Reflex threshold assessment methodology for evaluation of central sensitisation is vulnerable to EMG crosstalk

M.B. JENSEN; K.S. FRAHM; J.A. BIURRUN MANRESA; O.K. ANDERSEN

Aarhus

Encuentro; 2012 Joint Annual Meeting of the Scandinavian Association for the Study of Pain (SASP) and the Danish Pain Society (DSF); 2012

Background/aims: The nociceptive withdrawal reflex (NWR) is a robust measure of nociception highly correlated with subjective pain and several studies indicate that identification of reflex thresholds may be a valuable objective tool during diagnosis of chronic pain conditions related to central sensitisation. The NWR involves complex muscle synergies but most methodologies for assessment of human withdrawal reflexes evaluates surface electromyography (sEMG) measured over just one muscle and does not consider the possible interference of crosstalk from adjacent muscles. This study investigated if standardised reflex thresholds are robust with respect to crosstalk.Methods: Reflexes were elicited by electrical stimulation at the sole of the foot and measured using sEMG and intramuscular electromyography (iEMG) from the tibialis anterior and soleus muscles in 15 healthy subjects. 489 sEMG recordings were evaluated using a standardised reflex threshold parameter (interval peak z-score) proposed by Rhudy and France [1] to detect occurring reflexes. The outcome of this standardised reflex detection were compared to two different visual evaluations: (1) assessing only sEMG and (2) assessing both sEMG and iEMG (enabling distinction between genuine reflexes and crosstalk) for the production of two respective ROC-curves.Results: The first ROC-curve indicated that interval peak z-score allows reflex detection with high accuracy and has an optimal threshold around z = 12. However, the second ROC-curve based on evaluation 2 revealed that many recordings identified as reflexes were incorrectly categorised reflecting crosstalk. Crosstalk did severely reduce the accuracy of the performed reflex detection ? a z-score threshold of 12 was found associated with an extremely poor specificity (0.26).Conclusion: Crosstalk may severely reduce the accuracy of standardised reflex threshold assessment methodologies. Use of standard sEMG for reflex recording may cause an alarmingly low specificity of reflex detection and should be avoided whenever the presence crosstalk is suspected.