The Author Responds

We thank Drs. Journeay and Kumbhare for their comments and for the opportunity to clarify the points they raise in their letter to the editor. Although we acknowledge that surface electromyography has limitations, as do most physiological measures in human participants, the literature shows overwhelming evidence that differences exist in a number of electromyographic variables between individuals with low back disorders and healthy controls, supporting the potential value of surface electromyography in low back assessments. The point that we do completely agree with, however, is that the lack of standardization in electromyographic studies has resulted in difficulties comparing studies. This is exactly what we are attempting to address in our studies. We have described protocol details in our publications, including standardized electrode placements, amplitude normalization procedures, validation exercises, and controlled movement parameters along with signal averaging and pattern recognition techniques, that have been shown to improve the reliability of electromyographic signals. We conduct training and reliability testing among testers for all studies and have reported similar findings across data sets from a variety of healthy groups at different time points using this standardized protocol. Studies that use similar standardized protocols report more consistent results. For example, the study cited in Journeay and Kumbhare's letter to the editor reported lower interrater intraclass correlation coefficients (ICCs) than the intrarater as expected; however, the lowest values reported are for zero crossing measurements, which are not relevant to the present study. The amplitude measures at the 2 sites (iliocostalis and multifidus) for the exercises that were more controlled and closely related to the present study reported interrater ICCs ranging from 0.87 to 0.94, whereas intrarater ICCs were from 0.94 to 0.98. The lowest values reported for amplitude measures were for the balance exercises. They most likely do not represent reliability issues related to the electromyographic recordings themselves, rather the inconsistent task performance between days as suggested by the authors. To address the point on validity of the multifidus muscle surface recordings, the same authors cited their earlier 2001 study, indicating that high reliability was found for the mutifidus muscle surface recordings using their standardized protocol for electrode placement and normalization. Furthermore, although the Stokes et al article is often cited as evidence of weak relations between surface and indwelling sites, they reported high R² values (0.8–0.9) between surface and indwelling multifidus electrode recordings for specific tasks, with the lower values reported for the conditions when the forces were produced at angles and the electromyographic amplitudes were low. More recent work shows a strong linear relation (R²=.96) between the multifidus electromyographic intensity and magnetic resonance imaging intensity values. Other researchers and our previous work illustrate that this lower, more medial site exhibits both unique amplitude and temporal ^,  patterns that are distinct from other back sites that are positioned more lateral or at a higher spinal level. Therefore, consistent with electrophysiology and the volume properties of the tissues involved, electrodes over these sites are picking up the motor unit action potentials from the underlying muscles at that site, acknowledging that the depth of pick-up would preclude measures from deep fibers using electromyographic surface recordings.