Premium
Distortion product emission features and the prediction of pure tone thresholds
Author(s) -
Kimberley Barry P.
Publication year - 1995
Publication title -
the laryngoscope
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.181
H-Index - 148
eISSN - 1531-4995
pISSN - 0023-852X
DOI - 10.1288/00005537-199504000-00001
Subject(s) - audiology , categorization , linear discriminant analysis , tone burst , discriminant function analysis , tone (literature) , distortion (music) , data set , set (abstract data type) , mathematics , pattern recognition (psychology) , psychology , statistics , medicine , computer science , artificial intelligence , telecommunications , art , amplifier , literature , bandwidth (computing) , programming language
Distortion product emission (DPE) growth functions, demographic data, and pure tone thresholds (PTTs) were recorded in 229 normal‐hearing and hearing‐impaired ears. Half of the data set (115 ears) was used by a discriminant analysis routine to classify DPE and demographic features into either a normal PTT group or an impaired PTT group (PTT greater than 30 dB SPL [sound pressure level]) at six frequencies in the audiometric range. The six discriminant functions developed from this classification process were then used to predict PTT group membership in the remaining 114‐ear data set. Frequency‐specific prediction accuracy was approximately 85% overall. Of the 45 DPE and demographic variables evaluated, the DPE amplitude associated with an f 2 (a primary tone of frequency) of moderate level (50 dB SPL) and a frequency corresponding to PTT was generally most predictive. DPE features associated with frequencies immediately adjacent to the PTT frequency also appear to be useful. DPE level was found to be weakly correlated with subject age; perhaps for this reason, age was frequently included in discriminant functions. This study describes the DPE measures that can most reliably categorize PTTs as normal or impaired in large populations with varied cochlear hearing status.