The Stability Of Speech-evoked Auditory Brainstem Response

Background and ObjectivesShort latency auditory evoked potentials (AEP) elicited by clicks and tones are widely applied in clinics for their stability, repeatability and the imunity of conscious. By investiaging the short latency AEP, people have a certain understanding of how simple sounds like click are processed by auditory system. However, the process mechanism of complex sounds in auditory system is largly unknown.There is a highly correlated relationship between brainstem neuron discharge and speech acoustic characteristics, which suggests that brainstem is an important part of speech processing and the auditory brainstem evoked potential (ABR) recording is a convenient tool to study the auditory speech processing mechanism. Most of the researches on speech evoked ABR (Speech-ABR) are reported by research group led by Professor Kraus in Northwest University of USA. They presented a synthesis syllable/da/to elicit the brain stem responses. According to the report, speech ABR is composed of onset components (positive peak A and negative peak C) and frequency following responses (FFR) (negative peaks C, D, E and F). Children with learning problem were studied and found the children’s learning problem can be reflected by their Speech-ABRs. The latencies and amplitudes of AEPs are commonly used as clinical indicators, requiring a stable recording. This article is to explore the recording stability of speech-ABRs of Chinese healthy young adults. MethodThe subjects in this study are 40 young adults (20 males,20 females) whose ages are ranged from 19-26. All of the subjects are right-handed, with no auditory, neuron and mental diseases, and normal click-ABRs. The subjects were randomly divided into 2 groups (each group has 20 subjects), and the 2 groups were tested in succession with a time interval of 2 months. The right ears of subjects were stimulated with click and/da/as wildly used in speech ABR recording, so we can obtain 1 block of click-ABR and 3 blocks of speech-ABRs (1024 stimuli were presented in each block, and each block should be include 1024 sweeps). Notice that the sweep number of each block in average process should not be less than 1000, and the time interval of block testing is 1 minute. The analysis was applied to the single and average waveform of 3 blocks of response and the grand average of each group. By referring to the criteria developed previously, the latencies and amplitudes of these feature peaks detected by computer software developed on Matlab platform and the detection rates of feature peaks were analyzed statistically.ResultsThere is a highly correlated relationship between FFR of Speech-ABR and the quasi-periodic peaks in/da/, which can be concluded that the D-E-F periodic components were evoked by the corresponding quasi-periodic peaks in/da/. It is suggested that the FFR containing fundamental frequency and formant information of/da/reflects the phase-lock characteristic to/da/. Although there are noise-like waves in single speech-ABRs, the noise-like waves can be attenuated greatly after averaging. Both the single recording and the averaged recording of single subject contained most or all of feature peaks, and the latencies and amplitudes of these averaged recording in two groups were not significant difference (p>0.1). It is demonstrated that the recording to speech-ABR is stable. Among the feature peaks, the detection rates of V, A, C and F peaks were higher (40%-60%) than others (20%-30%), and the detection rates of transient peaks (V, A, C and O) are higher than periodic peaks (D, E and F). ConclusionThe peaks of D, E and F were phase-locking to the peaks of d, e, and f in/da/, suggesting that the acoustic information of/da/was reflected perfectly by Speech-ABR. The single Speech-ABR, averaged Speech-ABR of single subjects and the grand averaged Speech-ABR of two groups showed a high similarity, which demonstrated the recording is stable. The latencies and amplitudes of feature peaks in grand averaged Speech-ABRs were no significant difference, which means the recording is repeatable. The high detection rates showed that the V, A, C and F peaks were more stable than other peaks and can be mainly analyzed. In a word, the recording of Speech-ABR is stable and repeatable, and the V, A, C and F peaks can be used as important indicators of clinical observation, showing that Speech-ABR is a promising tool in the study of Speech perception mechanism.