Use of a Non-Linear Algorithm to Improve the Maximum Length Sequence Auditory Brainstem Response
Background: The auditory brainstem response (ABR) is the most preferred tool in Universal Newborn Hearing Screening (UNHS) because of its high sensitivity and specificity and its ability to detect abnormalities up to auditory nerve. However, the testing time is considerably long does increase the c...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2012
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/26739/ http://irep.iium.edu.my/26739/4/IRIIE_2012A.pdf |
| Summary: | Background:
The auditory brainstem response (ABR) is the most preferred tool in Universal Newborn Hearing Screening (UNHS) because of its high sensitivity and specificity and its ability to detect abnormalities up to auditory nerve. However, the testing time is considerably long does increase the cost of running UNHS. One of the possible techniques to shorten the ABR testing time is by using high stimulus repetition rate through linear Maximum Length Sequence (lMLS). This algorithm has been reported to have poor signal to noise ratio thus increase the testing time. The present study has developed a new non linear MLS (nlMLS) which aims to overcome the mismatch between the non-linearity of the auditory system and the linearity aspect of the algorithm.
Aim
This study is aimed to investigate the effect of using the novel nonlinear MLS (nlMLS) to the ABR results in newborn subjects.
Methodology
30 newborn subjects whom passed the UNHS were involved in this study. ABR were recorded using MLS using both linear and non linear reconstruction at 180, 250, 500 and 836 cps using a custom built evoked potential system. In addition, the ABR was recorded using vertical electrode montage (non-inverting: high forehead, inverting: nape of neck and ground: shoulder). Next, the wave V amplitude and signal to noise ratio (SNR) were calculated for each ABR. The ABR wave V amplitude was calculated from the peak of wave V to the following trough while SNR was derived from variance ratio at single point (Fsp) formula. MLS ABR were recorded using order 6 with 160 trains for 180 cps, 240 trains for 250 cps, 280 trains for 500 cps and 320 trains for 836 cps.
Results
Results show that the nlMLS has significantly higher amplitude and better SNR value than ABR recorded with lMLS (RM ANOVA, p<0.05) at the same stimulus repetition rate. The amplitude of wave V value increase from 13.3 to 53.6% when changing from lMLS to nlMLS with the maximum increment observed in MLS ABR at 500 cps. The SNR increase from 19.8 to 63 % when changing from lMLS to nlMLS with the maximum increment observed in MLS ABR at 500 cps.
Conclusion
With these results, nonlinear algorithm is considered as a good tool to be applied in ABR recording in particular for UNHS as it can improve the SNR, amplitude and further more can save the recording time.
* This paper has been presented at XXII Biennial International Evoked Response Audiometry Study Group Symposium, Moscow 2011
|
|---|