VECTOR SHORTAATENCY SOMATOSENSORY=EVOKEDPOTENTIALS Kaji and S ~ m n e rrecently ~ . ~ described the very interesting vector short latency somatomasensory-evoked potentials (SSEP), corresponding to the 3-channel Lissajous’ trajectory (3-CLT) described earlier.233,7They used the bipolar leads of 3-CLT to show, among other findings, that multiple simultaneous generators may contribute to SSEP peaks. The complex nature of 3-CLT leaves some doubt about the strength of the conclusions that can be drawn on its basis. T h e three components of the 3-CLT vector are formed by the potential differences that exist between the two electrodes of each of three bipolar leads, placed at right angles to one another. Each component of the vector describes changes in voltage in one direction. As the authors correctly state, the three-dimensional trace resulting from this method does not represent the anatomical path of conduction of the impulses, but the endpoint of a voltage vector instead. T h e origin of the vector is assumed to remain fixed. All sources of electrical activity that contribute to any of the three bipolar potential differences, regardless of size, number, origin, or orientation, will contribute to the three-dimensional vector. T w o or more simultaneous processes will become visible as a single vector. This line of reasoning suggests that 3-CLT will not always succeed in separating the effects of multiple, simultaneous generators. This is illustrated in Figure 1; vector plots are shown for rightsided, leftsided, and bilateral simultaneous median nerve SSEPs obtained in a healthy subject. T h e two unilateral vector plots show mirrorimaged vectors for the N9 peak, compatible with different generators for each side. Bilateral stimulation must have resulted in bilaterally active generators, but the 3-CLT method added their influences, resulting in a single equivalent vector instead of two vectors. Kaji and Sumner introduced the use of two sets of electrode positions, and thus of two different 3-CLT plots, in an attempt to overcome this problem: one used electrodes around the upper, and one around the lower

Letters to the Editor

part of the neck.435T h e N 13 peak had different vector orientations in the two 3-CLT plots: anteriorly in the lower set, and upward and anteriorly in the upper set.’ T h e conclusion that the N 1 3 peak consists of an N13a and an N13b peak, which represented different generators. Can this conclusion be drawn with certainty from this finding? T h e two vectors differed mainly in the vertical direction. This information was derived from the two vertical leads. These were a CVII to middle cervical lead (lower 3-CLT), and a middle cervical to Oz lead (upper 3-CLT). These two leads form a bipolar chain of leads, such as is commonly used in EEC. A synchronous wave with different polarity in adjacent leads in such a chain is said to show “phase reversal.”” This is usually not explained as pointing to separate dipoles, but rather to a single dipole, which is localized near the electrode around which reversal takes place. A peak with phase reversal will result in differing vector orientations in the two vector plots. T h e differing N13 vector orientation may be explained by assuming two generators, but just as easily by assuming a single generator near the middle cervical electrode. Other evidence provided by Kaji and S ~ m n e r ,not ~ based on 3-CLT, suggested that N 13 does have multiple generators; doubt remains as to whether this conclusion can be based on 3-CLT, however. 3-CLT plots have been used mostly to study the brain-stem auditory evoked potential (BAEP), whose generators lie closely together in or near the brainstem, and also closely to the region where the three electrode axes intersect.'^^.^ T h e effect of non-centrally located dipoles on 3-CLT is unknown””; the effect of using leads that d o not run through a common origin is also unknown. When a single dipole is analyzed with more than one set of electrode positions, dipole eccentricity in at least one of the sets must be the result. A difference in polarity or at least in vector magnitude is a likely consequence. Different vector orientations may be merely a consequence of the use of different electrode positions. T h e distribution of potential fields over the surface


August 1991


400 uV E - 2


400 u V E - 2

Mx C

FIGURE 1. Three-dimensional vector plots of the SSEP N9 peak. An SSEP was recorded in a healthy subject using three leads that were used as input for a 3D vector plot: CVll to front of neck (X), left to right side of the neck at the CVll level (Y), and a CVll to middle cervical lead (Z). These leads con~ . ~ resulting 3D vector is enclosed in a cube, form to the lower vector set of Kaji and S ~ r n n e r . The centered around the vector origin. This is then plotted in a perspective view to help visualization of the three-dimensional wave form. Projections of the 3D-waveform are drawn on the sides of the cube as an extra aid in visualisation. Three such SSEPs were recorded following stimulation of the median nerves: right sided (Figure lA), left sided (Figure l B ) , and bilateral (Figure 1C). The X, Y, and 2 traces and 3 0 waveforms are plotted for all three. Only a 7 to 14 ms period was drawn in the 3D views. The N9 vector (indicated with N, at 10.8 ms) consists mainly of a horizontal vector, pointing to opposite sides following right- and leftsided stimulation. Bilateral stimulation does not show both vectors as active simultaneously: The two Y-traces largely cancel one another instead, and an unimposing 3D-waveform is the result. (The software used to produce this figure was written in FORTRAN on a Nicolet Pathfinder 11; source codes may be obtained from the author.)

of the body is a complex matter and, considering the potential between any two electrodes as resulting from a dipole situated near their connecting line, may not d o full justice to this complexity. Other techniques, such as dipole localization*," and mapping,' represent other attempts to solve this problem. These techniques rest on


Letters to the Editor

other assumptions that 3-CLT, but they do give more attention to the complexity of potential distributions. J. Gert van Dijk, MD Dept. of NeurologyiClinical Neurophysiology Leiden University Hospital 2300 RC Leiden, The Netherlands


August 1991

1. Desmedt JE. Topographic mapping of generators of soma-

tosensory evoked potentials. In: Maurer K. ed. Topographic Brain Mapping of EEG and Evoked Potentials. Springer Verlag, New York 1989:76-89. 2. Jewett DL, Martin WH, Sininger YS, Gardi JN: The %channel Lissajous' trajectory of the auditory brain-stem response. I. Introduction and overview. Electroenceph C l i ~ t Neurophysiol 1987 ;68 :3 2 3- 3 2 6. 3 . Jewett DL: The 3-channel Lissajous' trajectory of the auditory brain-stem response. IX. Theoretical aspects. Electroenceph Clin Neurophysiol 1987;68:386- 408. 4. Kaji R, Sumner AJ: Bipolar recording of short-latency somatosensory evoked potentials after median nerve stimulation. Neurology 1987;37:410-4 18. 5. Kaji R, Sumner AJ: Vector short-latency somatosensoryevoked potentials after median nerve stimulation. Muscle Nerve 1990;13: 1 174- 1 182. 6. Niedermeyer E: The EEG signal: Polarity and field determination. I n : Niedermeyer E, Lopes da Silva E', etls. l < l w t r o r n c r p h a l o ~ ~ ~Hasic ~ ~ ) / zPrifici/)lP.\, ~. (;/iJiicuf Appiic(ilio)i\ nnd Related Fic.1d.S. Urban and Schwarzenberg. Ualtimore, Munich 3'387; 79-83. 7. Pratt H , Har'el Z,

Vector short-latency somatosensory-evoked potentials.

LETTERS TO THE EDITOR VECTOR SHORTAATENCY SOMATOSENSORY=EVOKEDPOTENTIALS Kaji and S ~ m n e rrecently ~ . ~ described the very interesting vector sho...
625KB Sizes 0 Downloads 0 Views