Evoked potential compression using AOTLC and DPCM

Neuro-electric signals, such as evoked potentials (EPs), have been widely used to quantify the conditions of neurological system. In applications like telemedicine, it is necessary to efficiently transmit the EP signals. Data compression has been widely used for other non-biomedical signals, such as speech coding and image compression. In this paper, we study and evaluate the performances of EP compression using two different methods. One is the differential pulse code modulation (DPCM) method, which is a waveform-based compression technique, the other is the adaptive orthogonal transform linear combiner (AOTLC), which is based on an explicit model of EP using orthogonal transform theory and an adaptive filter. Analysis and computer simulation show that the AOTLC can achieve a higher compression ratio than DPCM, and it also gives a robust performance for noise-contaminated EP signals. The criterion used for performance evaluation is the ability for the compressed EP to preserve the latency change information.