Assessing the initiation of failure is critical to understanding the progression of spinal injuries and to developing injury criteria. These incipient injuries are being assessed more frequently using acoustic emissions. However, the signals tend to be chaotic especially under high‐rate loading where simply the presence of acoustic emissions does not necessarily indicate injury. Therefore, this paper presents a technique to interrogate the acoustic emission signals to determine when cortical bone failure occurs. Isolated cortical bone specimens were loaded at 10 mm/min to failure and the resulting acoustic emissions had wideband response with peaks from 20 to 900 kHz, with the dominant peak being 59.6 ± 21.9 kHz (mean ± standard deviation). Whole cervical spines were then loaded at fast rate (1.2 m/s) to cause injury. A continuous wavelet transform using a Morlet mother wavelet showed the 59.6 kHz frequency component entered the time domain at many points in time, which could be calculated using the described algorithm. The first point was determined to indicate the initiation of failure due to its low amplitude, and the following two points suggested larger failure. This work presents a necessary set of tools for effectively utilising acoustic emissions in biological material testing.