Using touch to enhance auditory perception
Lead Research Organisation:
UNIVERSITY OF EXETER
Department Name: Mathematics
Abstract
Difficulty in following a single voice or conversation in noisy situations is one of the main frustrations for people with hearing impairment, even when fitted with modern hearing aids. Tactile (touch) sensation combined with sound is widely used in commonplace technology like mobile phones and video game controllers. This research investigates whether tactile stimulation (e.g. vibrations on the wrist), if synchronised with the sound around us, could improve our ability to separate out different sound sources (e.g. voices) in noisy situations. The research aims to demonstrate the benefit of this approach to help people with hearing loss, which leads to significantly reduced quality of life, impacts productivity in working life, affects 1 in 6 individuals in the UK and costs the NHS of 450 million pounds per year.
Sound sources can be separated based on differences in e.g. their pitch, location and timing: people talk at different speeds and start/stop stop talking at different times. One of the main features used to distinguish voices is their so-called fundamental frequency, e.g. male voices tend to sound lower than female voices. We are able to distinguish tactile vibrations in this same range of fundamental frequencies. If some important sound information, perhaps missing due to hearing loss, is automatically transformed into tactile stimulation, could this help to separate one voice from another? This research will use behavioural experiments with carefully designed combinations of sound and touch stimuli to answer this question. It aims to demonstrate the constraints for how this approach works best, firstly for normal-hearing listeners and then for people with hearing loss. The research will be supported by computational modelling to simulate processing in the brain that combines information from our senses of hearing and touch. This will help to link our own findings from behavioural experiments to research with brain imaging and other approaches used to understand how sound is processed with information from other senses.
In summary, the project investigates how tactile sensation affects sound source separation, how its effect is modified with impaired hearing and how information across the senses is combined through computations in the brain. These advances will provide a foundation of knowledge to improve strategies and technology to help with people impaired hearing by supporting the development of hearing aids that are enhanced by touch stimulation, e.g. on the wrist.
Sound sources can be separated based on differences in e.g. their pitch, location and timing: people talk at different speeds and start/stop stop talking at different times. One of the main features used to distinguish voices is their so-called fundamental frequency, e.g. male voices tend to sound lower than female voices. We are able to distinguish tactile vibrations in this same range of fundamental frequencies. If some important sound information, perhaps missing due to hearing loss, is automatically transformed into tactile stimulation, could this help to separate one voice from another? This research will use behavioural experiments with carefully designed combinations of sound and touch stimuli to answer this question. It aims to demonstrate the constraints for how this approach works best, firstly for normal-hearing listeners and then for people with hearing loss. The research will be supported by computational modelling to simulate processing in the brain that combines information from our senses of hearing and touch. This will help to link our own findings from behavioural experiments to research with brain imaging and other approaches used to understand how sound is processed with information from other senses.
In summary, the project investigates how tactile sensation affects sound source separation, how its effect is modified with impaired hearing and how information across the senses is combined through computations in the brain. These advances will provide a foundation of knowledge to improve strategies and technology to help with people impaired hearing by supporting the development of hearing aids that are enhanced by touch stimulation, e.g. on the wrist.
Publications

Fletcher MD
(2024)
Improved tactile speech robustness to background noise with a dual-path recurrent neural network noise-reduction method.
in Scientific reports

Fletcher MD
(2024)
Improved tactile speech perception and noise robustness using audio-to-tactile sensory substitution with amplitude envelope expansion.
in Scientific reports

Fletcher MD
(2024)
Improved tactile speech perception using audio-to-tactile sensory substitution with formant frequency focusing.
in Scientific reports

Fletcher MD
(2023)
Improving speech perception for hearing-impaired listeners using audio-to-tactile sensory substitution with multiple frequency channels.
in Scientific reports

Description | The project confirmed significant effects of haptic stimulation on auditory stream segregation. Mechanistic mathematical model developed in the project predicts that the timing of the haptic stimulation is one of the key characteristics that affects auditory processing. We experimentally validated the model prediction. The proposed model showed excellent agreement with the validation experiment. We also used new mathematical techniques to analyse model of auditory stream segregation that explicitly includes delays arising from signal transmission and processing in the auditory cortex. That is, a model described by a system of delay differential equations. Model analysis revealed good agreement between model and experimental data. This is an important step towards more detailed description of interactions between tactile and auditory stimuli. In addition, we identified speech features that can be transferred effectively through haptics. This allowed us to show that it is possible to use haptic stimulation to provide the basic speech cues needed to support different groups of people with hearing loss. Our experiments show that such haptic stimulation is robust to background noise. Altogether our findings imply that cross-modal synchronization, with carefully timed haptic cues, could improve auditory perception with potential applications in auditory assistive technologies aimed at enhancing speech recognition in noisy settings. |
Exploitation Route | Our findings have potential applications in auditory assistive technologies aimed at enhancing speech recognition in noisy settings and we are currently investigating this with our project partners. |
Sectors | Healthcare |
Title | Dataset for: Improved tactile speech perception using audio-to-tactile sensory substitution with formant frequency focusing |
Description | This dataset supports the publication: AUTHORS: Mark Fletcher, Esma Akis, Carl Verschuur, & Sam Perry TITLE: Improved tactile speech perception using audio-to-tactile sensory substitution with formant frequency focusing JOURNAL: Scientific Reports This dataset contains a CSV file with the participant number (matching the number used for the data presented in the published article associated with this dataset), dominant hand (left/right), wrist height, width, and circumference (mm), 31.5 Hz threshold and 125 Hz vibro-tactile detection threshold at the fingertip (ms/2), gender, age, and percentage correct for phoneme discrimination in each condition. The header name for each condition shows the frequency focusing method, the phoneme type (vowel or consonant), and the talker gender. Date of data collection: 01/01/2023 - 01/06/2023 All data was collected at the University of Southampton, U.K. |
Type Of Material | Database/Collection of data |
Year Produced | 2024 |
Provided To Others? | Yes |
URL | https://eprints.soton.ac.uk/id/eprint/487540 |
Title | Dataset supporting the publication "Improved tactile speech perception and noise robustness using audio-to-tactile sensory substitution with amplitude envelope expansion" |
Description | This dataset supports the publication by AUTHORS: Mark Fletcher, Esma Akis, Carl Verschuur, & Sam Perry, "Improved tactile speech perception and noise robustness using audio-to-tactile sensory substitution with amplitude envelope expansion" in Scientific Reports. This dataset contains a CSV file with the participant number (matching the number used for the data presented in the published article associated with this dataset), dominant hand (left or right), wrist circumference (mm), probe position on the wrist, screening vibro-tactile detection threshold at 125 Hz (ms/2), gender, age, and percentage correct for phoneme discrimination in each condition. The header name for each condition describes whether or not expansion has been applied ("NoExp" or "Exp") and whether discrimination is in quiet or in noise ("Noi" or "Qui"). Data is separated by consonants and vowels ("C" or "V") and whether speech is the male or female talker ("F" or "M"). |
Type Of Material | Database/Collection of data |
Year Produced | 2024 |
Provided To Others? | Yes |
URL | https://eprints.soton.ac.uk/id/eprint/491446 |
Title | Dataset supporting the publication "Improved tactile speech robustness to background noise with a dual-path recurrent neural network noise-reduction method" |
Description | This dataset supports the publication: Fletcher, M., Perry, S., Thoidis, I., Verschuur, C., & Goehring, T. (2024). Improved tactile speech robustness to background noise with a dual-path recurrent neural network noise-reduction method. Scientific Reports. This dataset contains three CSV files: one for the objective assessment of the audio, one for the objective assessment of the tactile signal, and one for the behavioural assessment. The objective audio CSV file shows the eSTOI and SI-SDR scores for each model at each of the SNRs tested for the four different noise reduction methods and with no noise reduction. Data is shown for the Party noise and ITASS noise. The objective tactile CSV file shows the SI-SDR scores for each model at each of the SNRs tested for the four different noise reduction methods and with no noise reduction. Rows show different sentences. Column headings stat the processing applied (no processing, log-MMSE, or DPRNN methods), the SNR, and whether the data is for the male ("M") or female ("F") talker. The behavioural CSV file shows the participant number (matching the number used for the data presented in the published article associated with this dataset), dominant hand (left/right), wrist height, width, and circumference (mm), 31.5 Hz threshold and 125 Hz vibro-tactile detection threshold at the fingertip (ms/2), wrist temperature (0C), gender, age, and percentage correct for sentence identification in each condition. The header name for each condition shows whether or not noise reduction ("NR") was applied, the SNR, and whether the talker was male or female. All data was collected at the University of Southampton, U.K. |
Type Of Material | Database/Collection of data |
Year Produced | 2024 |
Provided To Others? | Yes |
URL | https://eprints.soton.ac.uk/id/eprint/488218 |