rIntraoperative neurophysiological monitoring of the spinal cord
ical deficits). On the other hand, D wave changes – when recordable – have proven to be the strongest predictors of maintained corti-
cospinal tract integrity (and therefore, of motor function/recovery). Combining the use of muscle MEPs with D wave recordings provides
methodologies are significantly changing intraoperative
neurophysiological monitoring (IOM) of the spinal cord.
(Nash et al., 1977; Engler et al., 1978). Early enthusiasm
was stemmed by the presence of serious motor deficits
despite preserved SEPs. Thus, their capacity to monitor
the spinal cord motor tracts began to be questioned.
* Corresponding author. Tel.: +1 212 636 3281; fax: +1 212 636 3159.
E-mail address: vdeletis@chpnet.org (V. Deletis).
gythe most comprehensive approach for assessing the functional integrity of the spinal cord motor tracts during surgery for intramedullary
spinal cord tumors. However, muscle MEPs may suffice to assess motor pathways during other spinal procedures and in cases where the
pathophysiology of spinal cord injury is purely ischemic. Finally, while MEPs are now considered the gold standard for monitoring the
motor pathways, SEPs continue to retain value as they provide specificity for assessing the integrity of the dorsal column. However, we
believe SEPs should not be used exclusively – or as an alternative to motor evoked potentials – during spine surgery, but rather as a
complementary method in combination with MEPs. For intramedullary spinal tumor resection, SEPs should not be used exclusively
without MEPs.
 2007 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
Keywords: Intraoperative monitoring; Spinal cord surgery; Spine surgery; Motor evoked potentials; D wave
1. Introduction
Recent advances in technology and neurophysiological
Approximately 30 years have passed since somatosen-
sory evoked potentials (SEPs) were first used to monitor
the spinal cord during surgical correction for scoliosisduring spinal cord and spine surgery: A review focus
on the corticospinal tracts
Vedran Deletis a,*, Francesco Sala b
a Institute for Neurology and Neurosurgery, Beth Israel Medical Center – Singer Division, 170 East End Avenue, Room 311, New York, NY 10128, USA
b Section of Neurosurgery, Department of Neurological Sciences and Vision, University Hospital, Piazzale Stefani, 1, 3721 Verona, Italy
Accepted 7 September 2007
Abstract
Recent advances in technology and the refinement of neurophysiological methodologies are significantly changing intraoperative
neurophysiological monitoring (IOM) of the spinal cord. This review will summarize the latest achievements in the monitoring of the
spinal cord during spine and spinal cord surgeries. This overview is based on an extensive review of the literature and the authors’ per-
sonal experience. Landmark articles and neurophysiological techniques have been briefly reported to contextualize the development of
new techniques. This background is extended to describe the methodological approach to intraoperatively elicit and record spinal D
wave and muscle motor evoked potentials (muscle MEPs). The clinical application of spinal D wave and muscle MEP recordings is crit-
ically reviewed (especially in the field of Neurosurgery) and new developments such as mapping of the dorsal columns and the cortico-
spinal tracts are presented. In the past decade, motor evoked potential recording following transcranial electrical stimulation has
emerged as a reliable technique to intraoperatively assess the functional integrity of the motor pathways. Criteria based on the
absence/presence of potentials, their morphology and threshold-related parameters have been proposed for muscle MEPs. While the
debate remains open, it appears that different criteria may be applied for different procedures according to the expected surgery-related
morbidity and the ultimate goal of the surgeon (e.g. total tumor removal versus complete absence of transitory or permanent neurolog-Invited
Clinical Neurophysiolo1388-2457/$32.00  2007 International Federation of Clinical Neurophysiolo
doi:10.1016/j.clinph.2007.09.135eview
www.elsevier.com/locate/clinph
119 (2008) 248–264gy. Published by Elsevier Ireland Ltd. All rights reserved.

uropLesser et al. presented the first report in the literature of
a patient who suffered post-operative paraplegia despite
SEP preservation (Lesser et al., 1986). This extensively-
cited paper combines data dealing with this phenomenon
from 6 different centers. A review of the original papers,
which Lesser’s work is based on, shows that at least one
patient (from Ginsburg et al., 1985) had what he believed
were post-operatively preserved SEP recordings even while
the patient was paraplegic. Following Lesser et al., reports
in the literature continued to show the presence of pre-
served SEPs at the end of surgery in patients who suffered
post-operative paraplegia (Zornow et al., 1990; Minahan
et al., 2001; Jones et al., 2003; Pelosi et al., 1999). Most
of the patients from these four studies had anterior spinal
artery syndrome, which only affected the vascular territory
of the anterolateral column of the spinal cord. This is a pri-
mary factor that influences the preservation of the dorsal
column, which is a major contributor in the generation of
SEPs. Furthermore, preserved SEPs, coupled with the loss
of muscle motor evoked potentials (muscle MEPs), have
been documented in surgery for intracranial blood vessels
(Neuloh and Schramm, 2002; Szelenyi et al., 2003).
Lesions to the corticospinal tracts (CTs) or dorsal col-
umns can inadvertently occur as well during surgery for
intramedullary spinal cord pathology. If a lesion occurs
when approaching the intramedullary tumor during the ini-
tial dorsal myelotomy, SEPs can completely disappear, a
drawback that compromises the use of SEPs as the only
neurophysiological monitoring method for this type of sur-
gery (Kalkman et al., 1994; Deletis, 1999).
These limitations supported the opinions of some neuro-
surgeons who had come to believe that SEPs were useless
during surgery of the spinal cord. Although that statement
sounded rather extreme, it was worthy of notice given that
the recording of SEPs was the only available monitoring
technique at that time. In the last 10 years, however, the
development of specific methods for monitoring the func-
tional integrity of the CTs has opened a new field in the
monitoring of the motor tracts during spine and spinal
cord surgery. Two specific methods are: (a) Recordings of
the D wave from the spinal cord, and (b) Recordings of
motor evoked potentials from the limb muscles. The com-
bined use of these techniques has become useful in prevent-
ing surgically induced injury to the spinal cord and in
predicting post-operative motor outcome (Kothbauer
et al., 1997, 1998; MacDonald and Janusz, 2002). Today,
methods for monitoring the dorsal and the lateral columns
(CT) of the spinal cord are so refined that each long tract
can be monitored individually, with changes correlating
highly with post-surgical neurological outcome.
It should be mentioned that the terminology ‘‘false neg-
ative SEPs’’ is often inappropriately used to indicate the
occurrence of a post-operative motor deficit in spite of
unchanged intraoperative SEPs. A SEP result should be
labeled as ‘‘false negative’’ only when post-operative sen-
V. Deletis, F. Sala / Clinical Nesory deficits occur and were not predicted by intraoperative
SEP changes. Similarly, a MEP result should be labeled as‘‘false negative’’ if the patient wakes up with a new or wors-
ened motor deficit in spite of intraoperatively unchanged
MEPs. This review will summarize the latest achievements
in the monitoring of the spinal cord functional integrity
during surgery of the spine and spinal cord, as well as the
introduction of further developments in this field. This
review is written by a clinical neurophysiologist (VD) and
a neurosurgeon (FS) in order to give readers a balanced
view of the role that intraoperative monitoring plays in
the prevention and documentation of intraoperative injury
to the spinal cord.
2. A brief history of intraoperative monitoring (IOM) of the
spinal cord
Because of the previously mentioned disadvantages of
using SEPs as the only monitoring method during surgery
of the spine and spinal cord, a couple of different methods
have been developed with the hope of better evaluating the
functional integrity of the spinal cord’s long pathways. In
the course of developing the clinical use of these methods,
it has been shown that most of them cannot evaluate CT
functional integrity because they are not specific to the fast
neurons in the CT, which are essential elements for execu-
tion of precise voluntary movements.
2.1. Spinal cord-to-spinal cord technique
In this method, the spinal cord is stimulated with an epi-
dural catheter-type electrode, then the elicited compound
potentials are recorded over the spinal cord (Tamaki
et al., 1985, 1986). Stimulation can be performed cranially
and recorded caudally or vice versa. Due to the different
conduction properties (velocities) of the spinal cord path-
ways, the recorded potentials take the form of two distinct
waves. The recorded potentials are very robust and most
likely represent the combined activity of the dorsal columns
(DCs), the CTs, and other tracts of the spinal cord. It has
been claimed that one of these two waves belonged to the
DC and the other to the CT. Unfortunately, there is not
enough evidence to verify this statement. Koyanagi et al.,
could not find a clear correlation between results using this
method and the clinical outcome of 20 patients who under-
went surgery for intramedullary spinal cord tumors
(Koyanagi et al., 1993). This monitoring method has been
largely abandoned today, but it retains some value when
there is severe preexisting neuropathology, or in research
settings, when determination of the extent of conduction
may be important.
2.2. Spinal cord-to-peripheral nerve (‘‘Neurogenic MEPs’’)
On the basis of Machida’s work (Machida et al., 1985),
Owen and colleagues introduced a method that electrically
stimulates the spinal cord with trans-laminally-placed elec-
hysiology 119 (2008) 248–264 249trodes in order to elicit potentials that are then recorded
from the peripheral nerves (Owen et al., 1991). They named