A Preliminary Report on an Adaptive Word-Recognition Protocol That Generates Word-Recognition Functions

Background: The most important functions of the auditory system are monitoring the acoustic environment for signals of danger and oral/auditory communication using language. During the typical evaluation of the auditory system, only cursory attention is focused on the communication aspect of hearing that usually consists of one list of 25 monosyllabic words presented by monitored-live voice to each ear in quiet at a suprathreshold presentation level (e.g., 40-dB sensation level). The reason most often given for the limited study of the word-recognition ability is the time required to administer the test.Purpose: An adaptive software-controlled word-recognition protocol that provides word-recognition performances over a range of 40 to 60 dB was developed and a key assumption was validated. The assumption was that in an ascending presentation level paradigm, when a word is correctly recognized, it will be correctly recognized at the subsequent higher presentation levels. Clinically, the protocol will include a list of monosyllabic words at ≈40- to 80-dB hearing level (HL) in 10-dB increments that include the soft to loud speech range.Research Design: A quasi-experimental repeated measures design was used.Study Sample: Groups of 24 young adults (mean age = 24.2 years) with pure-tone thresholds of 250 to 6000 Hz at ≤15-dB HL and 8000 Hz at ≤20-dB HL participated in two experiments in which the VA-1 female speaker version of Northwestern University Auditory Test No. 6 was delivered by a speaker at 6- to 36-dB HL in 6-dB increments with unique randomization of each list presented.Data Collection and Analysis: During both experiments, a 50-word list was assigned to each ear, and both ears were tested at each level before the subsequent higher level was evaluated. After testing, 25-word list data were parsed from the alphabetized 50-word lists according to odd- and even-numbered words. During Experiment 1, 50 words were presented at each level. During Experiment 2, except at the lowest level, only the words that were incorrect at the previous lower presentation level were presented, with a cumulative percent correct used at each presentation level. A computer recorded several characteristics of the presentation/response cycle, including the correct/incorrect response input by the tester. Both the raw (actual) data and assumed data were analyzed using polynomial equations and measures of central tendency.Results: Experiment 1 revealed that the raw data and assumed data produced word-recognition functions that were identical for all practical purposes. Experiment 2 revealed that a version of the adaptive word-recognition protocol produced word-recognition functions that were essentially the same as the functions produced during Experiment 1. Both experiments revealed that the word-recognition threshold established for each word (that is, the lowest presentation level at which the word was recognized) ranged from 9.2-dB HL to 30.3-dB HL. There was a weak relation between the word-recognition thresholds and the root-mean-square amplitudes of the words.Conclusions: In Experiment 1, the close relation between the raw data and assumed data provided the supporting data for the assumption to be considered a presumption going forward. In Experiment 2, the adaptive protocol worked as intended and provided word-recognition functions comparable to the functions in Experiment 1. Both the underlying assumption of the adaptive protocol and the protocol itself must be documented for patients with sensorineural hearing loss.