11/14/2020 1:07:25 PM
Digital Antidromic SNAP Are NOT A Violation Of Volume Conductor Theory: A requested explnation
Digital Antidromic SNAPs Are NOT an Exception to Volume Conductor Theory
Please Note: In another post on “Atypical Waveform”, a question arose about antidromic SNAPs and I was requested to address why an antidromically recorded SNAP from upper limb digits is NOT an exception to volume conductor theory, i.e. it is biphasic initially negative and not triphasic initially positive as would be predicted for a propagating action potential. Rather than extend that particular post, I am initiating this post to address the previously noted issue as requested.
Specifically, it is well recognized that a propagating action potential (nerve or muscle) on a sufficiently long enough segment of excitable tissue can be represented as a tripole (+-+) and even better a quadrupole (+- -+) [see Dumitru et al: Electrodiagnostic Medicine 2nd Ed: Electrical Sources and Volume Conduction, pp 27-67]. Another way of saying this is that a negative sink (-) is preceded and followed by positive source currents “feeding” the negative sink. The associated current/voltage distribution associated with the three “poles” extends out into the volume conductor with the leading and trailing source currents extending out in front of, and behind the negative sink respectively thereby representing the “leading/trailing dipole: +- -+” model.
A recording electrode located at some location away from the propagating action potential will initially detect the leading positive source current by describing a downward or positive deflection on the instrument’s screen [Dumitru et el: Electrodiagnostic Medicine 2nd Ed: Figs. 2-5 & 2-6, pp 32-33]. This positive source voltage extends some distance physically both radially away from, and longitudinally in front of/behind, the negative sink. When the negative sink arrives at the recording electrode an upward or negative spike is described by the instrument. Further action potential propagation results in the recording electrode now detecting the terminal positive source current again defining a downward or positive deflection smaller in magnitude, but somewhat longer in duration than the leading positive spike (areas are the same) thereby defining the typical triphasic positive/negative/positive configuration of a propagating action potential traveling toward and passing by an electrode in a volume conductor. There are no exceptions to this theoretical construct.
Similarly, if an action potential is initiated at a location coincident with the recording electrode, then it is at the initiation site of the negative sink and the instrument first describes the upward negative deflection associated with the negative sink. With propagation, the action potential departs the electrode location in both directions with the terminal positive source currents now detected with the instrument defining a downward or positive deflection. The end result is a biphasic initially negative waveform. This is true for all waveforms initiated at a location coincident with the recording electrode (biphasic initially negative) to include: end-plate spike, CMAP, MUAP, Fibrillation potential, Myotonic Discharge, CRD, Fasciculation, etc. There are no exceptions to this fundamental aspect of action potential propagation and volume conductor theory.
There is, however, an “apparent exception”. Specifically, if you record an antidromic SNAP from the third digit, for example, subsequent to median nerve activation at the wrist, you will indeed record a biphasic initially negative waveform [Dumitru et el: Electrodiagnostic Medicine 2nd Ed: Figs. 2-7, pp 34]. Recall, volume conductor theory states that any waveform with an initial negative onset must originate at the recording electrode and propagate away. However, it is clear that the antidromic SNAP does not originate at the recording electrode. Rather, it originates away from the recording electrodes, that is at the wrist, and then propagates toward the two ring electrodes consistently describing the familiar biphasic initially negative waveform. Isn’t this either a violation, or at the very least, an exception to standard volume conductor theory? It sure seems like it is. Specifically, as noted above, when an action potential approaches, reaches, and then passes the recording electrode it must produce a triphasic waveform. The antidromic SNAP approaches, reaches, and passes the recording electrode to be sure, but generates a biphasic initially negative waveform. Herein lies the conundrum
Well, indeed no, it is not a violation of volume conductor theory. In order to understand what is going on one must conceptualize the voltage field distribution associated with the propagating action potential as it extends out into the volume conductor both longitudinally and radially. Imagine if you will, that the nerve is in the center of a cylinder filled with normal saline. Now, imagine the radius of the cylinder is reduced symmetrically about the center line [Dumitru et el: Electrodiagnostic Medicine 2nd Ed: Figs. 2-6, pp 33]. As the cross-sectional area of the cylinder declines, the associated voltage field lines become compressed radially and “stretched or distended” longitudinally. Imagine squeezing a water balloon. As you squeeze the balloon in one direction, the balloon expands in the direction perpendicular to the force applied, i.e. it grows longitudinally. If one were to reduce this imaginary cylinder’s radius until it was just a few millimeters thick around the nerve, then the action potential’s associated voltage field distribution would be concentrated around the nerve radially as well as extending some distance more longitudinally. If one were to locate two electrodes along the nerve within relative close proximity to each other (say 4 cm), then the approaching compressed positive source voltages would essentially be detected simultaneously at both electrodes and with very similar magnitudes. The compression of the longitudinal field doesn’t leave much voltage difference between these compressed isopotential positive voltage lines with respect to the distance within the spatial confines of our recording electrodes, i.e. 4 cm. In other words, both E-1 and E-2 detect virtually the same positive source voltages because they are comparatively close together with respect to the difference between the actual voltages of the compressed isopotential lines. Since E-1 and E-2 record very similar positive voltages, they are eliminated through differential amplification. So, no potential is described on the instrument, and the baseline remains flat.
However, the voltages transition between the leading positive source and subsequent negative sink, unlike the similar positive source voltages at the two recording electrodes, is detected. This is because the time frame of sodium activation is comparatively abrupt and manifests at
E-1 while E-2 still records the leading positive source current. This difference in voltage leads to the detection of the abrupt onset of the negative sink and manifest on the instrument’s screen as a rapid upward, or negative spike since E-1 is now coincident with the negative sink while E-2 is not. Just like the leading source voltages, the negative sink also has a spatial expanse. The spatial expanse of the leading source currents extends a considerable distance out in front of the negative sink (theoretically it’s along the entire volume conductor ahead of the sink), while that of the negative sink is rather confined. In other words, the negative sink is “sandwiched” between the leading and trailing source voltages (+ - +; or equivalently: +- -+). We can calculate the spatial expanse of the negative sink with respect to how much of the nerve it approximately occupies relevant to the parameter of interest to us as electromyographers: i.e. the SNAP’s amplitude. Amplitude, of course, is important because its magnitude is an approximate representation of the number of axons excited. This issue of amplitude and number of axons activated in a sensory nerve can be quite different from that of the CMAP: why?
So, if we are interested in how much distance along the nerve the negative sink occupies with respect to its maximum amplitude, then we need to know how long it takes from negative spike onset to its maximal peak. That time is close to about 0.8 ms. If the nerve is guestimated to propagate at roughly 50 M/s, then the distance of the negative sink from its initiation to maximum peak is about 4 cm. This is where that 4 cm number comes from. You want your two recording electrodes to be at least if not greater than 4 cm to avoid differential amplification from truncating the SNAP’s amplitude thereby yielding a possible false positive axonal loss lesion without demyelination (amplitude decline without latency prolongation). The onset latency will not be affected if the E-1 and E-2 difference are closer than 4 cm (E-1 remains a constant distance from the cathode), but the peak will be shortened (waveform stopped from reaching it maximal amplitude) and the amplitude reduced. Hence, no demyelination but supposed axonal loss.
There are several natural outcomes of the preceding issue of inter-electrode separation. Axons are certainly traveling faster than 50 M/s which has consequences of measuring the waveform’s true amplitude. Specifically, if the fastest axons are traveling at 60 M/s, then the optimal inter-electrode separation is 4.8 cm (NCV=D/T); and the 4 cm distance is too close which will result in differential amplification truncating the response magnitude. On the other hand, if a patient with a peripheral neuropathy and has a CV of 40 M/s, then 3.2 cm of inter-electrode separation is tolerable. The saving grace in all of this, is the concept of replicating the exact conditions under which the original reference population data was obtain. Again, we do not have “normal” values but rather “reference” values. If you use the exact filter, gain, inter-electrode separation, distance, temperature, etc. as that of the original data collection, then you should be just fine. If not, then your data cannot be compared to that of whatever reference population you are using. This is why it is critical for investigators to clearly define the instrument/environmental parameters under which they collected their data so it can be used by others.
Ok, now let’s return to reality. The sensory nerves in the digit occupy a relatively small region of the volume conductor compressed between the skin and underlying bone. Hence, the anatomy “squeezes” the propagating action potential’s voltage distribution as noted above.
E-1 and E-2 both record virtually the same voltage simultaneously associated with the approaching action potential. As a result, nothing is detected because of differential amplification. Then, the action potentials’ negative sinks arrive at E-1 but not yet at E-2. There is an ensuing initially upward negative spike with subsequent propagation and termination at the fingertip, thereby describing a biphasic initially negative and NOT triphasic initially positive waveform as would be predicted by theory. But wait, there’s more! You can accept the above explanation and move on, or you can rightfully expect some evidence for all this theory stuff. Fortunately, this can all be “proved” quite easily. First, record the typical antidromic SNAP with your two ring electrodes placed as noted above [Dumitru et el: Electrodiagnostic Medicine 2nd Ed: Figs. 2-7, pp 34]. You then obtain the anticipated biphasic initially negative SNAP. Now, quite simply, move E-2 to the fifth digit, stimulate the median nerve again; low and behold the waveform now becomes larger (less differential amplification) and clearly displays an initial positive deflection fully supporting volume conductor theory. All is well! (An alternative approach would be to move E-2 further along the finger until such a distance is reached so as to manifest the small voltage differences between E-1 and E-2 pursuant to the initial source current. The problem is that you would need a really really long digit.)
Now you can also ask, why do I get an initial positive deflection (triphasic waveform) when doing orthodromic or mixed nerve studies: e.g. stimulating the proper digital nerves and recording the response from the wrist? Well you do most times but not always. It depends how much subcutaneous tissue is present so as to permit the radial expansion of the source currents ahead of the negative sink allowing them to manifest differently at the two recording electrodes. The more subcutaneous tissue present, the more the source currents expand radially and the less compressed longitudinally, thereby permitting their associated voltages to be recorded somewhat differently and hence now observed. This concept applies to all recordings whether using bar, ring, or separate disc electrodes. Volume conductor theory is preserved.
[Note: I hope this helps. I apologize for the lengthy explanation, but this is rather complicated stuff and I am attempting to address this issue at the request of a reader. I tried to include practical examples which also leads to a more lengthy, but hopefully more clear explanation. Also, I could not include figures and hence only refer to them because of copyright issues.]