Automated Analysis of the Auditory Brainstem Response Using Derivative Estimation Wavelets

Abstract

In this paper, we describe an algorithm that automatically detects and labels peaks I–VII of the normal, suprathreshold auditory brainstem response (ABR). The algorithm proceeds in three stages, with the option of a fourth: (1) all candidate peaks and troughs in the ABR waveform are identified using zero crossings of the first derivative, (2) peaks I–VII are identified from these candidate peaks based on their latency and morphology, (3) if required, peaks II and IV are identified as points of inflection using zero crossings of the second derivative and (4) interpeak troughs are identified before peak latencies and amplitudes are measured. The performance of the algorithm was estimated on a set of 240 normal ABR waveforms recorded using a stimulus intensity of 90 dBnHL. When compared to an expert audiologist, the algorithm correctly identified the major ABR peaks (I, III and V) in 96–98% of the waveforms and the minor ABR peaks (II, IV, VI and VII) in 45–83% of waveforms. Whilst peak II was correctly identified in only 83% and peak IV in 77% of waveforms, it was shown that 5% of the peak II identifications and 31% of the peak IV identifications came as a direct result of allowing these peaks to be found as points of inflection.

On the dual structure of the auditory brainstem response in dogs

“…These waveforms were transferred in ASCII format from their location on the Biologic Evoked Potential system to a personal computer running Mathworks Matlab Ó Software Version 6.5. A Matlab Ó M-file [adapted from Wilson (2004) and Bradley and Wilson (2005)] was then used for all further signal processing.…”

“…Both the ABR and the reconstructed ABR OCDWT signals were finally passed through a previously written peak and trough finding algorithm (Bradley and Wilson, 2005) to identify absolute wave latencies and the precedingtrough-to-peak ''a'' and peak-to-following-trough ''b'' amplitudes of all major waves (peaks). To confirm the accuracy of these analyses, all calculations were checked manually and corrected where required.…”

“…The discrete wavelet transform (DWT), and its use as a multiresolution analysis (MRA) tool, has been widely described in the literature (Jawerth and Sweldens, 1994;Hess-Nielsen and Wickerhauser, 1996;Unser and Aldroubi, 1996;Blinowska and Durka, 1997;Samar et al, 1999;Wilson, 2002;Bradley and Wilson, 2004;Wilson, 2004;Zhang et al, 2004;Bradley and Wilson, 2005;Zhang et al, 2005). In summary, the DWT is a form of digital filtering capable of deconstructing a signal into its component scales (frequency ranges), and then detailing how each scale evolves over time.…”

Over-Complete Discrete Wavelet Transformation of the Normal Auditory Brainstem Response Improves Prediction of Outcome following Severe Acute Closed Head Injury

“…This both reduced the number of signals to be analysed and improved the fi nal signal-to-noise ratios. These averaged ABRs were baseline shifted to a starting baseline of 0 V and a previously written peak and trough fi nding algorithm [Bradley and Wilson, 2005] was used to identify the latencies and preceding-troughto-peak 'a' and peak-to-following-trough 'b' amplitudes of each ABR's wave I, III and V. The overall root mean square amplitude of each ABR was also determined. To confi rm the accuracy of these analyses, the authors manually checked all ABR signals prior to accepting their data.…”

“…The resulting wavelet coeffi cients were then used to reconstruct the ABR signal at scales A6 (512 coeffi cients, 0-267 Hz), D6 (512 coeffi cients, 267-533 Hz), and D5 (512 coeffi cients, 533-1067 Hz) only, as these scales ensured access to the frequency range of relevance to the ABR signal (0-1500 Hz). These reconstructed ABR OCDWT signals were then passed through a previously written peak and trough fi nding algorithm [Bradley and Wilson, 2005] to identify the latencies and preceding-trough-to-peak 'a' and peak-to-following-trough 'b' amplitudes of the reconstructed ABR OCDWT peaks corresponding in latency to the ABR waves I, III and V, that is, waves A6V, D6I, D6III, D6V, D5I, D5III and D5V. The overall root mean square amplitude of each reconstructed ABR OCDWT signal was also calculated.…”