The parallel auditory brainstem response paradigm provides accurate and fast hearing thresholds in a clinic-like
Melissa J Polonenko1, Ross K Maddox2
1Department of Speech-Language-Hearing Sciences, University of Minnesota.
Objectives:
The auditory brainstem response (ABR) is an essential tool in screening for and diagnosing infant hearing loss, and its results drive decisions regarding interventions and hearing habilitation with impacts extending far into a child's future. Despite the traditional ABR exam's usefulness, there is an identified need to develop faster, more informative exams. The parallel ABR (pABR) measures responses to all frequencies of interest in both ears all at once, rather than the traditional series of single-frequency measurements in one ear at a time, greatly speeding the diagnostic exam. The pABR has been shown to be effective at quickly measuring frequency-specific responses in adults with normal hearing, but it has not yet been tested in people with hearing loss. The goal of this study was to determine the accuracy and speed of the pABR for estimating hearing thresholds in a clinic-like setting.
Design:
Seventy adults with widely varying sensorineural hearing loss configurations were recruited to participate in this study. We measured thresholds at octave frequencies in two ways: the behavioral audiogram, serving as the ground truth, and using the pABR with a custom-designed interactive user interface. Accuracy was determined through threshold correlation coefficients as well as absolute error in decibels. Acquisition time was assessed as the time from measurement start to determination of the final threshold. To determine the pABR's speed advantages, a subset of participants was invited back and their thresholds estimated a third time, using a commercially available clinical system to serially measure ABR waveforms. Speedup was assessed in terms of the raw difference in acquisition time in minutes and as the ratio between measurement times made with the two ABR paradigms.
Results:
Thresholds estimated with pABR highly correlated with the behavioral audiogram ground truth. The correlation was 0.90 (0.88-0.92, 95% confidence interval) across all ears and frequencies. 79% of pABR thresholds were within one 10-dB step-size of the behavioral threshold. The pABR was faster in all ten participants where traditional serial ABR was also recorded, with a mean recording time of 28 minutes to estimate ten pABR thresholds (500-8000 Hz in each ear) versus 70 minutes to estimate eight serial thresholds (500-4000 Hz in each ear), or a mean reduction of 42 minutes. The median speedup ratio was 2.5×.
Conclusions:
The pABR provides accurate threshold estimates with greatly reduced measurement time compared to traditional methods. Given these results and other advantages related to its design, the pABR holds promise as a clinical tool that can be deployed to commercial systems in the near future.
