Trial-to-trial latency variability of somatosensory evoked potential in early stage of cervical myelopathy

Somatosensory evoked potentials (SEP) have been widely used in monitoring spinal cord function during surgery. However, its clinical diagnostic value was limited because of big test-retest variability in SEP measurement. This study applied constrained second order based blind source separation (CSOB-BSS) algorithm to extract single trial SEP. This novel method provides a new measurement of trial-to trial latency variability (TTLV) of SEP. Ten healthy subjects and ten patients with cervical myelopathy (CM) at early stage were recruited in this study. Latency in averaging SEP did not show significant difference between healthy and CM. However, median SEP showed significant lower TTLV (4.9±1.6%) in healthy subjects in comparison with TTLV (10.8±1.0%) in CM (p<;0.05). To investigate the meaningful interpretation of TTLV, clinical measures of disease severity including modified Japanese Orthopaedic Association (mJOA) and transverse area ratio between spinal cord to canal in MRI were performed in CM patients. There is no correlation among TTLV, mJOA and MRI measurement. Results suggested the diagnostic value of TTLV to discriminate CM from healthy subjects in early stage, as well, TTLV provides different features in CM other than functional evaluation in mJOA and anatomic features in MRI.