Validity of the EMG signal

(Kinesiological) EMG has good logical validity (i.e. it measures the presence/absence of striated muscle activity), but technical expertise is required. EMG recording has to be differentiated from artifacts, which is very straightforward for intramuscular recordings (particularly if also amplified as an acoustic signal). In surface recordings, the artefacts are much more difficult to sort out. With surface electrodes there may be a problem of changing contact quality, particularly with prolonged recordings, creating a variability of recording quality. For detection with surface EMG there are studies claiming sound reliability and clinical predictive validity for intra-anal electrodes. Content validity of kinesiological EMG recording implies a continuous recording from the same defined source; needle electrodes may become dislodged, therefore intramuscular wire electrodes are much more intrinsically reliable for long-term recordings. The content validity of recordings with surface-type electrodes depends on the type of electrode, and the possibility of their movement (displacement). A particular problem is content validity of repeated EMG recordings, which should sample the same source; this is intrinsically better for surface recordings, which are less selective. On the other hand, content validity of surface recordings may be questioned if the source of EMG activity is claimed to be only one of several muscles in the vicinity of the electrodes.

With surface-type electrodes the overall anatomical source of the EMG signal in the pelvic region is often uncertain – is the EMG really derived only from the muscle which is claimed as the source? The other relevant issue is the question of representativeness – is the EMG signal really representative for the muscle or muscle group for which it is being claimed to be representative of? In other words, for all electrode types, content validity needs to be established for the physiological relevance of the particular source recording. Thus, for example, anal sphincter recording may not be conclusive for urethral sphincter behaviour. The kinesiological EMG (as obtained during polygraphic urodynamic recording) has accepted diagnostic validity to detect detrusor/striated sphincter dyssynergia, but this has not been formally researched much and probably holds true particularly for intramuscular recordings for the urethral sphincter. Indeed, the test has not yet been standardized. Its sensitivity and specificity are not known, but are far from ideal. The logical and content validity of CN EMG to diagnose muscle reinnervation is good, but is usually not discussed in these terms.

The diagnostic validity of CN EMG to detect striated muscle denervation and reinnervation is generally accepted. CN EMG sensitivity and specificity to detect moderate to severe denervation and reinnervation is accepted as good. These statements are supported by a large body of experience with nerve and muscle lesions as defi ned clinically, electrophysiologically and histopathologically. The sensitivity and specificity to diagnose changes of reinnervation may vary for different types of CN EMG signal analysis. SFEMG has good logical and content validity, and good diagnostic validity to detect changes due to reinnervation, but does not seem to be used clinically for PFM.

Responsiveness

A technically good EMG recording is capable of demonstrating absence of electrical activity in a non-active muscle (‘electrical silence’) and a graded response to increasing muscle activation. There is no difficulty in detecting even small differences in EMG activity from a given source with a good (technically reliable) technique.

Reliability

The consistency and reproducibility of results of diagnostic EMG (quantitative techniques using concentric and single fibre needle electrodes) is accepted as good if performed by experienced testers. Extensive experience is needed for either method, possibly even more for CN EMG. Straightforward parameters from surface EMG recordings (presence/absence of muscle activation) are much easier to interpret than CN EMG, and the consistency and reproducibility of such recordings (if technical issues are solved) are good. The overall consistency and reproducibility of results of the kinesiological EMG as a tool to investigate physiology lies more with the reproducibility of the ‘physiology’ that is being assessed (i.e. reproducibility of muscle ‘behaviour’).