A vascular subtraction method for improving the variability of evoked tympanic membrane displacement measurements

Evoked tympanic membrane displacement (TMD) measurements show a correlation with intracranial pressure (ICP). Attempts to use these measurements for non-invasive monitoring of ICP in patients have been limited by high measurement variability. Pulsing of the tympanic membrane at the cardiac frequency has been shown to be a significant source of the variability. In this study we describe a post processing method to remove the cardiac pulse waveform and assess the impact of this on the measurement and its repeatability.

Three-hundred and sixteen healthy volunteers were recruited for evoked TMD measurements. The measurements were quantified by V _m, defined as the mean displacement between the point of maximum inward displacement and the end of the stimulus. A sample of spontaneously pulsing TMDs was measured immediately before the evoked measurements. Simultaneous recording of the ECG allowed a heartbeat template to be extracted from the spontaneous data and subtracted from the evoked data. Intra-subject repeatability of V _m was assessed from 20 repeats of the evoked measurement. Results with and without subtraction of the heartbeat template were compared. The difference was tested for significance using the Wilcoxon sign rank test.

In left and right ears, both sitting and supine, application of the pulse correction significantly reduced the intra-subject variability of V _m (p value range 4.0 × 10^-27 to 2.0 × 10^-31). The average improvement was from 98 ± 6 nl to 56 ± 4 nl.

The pulse subtraction technique substantially improves the repeatability of evoked TMD measurements. This justifies further investigations to assess the use of TMD measurements in clinical applications where non-invasive tracking of changes in ICP would be useful.