EarSwitch, including the ease of use, any product design observations. Part 6: Training for home use (10 minutes) The researcher will provide the subject with the EarSwitch device used in the home testing, and instruction sheets including QR code - detailing how to access the trial software on their own device and a training schedule. If the subject has brought their own device, tablet or laptop then they will be asked to connect to the PHU Wi-Fi and follow the instructions (prompted by researcher if necessary) to access the EarSwitch online trial interface, connect the EarSwitch device to the Bluetooth or USB of their own device, tablet or laptop, and operate a demonstration interface. The researcher will train the user on aspects of the device - including configuration using application, hardware controls, charging and connection. Next, the subject will be asked to run through set up procedure which will be demonstrated by the researcher to set sensitivity and alignment of the EarSwitch device. Finally, participants will complete a questionnaire about the setup of the software. 2. At-home setup (Clinical group 1) Participants in clinical group 1 (recruited by UoB) will not attend any face-to-face assessments. Participants will complete the demographics and pre-study questionnaire online, followed by the QUEST2 questionnaire based on their current assistive technology device. Participants will be sent the EarSwitch device to their home address via post. Instructions will be included to allow a self-guided walkthrough of the setup and installation process. Remote technical support will be on offer to participants via telephone or video call to assist them in setting up the device. Participants will complete online questionnaires to gain feedback regarding the setup and installation process. 3. At-home testing (Clinical groups 1, 2, 3, and 4) The at-home testing phase will involve five parts: Weekly tasks that are to be completed at the start of weeks 1, 2, 3, and 4, and finally at the end of week 4 (five in total). Daily tasks that are to be completed four times per week. Progression phone calls Post-testing questionnaires Exit interview and device return (clinical group 1) The subjects will be asked to consent (or not) to push notifications as reminders to complete tasks & use the EarSwitch. Part 1: Weekly tasks Once a week, all participants will be asked to complete a weekly interactive click distribution analysis (CDA) task, comprising of four tasks: Reactive task - a target will appear after a random time period and participants are asked to ear rumble as soon as they see the target. This will be repeated ?? THREE times in succession Predictive task - a target will appear after a predetermined amount of time which is communicated to the participant, the participant is asked to ear rumble when the target appears. This will be repeated ?? THREE times in succession Double rumble - same as the reactive task, except the participant is asked to perform a double ear rumble (i.e., two in a row). This will be repeated ?? THREE times in succession Consecutive rumble - participants will be given a duration (e.g., 5 seconds) where they are asked to ear rumble as many times as possible consecutively.This will be repeated ?? THREE times in succession These are the same tasks as in the face-to-face appointment (minus the hold rumble task) however r these are interactive which means the EarSwitch system will be actively running allowing the participants to interact with the interface using the EarSwitch and providing participants with real-time feedback. A summary of their performance will be shown upon completion. In all weekly tasks participants will be asked about their perceptions of ear rumbling using a quantitative online questionnaire. In the first and final weekly tasks, participants will be asked about comfort of the EarSwitch device and usability of the system using a quantitative online questionnaire. Part 2: Daily Tasks In addition to the weekly tasks, participants will be encouraged to use the EarSwitch device and software for approximately 10 minutes per day, four times per week (allowing for the weekly test day and two rest days). For each daily task, the participant will be asked to complete the following: One specified (of the four) interactive CDA tasks Type two specified words on an interactive keyboard (OPTIONAL) Play a "Dino Run" game which is a side-scrolling game featuring a dinosaur that has to jump over obstacles Part 3: Progression Phone Calls Participants will receive two phone calls from the research team during weeks 1 and 3 (exact time and day according to participant preference) to see if they require any support and to support and understand adherence to the protocol. Part 4: Post-testing questionnaires After the four-week period, participants will complete a set of quantitative questionnaires about their experience including their overall perception, usability, and comfort of the EarSwitch device. In addition, clinical group 1 will complete the QUEST2 based on the EarSwitch device and complete a health economic questionnaire. Part 5: Exit interview and device return (clinical group 1) An exit interview will be conducted with the participant to gain qualitative feedback about their experience using the EarSwitch. Arrangements to return the EarSwitch device to the University of Bath will subsequently be made. Final face-to-face appointment (Clinical groups 2, 3, and 4) Participants from clinical groups 2, 3, and 4 will finish the study with a final in-person assessment.. This will involve the same activities as the initial face-to-face appointment minus the demographics and training for home use, therefore: EarSwitch insertion Vital sign measurements during EarSwitch video recording EarSwitch Click Distribution Analysis tasks Questionnaires on insertion and data collection Finally, participants will be asked to complete some questionnaires to understand the participant's perception of interacting with the EarSwitch, including the comfort of the device and how easy it was to use. On this final visit, participants will be required to give back their Earswitch device to the research team at PHU. Follow up: There will be no long-term follow-up assessment of participants. During the 4-week testing period, the follow-up procedures are as followed: Intermittent testing at home using the web-based interface controlled by the EarSwitch, on a weekly basis (minimum) will be encouraged by reminders (if consented) by text/email (as per patient preference). Telephone calls during weeks 1 and at 3 weeks (approximately 20 min total per subject per call).

Participants in all clinical groups of the study will perform identical at-home procedures and be asked to complete the same weekly and daily tasks and relevant questionnaires. Participants in clinical groups 2,3 and 4 at PHU (Arm ) of the study will complete two additional in-person visits to Queen Alexandra Hospital at the beginning and end of the study, whereby additional data on the robustness of the EarSwitch will be tested.

Participants will be Assistive Technology Users (clinical group 1) and the study will be conducted remotely at their homes. The device will be sent via post to their address.

Three clinical groups will be recruited in Arm 2: 40 with mild-to-moderate motor neurodisabilities without communication assistive technology needs, 40 healthy participants and 20 who do not have the ability to ear rumble. Participants will be required to attend two in-person visits at the Queen Alexandra Hospital at the beginning and end of the study. This is in addition to completing the home procedure using the Earswitch device at home.

The Earswitch™ device consists of an earpiece made from silicon or similar bio-compatible material and incorporates a miniature camera. The earpiece sits within the ear canal and is not in direct contact with the eardrum and the camera is not in contact with the ear-canal. The camera detects voluntary movement from the tensor tympani muscle as presented at the eardrum and the Earswitch system translates this movement to an external virtual keypad or similar to allow the patient to communicate nonverbally.

To determine whether the EarSwitch system can be used for interaction by detecting in-ear movements (IEMs), and whether this differs with motor neurological disability and to assess its change over 4 weeks, in participants from three clinical groups.

The primary outcome is the success rate of how well individuals completed the interactive click distribution analysis (ICDA) tasks after the four-week period which indicates how well they can interact using the EarSwitch device. Comparisons will be made within each clinical group between baseline (start of week 1), intermediate (end of weeks 1, 2 and 3), and end-of-trial (end of week 4) to see whether repeated use increase the success rate of interacting with the EarSwitch. Comparisons will be made across the clinical groups to see if there is any difference in the success when using the EarSwitch device.

The frequency of EarSwitch usage as assessed by: average number of logins to the EarSwitch system. total number of tasks completed on the EarSwitch system Completed at the end of trial. Comparisons will be made across the clinical groups to see if there is any difference in frequency of usage when using the EarSwitch device.

The frequency of EarSwitch usage as assessed by time (in minutes) spent using the EarSwitch system. Completed at the end of trial. Comparisons will be made across the clinical groups to see if there is any difference in frequency of usage when using the EarSwitch device.

To determine the acceptability of the EarSwitch as assessed quantitively using a 7-point Likert item scale questionnaire.

Participants to answer a 7-point Likert items, exploring: how easy rumbling is to perform, how comfortable rumbling is, how successful participants perceived themselves & how physically/mentally demanding rumbling is. 1 reflects Strongly disagree - to 7 reflecting Strongly Agree. Completed baseline, end of weeks 1, 2 & 3, & end-of-trial. Comparisons will be made across the clinical groups to see if there is any difference in acceptability when using the EarSwitch device.

Participants to complete the pre-validated SUS questionnaire, consisting of 10 statements related to device usability and a 5-point likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). To be completed at baseline and at the end of the study. Comparisons will be made across the clinical groups to see if there is any difference in usability when using the EarSwitch device.

Qualitative questionnaire will be asked to gain detailed feedback with questions related to positioning of the wires, mounting of the device, design of the buttons, ease/ ability to clean the product, and support infrastructure (e.g., configuration application) completed at the end of study.

To determine how physically comfortable the EarSwitch device is, as assessed by quantitative measures based on subjective questionnaires measuring the comfort and fit of the device.

This will be measured using the Comfort Rating Scale. This is measured on a scale of low (1) to high (20) exploring five statements of: emotion, attachment, harm, perceived change, movement and anxiety. The participant will give a score of up to 20 for each factor based on their experience with the EarSwitch. To be completed at baseline and at the end of the trial. Comparisons will be made across the clinical groups to see if there is any difference in comfort when using the EarSwitch device.

To determine how physically comfortable the EarSwitch device is, as assessed by quantitative measures based on subjective questionnaires measuring the comfort and fit of the device.

This will be measured using a Bipolar Comfort Rating Scales. This is a pre-validated questionnaire which involves 14 different comfort indexes. Participants must rate a number between 1 and 7 for each index. A score of 4 would indicate a neutral response between each extremity, for example between painful - painless. To be completed at baseline and at the end of the trial. Comparisons will be made across the clinical groups to see if there is any difference in comfort when using the EarSwitch device.

To determine feasibility for the EarSwitch to be incorporated into routine clinical care, as assessed by quantitative questionnaires.

Questionnaires for set-up and installation of using the EarSwitch system. This will incorporate quantitative questions using a 5-point Likert scale whereby the scale 1 (Not very easy) - 5 (Very easy)] is used. This will be completed at the beginning of the trial, after the participants have completed the setup procedure

To determine feasibility for the EarSwitch to be incorporated into routine clinical care, as assessed by quantitative questionnaires.

The Health Economist questionnaire, a specialised survey designed by the health economist using quantitative tick box answers will be used to analyse the benefits of the EarSwitch use for assistive tech users. This will be completed at the end of the trial.

To determine feasibility for the EarSwitch to be incorporated into routine clinical care, as assessed by number of participant incidents across the 4 week period.

This will be measured by Number of participants completing to 4 weeks. Number of people who cannot use the device at baseline (including after micro-suction for clinical groups 2, 3, and 4) due to inadequate view of the ear drum and description of the obstruction (e.g., wax, debris, other anatomy). Number of support calls from patients to trial teams. Number of people who require ear micro-suction at baseline to be able to use the device (Clinical groups 2, 3, and 4). Number of serious or adverse events, number of ear infections. Comparisons will be made across the clinical groups to see if there is any difference when using the EarSwitch device.

Determine whether the ability to voluntarily contract the tensor tympani i.e., ear rumble in isolation is trainable

The following outcome measures will be recorded: Quantitative measures of success rate for the non-interactive CDA tasks at baseline (initial face-to-face appointment) and end-of-trial (final face-to-face appointment) for clinical group 4 only.

Determine whether the ability to voluntarily contract the tensor tympani i.e., ear rumble in isolation is trainable

The following outcome measures will be recorded: Numbers of participants able to self-identify whether they can rumble in isolation at the end of the 4 weeks for clinical group 4 only.

Determine whether the ability to voluntarily contract the tensor tympani i.e., ear rumble in isolation is trainable

An interview at the end of the 4 week testing period will be used to collect qualitative answers (from clinical group 4) to the following question: how they have been able to train or improve their ability to contract their tensor tympani?

• Quantitative metrics of how well individuals in clinical groups 2, 3, and 4 can ear rumble calculated using a custom piece of software which uses the manually labelled data from videos of the participants' ear drum combined with the time logs of the experimental software. This includes: (1) Reaction time from when the stimulus was shown until when the tensor tympani contracts, measured in milliseconds. Comparisons will be made across the clinical groups to see if there is any difference in reaction time when using the EarSwitch device.

• Quantitative metrics of how well individuals in clinical groups 2, 3, and 4 can ear rumble calculated using a custom piece of software which uses the manually labelled data from videos of the participants' ear drum combined with the time logs of the experimental software. This includes: (2) Time between ear rumbles (for double rumbles) measured in milliseconds. Comparisons will be made across the clinical groups to see if there is any difference using the EarSwitch device.

• Quantitative metrics of how well individuals in clinical groups 2, 3, and 4 can ear rumble calculated using a custom piece of software which uses the manually labelled data from videos of the participants' ear drum combined with the time logs of the experimental software. This includes: (3) Average number of consecutive ear rumbles (for consecutive rumbles) which do not have units (number of ear rumbles). Comparisons will be made across the clinical groups to see if there is any difference when using the EarSwitch device.

• Quantitative metrics of how well individuals in clinical groups 2, 3, and 4 can ear rumble calculated using a custom piece of software which uses the manually labelled data from videos of the participants' ear drum combined with the time logs of the experimental software. This includes: (4) Average duration of the hold rumbles measured in milliseconds. Comparisons will be made across the clinical groups to see if there is any difference when using the EarSwitch device.

How robust is the Earswitch technology in detecting voluntary eardrum movements? This will be measured by the F-score.

• This will be measured by the F-score .The F-score represents the robustness of the Earswitch in detecting TT contraction and is the harmonic mean of the precision and recall of the binary classifier used to assess whether an ear rumble has taken place.

How robust is the Earswitch technology in detecting voluntary eardrum movements? This will be measured by the Matthews correlation coefficient (MCC).

• This will be measured by the Matthews correlation coefficient (MCC). The MCC measures the quality of binary classifications, calculated from the confusion matrix which includes true positives, true negatives, false positives and false negatives.

How robust is the Earswitch technology in detecting voluntary eardrum movements? This will be measured by the receiver operating characteristics (ROC) curve.

• This will be measured using A ROC curve, which shows the performance of the classification model at all classification thresholds.

How robust is the Earswitch technology in detecting voluntary eardrum movements? This will be measured by the Area Under the ROC curve (AUC).

• This will be measured from the Area Under the ROC curve (AUC). The AUC metric will be calculated providing an aggregate measure of performance across all classification thresholds. AUC ranges from 0-1, with greater values indicating more correct. This gives an overall metric of the model that is invariant to the classification-threshold and can be used to assess different classification models.

Assess optimum design and fit of the EarSwitch in-ear device at baseline and end of week 4 visits in Clinical Groups 2, 3 and 4

The outcomes for assessing the optimum design and fit are: Range of depths (measured in mm) of EarSwitch device once removed from the participant's ear to inform future design decisions.

Assess optimum design and fit of the EarSwitch in-ear device at baseline and end of week 4 visits in Clinical Groups 2, 3 and 4

The outcomes for assessing the optimum design and fit are: Range of angles (measured in degrees) of EarSwitch device once removed from the participant's ear to inform future design decisions using a jig to ensure accurate alignment. The researcher will record a 3-axis photo of the EarSwitch in the jig for picture reference.

For all participants (Clinical groups 1, 2, 3, and 4) the following inclusion criteria will apply: Aged 18 and older. Able to give informed consent. Access to an internet-connected device (e.g., laptop/PC/ tablet/smart phone) allowing connection to the Earswitch web based trial system. Ability to operate, or have someone operate on their behalf, a web browser-based interface to respond to questionnaires. Normal or corrected-to-normal vision. Require the use of an assistive device (defined in Appendix….) in everyday life (Clinical group 1 only). For clinical groups 1, 2, and 3 the following will also apply: Able to voluntarily contract the tensor tympani muscle "in isolation" by self-identification, i.e., able to rumble the ear(s) "in isolation". For clinical group 4 the following will also apply: Unable to voluntarily contract their tensor tympani muscle in isolation by self-identification, i.e., unable to rumble the ear(s) "in isolation". For clinical group 1 the following will also apply: Individual's must require, or benefit from, the use of an assistive device in everyday life. For example, for use with communication, environmental control and/or mobility (excluding simple mechanical devices for example manual wheelchairs/walkers and frames) and excluding hearing-aids and visual aids. For clinical group 2 the following will also apply: Individuals must have a mild to moderate motor neurodisabilities and not require electronic assistive or communication (AAC) technology. The definition of mild to moderate motor-neurological disability is broad and will include participants with mild to moderate motor disabilities not requiring assistive technology (including MND/ Multiple Sclerosis/Parkinson's /Cerebral Palsy /Stroke / Post head injury /other rare degenerative diseases). Exclusion criteria The following exclusion criteria will apply: Those who do not have the capacity to understand the study and consent (e.g., severe learning disabilities). Inability to communicate either directly to the researcher or through a carer or communication partner and therefore inability to provide consent. Contraindication to use of EarSwitch including receiving treatment for an intercurrent ear infection or experiencing ear discharge or pain. Unable to view eardrum despite removal of ear-wax by ear micro-suction.

Participant data will be automatically anonymised and therefore individual participant data will not be available to other researchers. Anonymised data will be transferred to Earswitch Ltd and analysed in the future by Earswitch Ltd for the development of more robust algorithms for detecting voluntary tensor tympani movement.