Objective: A clinical protocol for diagnosing hearing loss (HL) in infants designed to meet early intervention guidelines was used with the goals of providing normative data for (1) frequency-specific tone burst auditory brain stem response (TBABR) thresholds by air conduction (AC) and bone conduction (BC) in early infancy used to classify type and severity of HL, (2) ear-specific behavioral thresholds for these same infants by 1 yr of age, and (3) the relationship between TBABR thresholds and behavioral thresholds for this group of infants.

Design: AC- and BC-TBABRs were measured in young infants (mean age, <3 mo) under natural sleep to classify the type and severity of HL (conductive, sensorineural, or mixed). A small group of normal-hearing adults undergoing the same TBABR protocol served as a control group. Threshold and latency data for AC- and BC-ABR were analyzed for infants classified as having normal hearing and for those with and without conductive HL. The ability to detect conductive HL based on ABR latencies evoked by clicks presented at 80 dB nHL was assessed. Behavioral thresholds using visual reinforcement audiometry (VRA) were measured in infants at a mean age of approximately 10 mo. The relationship between TBABR and behavioral thresholds obtained in infancy was analyzed, and the prediction of behavioral thresholds from TBABR thresholds was examined.

Results: Mean TBABR thresholds in young infants with normal hearing tested under natural sleep were similar to previously published data. The relationship between AC- and BC-TBABR thresholds differed as a function of stimulus frequency for infants but not adults. A mean air-bone gap (ABG) of 15 dB was present at 500 Hz even in normal-hearing infants, with those infants classified as having conductive HL presenting with substantially larger ABGs. Wave V latency functions for AC- and BC-TBABR also differed between infants and adults as a function of frequency. Infant BC-TBABR latencies were well matched between those with normal hearing and conductive HL, whereas AC-TBABR latency functions separated these groups. Mean VRA thresholds using insert phones in normal-hearing infants tested were between 14 and 17 dB HL for all three test frequencies at a mean age of 9.7 mo. Correlations between TBABR and VRA thresholds, both obtained during infancy, were strong for all three test frequencies (r = 0.86, 0.90, and 0.91 for 500, 2000, and 4000 Hz, respectively).

Conclusions: AC- and BC-TBABR results can be readily obtained in young infants under natural sleep and were used to classify the type of HL based on the absolute threshold and the size of the ABG. Differences in wave V latency functions for TBABR by AC and BC and wave I and V latencies of the high-level click ABR also distinguish between infants with and without TBABR ABGs. Ear-specific behavioral responses can be obtained at levels under 20 dB HL in normal-hearing infants younger than 1 yr using VRA, and these behavioral thresholds correlate well with TBABR thresholds obtained on average 6.5 mo previously in this population. The current results suggest that protocols for obtaining AC- and BC-TBABR and behavioral thresholds that meet guidelines for early intervention are clinically feasible.