Stream Segregation and Speech Recognition in Noise in Individuals With Cochlear Implants

Status

Completed

Phases

Unknown

Study type

Observational

Source

ClinicalTrials.gov

Registry ID

NCT04854031

Enrollment

Registered

2021-04-22

Start date

2021-08-02

Completion date

2022-03-31

Last updated

2023-05-18

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Cochlear Implants

Brief summary

Individuals with cochlear implants will complete tasks which measure auditory resolution, working memory, stream segregation, and speech recognition in the presence of competing speech using their everyday clinical device settings. The relationship between these tasks will be examined to identify the factors which predict successful speech recognition in the presence of competing speech.

Detailed description

Individuals with cochlear implants struggle to understand speech in the presence of competing talkers because they have trouble segregating auditory streams. The premise of this project is that individual differences in auditory resolution and cognitive ability across individuals with cochlear implants determine the extent to which they can segregate auditory streams from one another and hear out target speech embedded in competing talkers. Our goal is to test whether the link between speech recognition in noise and individual differences in auditory resolution and working memory in individuals with cochlear implants is due to the limitations that these individual differences place on stream segregation. The outcome measure to be predicted is sentence recognition in two-talker babble. Previous work has found that spectral and temporal modulation detection thresholds (measures of auditory resolution) and performance on the reading span task (a measure of working memory that is closely linked to fluid intelligence) are predictors of speech recognition in quiet. To account for these sources of variability in speech recognition, we will verify that these tasks jointly predict individual differences in sentence recognition in quiet. Adding competing talkers during the speech recognition task will introduce additional variability beyond the variability of speech recognition in quiet. We hypothesize that this additional variability with competing talkers should be predicted by individual differences in stream segregation ability, which will in turn be predicted by auditory resolution and working memory. Obligatory and voluntary stream segregation ability will be measured using rapidly presented digit sequences manipulated to have alternating fundamental frequencies (F0) for each digit. Participants will resist integration and repeat back only the digits presented with the higher F0. The magnitude of the F0 alternation will be manipulated to control the difficulty of segregating streams. The predicted relationship between stream segregation and sentence recognition will be tested for auditory resolution and working memory in independent and combined models. Completing this goal will identify the auditory and cognitive factors that support stream segregation in post-lingually deafened adults with cochlear implants. This identification will enable development of cochlear implant design and rehabilitation strategies to facilitate stream segregation in these patients as well as investigation of the developmental trajectories of these factors in children with cochlear implants. We plan to implement this study in at-home testing conditions to avoid risking coronavirus disease 2019 transmission in the lab, so this work will also determine the feasibility of at-home testing of individuals with cochlear implants. At-home testing would expand the amount and diversity of participants we are able to recruit for future studies.

Interventions

DEVICECochlear Implants

Individuals with cochlear implants will be tested on their hearing ability. Auditory recordings of speech will be played to participants from a loudspeaker. Recordings will be edited to add competing noise sources and to adjust talker voice pitch. Synthetic sounds will be manipulated to control auditory cue salience in detection tasks.

Study design

Observational model

CASE_ONLY

Time perspective

CROSS_SECTIONAL

Eligibility

Sex/Gender

ALL

Age

19 Years to 80 Years

Healthy volunteers

Inclusion criteria

* Has at least one cochlear implant. * Lost their hearing during adulthood. * Native English speaker.

Exclusion criteria

* Cognitive impairment

Design outcomes

Primary

Measure	Time frame	Description
Speech Recognition	Up to 30 minutes in each of two listening conditions.	The metric for speech recognition is the percentage of Perceptually Robust English Sentence Test Open-set (PRESTO) sentence keywords that were correctly repeated in order.
Temporal Modulation Detection Threshold	Up to 30 minutes	The metric for temporal modulation detection is the modulation depth relative to 100% modulation which the participant can detect 71% of the time.
Spectral Modulation Detection Thresholds	Up to 30 minutes	The metric for spectral modulation detection is the magnitude of the peak-to-valley ratio of the spectrally modulated stimulus which the participant can detect 71% of the time.
Reading Span Task Performance	Up to 20 minutes	The outcome is the percentage of letters recalled in the correct position across all trials, out of a total of 75 letters.
Digit Span Task Performance	Up to 20 minutes	The outcome is the total percentage of digits recalled in the correct position across all trials, out of a maximum of 220.
Free Recall Task Performance	Up to 10 minutes	The outcome is the average number of words recalled from each list, with a possible maximum of 12.
Digit Updating Task Performance	Up to 20 minutes	The outcome is the percentage of numbers correctly recalled across boxes, out of a total of 49.
Running Digit Span Task Performance	Up to 15 minutes	The outcome is the average number of digits recalled in the correct position relative to the last item in the sequence across lists.
Stream Segregation	Up to 1 hour	The metric for stream segregation is the percent change in digit recall that occurs when voice pitch differs from the distractor digits (85, 120, and 150 Hz) relative to when voice pitch between target and distractor digits is the same (200 Hz).

Countries

United States

Participant flow

Participants by arm

Arm	Count
Cochlear Implant Recipients 8 participants who lost their hearing and received one or two cochlear implants as adults will participate in this study. We will include unilaterally and bilaterally implanted individuals listening with their everyday hearing configuration. Individuals with residual acoustic hearing better than 60 A-weighted decibels at any audiometric frequency will be excluded. Participants will range in age between 19 and 80 years old, although most are expected to be within 50 - 75 years of age. Cochlear Implants: Individuals with cochlear implants will be tested on their hearing ability. Auditory recordings of speech will be played to participants from a loudspeaker. Recordings will be edited to add competing noise sources and to adjust talker voice pitch. Synthetic sounds will be manipulated to control auditory cue salience in detection tasks.	8
Total	8

Arm

Count

Cochlear Implant Recipients

8 participants who lost their hearing and received one or two cochlear implants as adults will participate in this study. We will include unilaterally and bilaterally implanted individuals listening with their everyday hearing configuration. Individuals with residual acoustic hearing better than 60 A-weighted decibels at any audiometric frequency will be excluded. Participants will range in age between 19 and 80 years old, although most are expected to be within 50 - 75 years of age. Cochlear Implants: Individuals with cochlear implants will be tested on their hearing ability. Auditory recordings of speech will be played to participants from a loudspeaker. Recordings will be edited to add competing noise sources and to adjust talker voice pitch. Synthetic sounds will be manipulated to control auditory cue salience in detection tasks.

Total

Baseline characteristics

Characteristic	Cochlear Implant Recipients
Age, Categorical <=18 years	0 Participants
Age, Categorical >=65 years	4 Participants
Age, Categorical Between 18 and 65 years	4 Participants
Age, Continuous	67 years
Ethnicity (NIH/OMB) Hispanic or Latino	1 Participants
Ethnicity (NIH/OMB) Not Hispanic or Latino	7 Participants
Ethnicity (NIH/OMB) Unknown or Not Reported	0 Participants
Race (NIH/OMB) American Indian or Alaska Native	1 Participants
Race (NIH/OMB) Asian	0 Participants
Race (NIH/OMB) Black or African American	0 Participants
Race (NIH/OMB) More than one race	0 Participants
Race (NIH/OMB) Native Hawaiian or Other Pacific Islander	0 Participants
Race (NIH/OMB) Unknown or Not Reported	0 Participants
Race (NIH/OMB) White	7 Participants
Region of Enrollment United States	8 participants
Sex: Female, Male Female	3 Participants
Sex: Female, Male Male	5 Participants

Adverse events

Event type	EG000 affected / at risk
deaths Total, all-cause mortality	0 / 8
other Total, other adverse events	0 / 8
serious Total, serious adverse events	0 / 8

Outcome results

Primary

Digit Span Task Performance

The outcome is the total percentage of digits recalled in the correct position across all trials, out of a maximum of 220.

Time frame: Up to 20 minutes