Efficient extraction of data from intra-operative evoked potentials: 1.-Theory and simulations

Quickly and efficiently extracting evoked potential information from noise is critical to the clinical practice of intraoperative neurophysiologic monitoring (IONM). Currently this is primarily done using trained professionals to interpret averaged waveforms. The purpose of this paper is to evaluate and compare multiple means of electronically extracting simple to understand evoked potential characteristics with minimum averaging. A number of evoked potential models are studied and their performance evaluated as a function of the signal to noise level in simulations.

which extract the least number of parameters from the data are least sensitive to the effects of noise and are easiest to interpret. The simplest model uses the baseline evoked potential and the correlation receiver to provide an amplitude measure. Amplitude measures extracted using the correlation receiver show superior performance to those based on peak to peak amplitude measures. In addition, measures of change in latency or shape of the evoked potential can be extracted using the derivative of the baseline evoked response or other methods. This methodology allows real-time access to amplitude measures that can be understood by the entire OR staff as they are small, dimensionless numbers of order unity which are simple to interpret. The IONM team can then adjust averaging and other parameters to allow for visual interpretation of waveforms as appropriate.