Ultrafast ultrasound imaging is used to capture driven spinal vibrations as a new method for noninvasive spinal testing in living subjects. Previously, it has been shown that accelerometer-based vibration testing in cadaveric models can reveal the presence, location and magnitude of spinal pathology. However, this process remains an invasive procedure as current non-invasive sensors are inadequate. In this paper, the ability of non-invasive ultrafast ultrasound to quantify in vivo vertebral vibration response across a broad range of frequencies (10-100Hz) in anesthetized pig models is investigated. Close agreement with invasive accelerometer measurements is achieved using the noninvasive ultrasound method, opening up unique opportunities to investigate spinal pathologies.Globally, low back pain is the leading cause of years lived with disability 1 . As a result, the direct and indirect costs of low back pain are enormous adding to the significant morbidity caused by this chronic, recurrent condition. Unfortunately, only 10 percent of low back pain cases have an identifiable cause 2 . Without knowing its cause, the prevention, identification and treatment of low back pain is ineffective at best which perpetuates its impact on society.While there are many ways to measure spinal function (e.g. range of motion), current limitations in diagnosing back pain are thought to arise largely from the failure of current tests to measure spinal functions that are linked to meaningful clinical outcomes. Similarly, current imaging techniques display static anatomy very well, but reveal little about the function of a mechanical structure like the spine.Given the above, we have looked to examples of technologies able to evaluate the functional status of mechanical or geological targets. One such technology is non-destructive vibration testing which is used to identify the location, presence and quantity of oil deposits in seismic testing and structural flaws in aerospace and civil construction projects [3][4][5][6][7][8] .With this approach, we have shown previously that when vibration is applied to cadaveric spines, the response of accelerometers implanted within lumbar vertebrae are unique and can be used to identify a range of simulated pathologies 9-11 . For example, the vibration spectrum approach was capable of differentiating vertebrae tied together (to reduce inter-vertebral motion) which simulated disk degeneration. Additionally, progressive levels of injury simulating increased vertebral motion (modeling, for example, herniated discs) could be detected from vibration signatures 9,10 . Unfortunately, this invasive approach has limited clinical application. In addition to model experiments, a previous clinical study using surface-based accelerometers was able to demonstrate that signals from different pathologies are unique in human twins 12 . In this study, five sets of discordant twins showed unique vibration signals in contrast to five sets of concordant twins, where signals overlapped. For example, pairs of healthy twi...