RESEARCH ARTICLE
Spreading Depression in Continuous Electroencephalography of Brain Trauma Jed A. Hartings, PhD,1,2,3 J. Adam Wilson, PhD,1 Jason M. Hinzman, PhD,1 Sebastian Pollandt, MD,4 Jens P. Dreier, MD, PhD,5 Vince DiNapoli, MD, PhD,1,3 David M. Ficker, MD,4 Lori A. Shutter, MD,1,2,3,4 and Norberto Andaluz, MD1,2,3 Objective: Cortical spreading depolarizations are a pathophysiological mechanism and candidate target for advanced monitoring in acute brain injury. Here we investigated manifestations of spreading depolarization in continuous electroencephalography (EEG) as a broadly applicable, noninvasive method for neuromonitoring. Methods: Eighteen patients requiring surgical treatment of traumatic brain injury were monitored by invasive electrocorticography (ECoG; subdural electrodes) and noninvasive scalp EEG during intensive care. Spreading depolarizations were first identified in subdural recordings, and EEG was then examined visually and quantitatively to identify correlates. Results: A total of 455 spreading depolarizations occurred during 65.9 days of simultaneous ECoG/EEG monitoring. For 179 of 455 events (39%), depolarizations caused temporally isolated, transient depressions of spontaneous EEG amplitudes to 57% (median) of baseline power. Depressions lasted 21 minutes (median) and occurred as suppressions of high-amplitude delta activity present as a baseline pattern in the injured hemisphere. For 62 of 179 (35%) events, isolated depressions showed a clear spread of depression between EEG channels with delays of 17 minutes (median), sometimes spanning the entire hemisphere. A further 188 of 455 (41%) depolarizations were associated with continuous EEG depression that lasted hours to days due to ongoing depolarizations. Depolarizations were also evidenced in EEG as shifts in direct current potentials. Interpretation: Le~ ao’s spreading depression can be observed in clinically standard, continuous scalp EEG, and underlying depolarizations can spread widely across the injured cerebral hemisphere. These results open the possibility of monitoring noninvasively a neuronal pathophysiological mechanism in a wide range of disorders including ischemic stroke, subarachnoid hemorrhage, and brain trauma, and suggest a novel application for continuous EEG. ANN NEUROL 2014;76:681–694
C
linical treatment of stroke and traumatic brain injury (TBI) is limited by the lack of methods to monitor pathologic mechanisms of secondary injury. Presently, intracranial pressure is the only cerebral variable that is widely monitored in clinical practice, but remains controversial and is used mainly in TBI. Other monitoring modalities are available but are not widely used due to their invasive nature and uncertain clinical value. Nonetheless, advances in neuromonitoring are required to identify pathophysiological mechanisms that are active in individual patients who might benefit from neuroprotective therapies. In epilepsy and cardiology, for
instance, continuous electrophysiologic monitoring allows diagnosis of disease subtype, pathologic localization, and evaluation of treatment effects. In stroke, the concept of the ischemic penumbra was originally developed on the basis of electrophysiologic monitoring in animals.1,2 As currently understood, the core ischemic region experiences a loss of spontaneous electrical activity followed by mass terminal depolarization, which develops in a spreading fashion within minutes of arterial occlusion. The penumbral tissue surrounding the core may also experience electrical silencing and loss of function, but here cellular membrane
View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.24256 Received Jun 5, 2014, and in revised form Aug 15, 2014. Accepted for publication Aug 19, 2014. Address correspondence to Dr Hartings, 231 Albert Sabin Way, ML0517, Department of Neurosurgery, University of Cincinnati, Cincinnati, OH 45267. E-mail: [email protected] From the 1Department of Neurosurgery, University of Cincinnati College of Medicine, Cincinnati, OH; 2Neurotrauma Center at University of Cincinnati Neuroscience Institute, Cincinnati, OH; 3Mayfield Clinic, Cincinnati, OH; 4Department of Neurology, University of Cincinnati College of Medicine, e University Medicine Berlin, Berlin, Cincinnati, OH; and 5Center for Stroke Research, Departments of Experimental Neurology and Neurology, Charit Germany.
C 2014 American Neurological Association V 681
ANNALS
of Neurology
polarization is initially preserved. Over time, transient mass depolarizations arise spontaneously and periodically and spread throughout the penumbra and surrounding, normally perfused tissue. These spreading depolarizations eventually become prolonged and evolve to terminal depolarizations, recruiting the vulnerable penumbra into the expanding ischemic core.3,4 Thus, electrophysiologic characterization of time-dependent lesion growth yielded the concept of secondary injury and raised the possibility of neuroprotective interventions. The phenomena of transient and terminal spreading depolarizations of cerebral cortex were discovered in the 1940s by A. P. P. Le~ao.5,6 When a strong stimulus was applied to rabbit cortex, he described a “spreading depression of activity in the cerebral cortex,” indicating loss of function, accompanied by a negative shift of the direct current (DC) potential, indicating mass tissue depolarization. It is now understood that the spreading depression of activity is a consequence of the spreading depolarization. Since 2002, these same electrophysiological methods have been applied clinically to investigate spreading depolarizations in the pathophysiology of acute brain injury in humans.7,8 In patients requiring craniotomies, electrode strips are placed on the cortical surface for subsequent electrophysiologic monitoring during intensive care.9 As in animals, frequent spreading depolarizations are observed as propagating shifts of DC potential and spreading depression of spontaneous cortical activity.10–12 In patients with aneurysmal subarachnoid hemorrhage (SAH), clusters of repetitive spreading depolarizations are associated with delayed cerebral ischemia, characterized by new neurologic deficits and cortical infarcts.13,14 In TBI, the occurrence of spreading depolarizations is an independent predictor of poor clinical outcome.11,15 Accumulating evidence thus suggests that monitoring of spreading depolarizations could have similar value in acute brain injury as electrophysiological tools have in epilepsy and cardiology, because real-time diagnosis of secondary injury becomes possible. However, the present requirement for invasive procedures has limited the application of such monitoring to a small minority of patients. Therefore, here we combined invasive monitoring with scalp electroencephalography (EEG) to determine whether spreading depolarizations could be detected noninvasively in TBI patients. Our results suggest new applications and methods for continuous EEG.
Patients and Methods Eighteen TBI patients were enrolled at the University of Cincinnati Medical Center in an observational study registered at ClinicalTrials.gov (number NCT00803036). Inclusion criteria
682
were the clinical decision for craniotomy and age 18 years. Patients with fixed, dilated pupils were excluded. The study was approved by the institutional review board, and written informed consent for research was obtained from patients’ legally authorized representatives. At the conclusion of surgery for lesion evacuation or decompression, an electrode strip was placed on the surface of viable perilesional cortex judged to be at highest risk for secondary injury.9,10,15 Placement thus varied by injury location. After surgery, patients were transferred to the intensive care unit, where continuous electrocorticography (ECoG) and simultaneous scalp EEG were performed. All patients were managed in accordance with the local standard of care for TBI and the Guidelines for Management of Severe Traumatic Brain Injury.16 Prophylactic antiseizure drugs were administered for 7 days after injury, with levetiracetam as the first choice. Midazolam was preferred for long-term sedation in patients requiring mechanical ventilation. A complete list of drugs is provided in Table 1. Intracranial pressure (ICP) was monitored and recorded hourly in 17 patients. ECoG recordings were terminated and electrode strips were removed at the bedside by gentle traction when invasive neuromonitoring was no longer clinically required or after a maximum of 7 days.
Electrophysiology Monitoring The subdural ECoG strip contained 6 platinum electrodes with 10mm spacing between contacts (Wyler; Ad-Tech Medical, Racine, WI). For 13 patients, electrodes were connected in a sequential bipolar fashion to an alternating current (AC)coupled amplifier (0.01Hz high-pass cutoff, GT205; Guger Technologies, Graz, Austria) and data were digitized and recorded with Powerlab 16/SP and LabChart software (ADInstruments, Bella Vista, Australia). In 5 patients, monopolar signals were recorded with a DC-coupled amplifier (g.USBamp, Guger Technologies) using a platinum subdermal needle (Natus Neurology–Grass Products, Warwick, RI; Astro-Med, West Warwick, RI) at the mastoid as reference. Data were displayed and acquired using a custom multimodal monitoring program based on BCI2000, a brain–computer interface tool written in C11 to support a variety of data acquisition systems.17,18 Scalp EEG recordings were obtained in the intensive care unit as clinical standard of care for management of severe TBI. Sixteen electrodes (Ag/AgCl-impregnated plastic; Ives EEG Solutions, Newburyport, MA) were placed bilaterally according to the international 10–20 electrode system, and data were acquired in a double-banana bipolar montage with a Grass (Natus Neurology) amplifier system. The high-pass cutoff was 0.5Hz, which allowed assessment of EEG signals only in the conventional frequency ranges. In 5 patients, EEG leads were also jumped to the g.USBamp DC-coupled amplifier for simultaneous data acquisition with ECoG, which also permitted assessment of full-band EEG activity.
Data Processing and Analysis Spreading depolarizations were identified in ECoG recordings according to methods previously described.10,15 Briefly, the
Volume 76, No. 5
November 2014
21
23
74
44
36
38
57
62
46
65
37
29
27
29
30
74
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Fall
MVA
Assault
MVA-P
Fall
Fall
Fall
MC
0
2
2
7
15
3
7
15
9
3
8
13
10
6
6
2
2
1
2
2
2
2
2
0
2
1
0
2
N/A 0
3
3
Unknown 6
Fall
GSW
MVA
Fall
Fall
MVA
MVA
Assault
R
R
R
R
L, R
R
L T Cont, SDH
R Fr ICH
L SDH
Bilat Fr Cont (L > R)
Bilat Cont
L Fr Cont, SDH
L SDH, Cont
R Fr Cont, SDH
R SDH
R Fr Cont, ICH
L
R
L
Bif
Bif
L
L
R
R
R
L MCA/ACA infarct L
R SDH
Bilat SDH (R > L)
R Fr Cont, SDH
EDH, SAH
Bilat Fr Cont
Bilat Fr Cont (R > L)
L
No
No
No
No
No
No
No
Yes
No
No
No
No
No
No
No
No
No
No
1.3
0.5
1.4
1.7
0.8
2.4
0.9
1.0
1.0
1.1
2.8
0.8
0.3
0.9
0.6
3.3
0.6
0.6
4.9
6.2
4.9
6.5
0.9
2.0
3.4
4.4
2.8
3.9
3.0
4.9
2.4
2.2
5.3
2.8
1.9
3.7 Ph, Lev
Ph, Lev
Pr, Fe
Pr, Mi, Fe
Pr, Fe, Lor
Fe, Lor
Pr, Mi, Fe
Fe
Mi, Fe
Pr, Fe
Mi, Fe
Pr, Mi, Fe
Pr, Mi, Fe
Pr, Mi, Fe
Pr, Mi, Fe
Fe
Pr, Mi, Lor
1
71
0
134
26
1
54
12
4
126
59
0
Lev
7
Ph, Lev, Lor 122
Lev
Lev
Ph, Lev, Lor 0
Lev
Lev
Lev
Lev, Lor
Lev
Lev
Lev
Ph, Lev
34
1
19
Sev
Sev
Good
Veg
Mod
Sev
Sev
Sev
Veg
Sev
Dead
Sev
Sev
Dead
Sev
Veg
Mod
Good
Antiseizure SDs 6-Month in Outcome ECoG by GOS
Pr, Mi, Fe, Lor Ph, Lev
Mi, Fe
Pr, Mi, Fe
Analgesia/ Craniotomy Bone Flap EEG/ EEG/ Location Replacement ECoG ECoG Sedation Start, Duration, Days Days
GCS and pupillary reactivity were assessed at admission to the study hospital following resuscitation. For pupils, 2 5 both reactive, 1 5 1 reactive, 0 5 neither. The number of SDs observed throughout the whole duration of ECoG recordings is given. ACA 5 anterior cerebral artery; Bif 5 bifrontal; Bilat 5 bilateral; Cont 5 contusion; ECoG 5 electrocorticography; EDH 5 epidural hematoma; EEG 5 electroencephalography; F 5 female; Fe 5 fentanyl and its analogues; Fr 5 frontal; FT 5 frontotemporal; GCS 5 Glasgow Coma Scale; GOS 5 Glasgow Outcome Score; GSW 5 gunshot wound; ICH 5 intracerebral hematoma; L 5 left; Lev 5 levetiracetam; Lor 5 lorazepam; M 5 male; Mi 5 midazolam; MC 5 motorcycle; MCA 5 middle cerebral artery; Mod 5 moderate; MVA 5 motor vehicle accident; MVA-P 5 pedestrian involved in motor vehicle accident; N/A 5 not applicable; P 5 parietal; Pr 5 propofol; Ph 5 phenytoin or fosphenytoin; R 5 right; SAH 5 subarachnoid hemorrhage; SD 5 spreading depolarization; SDH 5 subdural hematoma; Sev 5 severe disability; T 5 temporal; TBI 5 traumatic brain injury; Veg 5 vegetative state.
F
F
M
M
M
F
F
M
M
M
M
F
F
F
F
M
M
10
43
L FT Cont, SDH
2
2
8
MVA-P
23
1
F
GCS Pupil Lesions Reactive
No. Age, Sex Cause yr of TBI
TABLE 1. Demographics and Clinical Characteristics
Hartings et al: Spreading Depression in EEG
683
ANNALS
of Neurology
signature of spreading depolarization is a negative shift of the DC potential, approximately 5 to 15mV in amplitude in recordings with platinum electrodes and DC amplifiers.14,17,19,20 With AC amplifiers using 0.01Hz high-pass cutoff, the DC shift appears as a slow-potential change of 1 to 5mV peak-to-peak amplitude in the near-DC (6 hours. This pattern of continuous, prolonged EEG depression during repetitive spreading depolarizations was also evidenced in 3 patients (Nos. 8, 11, 15) from the onset of EEG recordings. In these cases, depolarizations were ongoing at the start of EEG recordings and therefore the onset of EEG depression with the first depolarization was not witnessed. However, in each case, the EEG Volume 76, No. 5
November 2014
R FP
RT
R FP
R Sup P
LP
R FP
R FB
LF
RF
L FP
R FP
LO
2
4
5
6
8
9
11
12
15
16
17
18
T7-P7, P7-O1
Fp2-F4, F4-C4, C4-P4, P4-O2, Fp2-F8, F8-T8, T8-P8
Fp1-F3, F3-C3, C3-P3, P3-O1
Fp2-F4, F4-C4
Fp1-F3
Fp2-F4
Fp2-F4, F4-C4
F3-C3, C3-P3, P3-O1, F7-T7, T7-P7
Fp2-F4, F4-C4, C4-P4, P4-O2, T8-P8, P8-O2
Fp2-F8
—
—
EEG Channels with Depressions
5
115
Yes (28)
Yes (2)
125
45
1
16
83
24
36
1
3
1
ECoG SDs, No.b
Yes (7)
No
No
No
No
Yes (6)
Yes (15)
No
—
—
Spread in EEG (No.)a
2
66
45
3
1
5
19
10
27
1
—
—
Isolated
—
11
5
42
—
11
14
14
—
—
—
—
Continuous
—
21
24
—
—
—
45
—
1
—
—
—
Intermediate
EEGc
—
16
49
—
—
—
3
—
8
—
—
—
Artifact
3
1
2
—
—
—
2
—
—
—
3
1
No Change
Indicates whether depression of EEG activity was observed as spreading between 2 or more channels for any ECoG-identified spreading depolarizations; the number is given in parentheses. b Number of ECoG-identified spreading depolarizations during period of simultaneous EEG/ECoG monitoring. For each patient, the number of depolarizations with each type of EEG manifestation is given. c Isolated types induced EEG depression and were scored for depression depth and duration. Continuous types exhibited sustained depression from a prior depolarization. Intermediate types were associated with only partial recovery from prior depolarizations and no clear further depression of amplitudes. Recordings were sometimes contaminated with artifact, precluding assessment. For some depolarizations, no EEG changes were observed. ECoG 5 electrocorticography; EEG 5 electroencephalography; F 5 frontal; FB 5 frontobasal; FP 5 frontoparietal; L 5 left; O 5 occipital; P 5 parietal; R 5 right; SD 5 spreading depolarization; Sup 5 superior; T 5 temporal.
a
ECoG Strip Location
Patient No.
TABLE 2. Summary of EEG Manifestations of Spreading Depolarizations
Hartings et al: Spreading Depression in EEG
685
ANNALS
of Neurology
FIGURE 1: Isolated, continuous, and intermediate depressions of electroencephalographic (EEG) amplitude induced by spreading depolarizations. (A) Eight hours of time-compressed recordings from Patient 8 illustrate the 3 types of EEG depressions for spreading depolarizations labeled 1 to 7. Depressions are seen in raw traces C3-P3 and T7-P7, but not in other channels of the ipsilateral hemisphere or corresponding contralateral channels. Depolarizations #6 and #7 each produce unique, isolated depressions in EEG channels C3-P3 and T7-P7 followed by full amplitude recovery before the next depolarization. Depolarization #2 illustrates a continuous-type depression, because the EEG depression initiated by depolarization #1 is continued when the second occurs 29 minutes later. Depolarizations #4 and #5 illustrate intermediate cases in which EEG amplitude recovery from prior depolarizations is only partial and a relative depression is maintained. Power integrals for channels C3-P3 and T7-P7 are shown, illustrating the depression periods and spreading nature (arrows) of depolarizations #6 and #7. (B) Representative traces from time points marked with red dashed lines in A, shown on an expanded time scale (10 seconds each). Note that the 2 EEG channels in which depressions are observed have high baseline amplitudes with prominent polymorphic delta activity. In the depressed state, these waves are suppressed such that amplitudes are similar to other EEG channels. ECoG 5 electrocorticogram. (C) Pie chart shows the relative proportion of all 455 ECoG-identified spreading depolarizations with EEG manifestations characterized as the 3 types illustrated in A.
amplitude recovered when repetitive depolarizations stopped. An example is shown in Figure 2B. In total, 97 of 455 ECoG-identified depolarizations (21%) were associated with continuous EEG depressions sustained by multiple depolarizations. Intervals between depolarizations with continuous-type EEG depressions (33 minutes; IQR 5 28–38) were significantly shorter than intervals preceding depolarizations with isolated-type EEG depressions (53 minutes; IQR 5 34–90; p < 0.001, Mann–Whitney U test [M-W]). The third category was intermediate between isolatedtype and continuous-type EEG depressions. In these instances, which accounted for 91 of 455 ECoG-identified depolarizations (20%), recovery of EEG amplitude from a prior depolarization was only partial. Therefore, maximal depression was not considered to be continuing, but distinct EEG depressions could not be identified as clearly associated 686
with individual depolarizations (see Fig 1A, examples 4, 5). In some cases, modulations of EEG amplitude were evident but too poorly defined to be scored. These depolarizations with intermediate EEG depression occurred at intervals (median 5 33 minutes, IQR 5 22–45) similar to continuous-type events but significantly shorter than the isolated type (p < 0.001, M-W). For 76 depolarizations, the EEG signal, and therefore the power integral, was contaminated by highamplitude artifact from patient movement, nursing procedures, or changes in electrode impedance. This noise precluded observation and/or scoring of the events. In a further 12 cases, no EEG depressions were observed to be induced by the depolarizations or to be continuing from previous ones. To determine whether depressions of EEG amplitude might be better identified in monopolar recordings, Volume 76, No. 5
Hartings et al: Spreading Depression in EEG
FIGURE 2: Slow potential changes accompany amplitude depressions as electroencephalographic (EEG) signatures of spreading depolarizations. (A) A 3.3-hour epoch from Patient 15 shows isolated depressions and direct current (DC) shifts in scalp EEG traces for 2 spreading depolarizations. Electrocorticography (ECoG): ECoG traces show spreading shifts of the DC potential (circles) and depressions of high-frequency (HF; 0.5–50Hz) activity. EEG: Full-band (DC) referential recordings from EEG electrode F4 show corresponding DC shifts (circles) and high-pass filtered (0.5Hz) traces show isolated amplitude depressions in Fp2-F4 and F4-C4, but not other ipsilateral channels. (B) DC shifts are also observed in EEG during continuous depression. From the start of EEG recordings in the same patient, 42 spreading depolarizations were identified in ECoG occurring at intervals of 34 minutes (median; interquartile range 5 29–38) over a period of 27 hours. B shows 7.5 hours spanning the end of this period, marked by the asterisk; channels and conventions are the same as in A. During repetitive ECoG-identified spreading depolarizations, clear EEG amplitude depressions could not be identified, but EEG amplitudes increased in channels Fp2-F4 and F4-C4 with the onset of polymorphic delta activity when depolarizations stopped. Correspondingly, DC recordings from Fp2 (not shown) and F4 showed an unstable baseline with slow potential fluctuations during ECoG-identified depolarizations, with a marked transition to a stable baseline when depolarizations stopped and high-frequency amplitudes increased. The epileptologist report on continuous EEG findings states that “1–4Hz frequency delta background is seen. At times, it appears to be relatively suppressed.” These suppressions are clearly the isolated (A) and continuous (B) depressions induced by spreading depolarizations. Scale bars apply to all traces of the same type.
depression magnitudes were compared between bipolar and monopolar (referential) channel derivations in the 7 patients with at least 3 isolated-type depolarizations. Amplitudes were significantly more depressed in bipolar recordings, reaching 47% of baseline values, compared to monopolar recordings with depression to only 55% of baseline (p < 0.01, paired t test, n 5 3 per patient). EEG Depressions Are Suppressions of HighAmplitude Delta Activity Table 3 summarizes the EEG abnormalities noted in clinical reports, including epileptiform discharges and seizures. All patients exhibited focal slowing consisting of high-amplitude polymorphic delta activity. Depolarization-induced amplitude depressions of all 3 types were caused primarily by suppression of this highamplitude delta activity, rather than by suppression of a normal-amplitude baseline activity. In Figure 1, for instance, polymorphic delta activity elevates baseline amplitudes in several channels of the injured hemisphere November 2014
(eg, C3-P3, T7-P7), in contrast to lower amplitude signals of other ipsilateral (eg, Fp1-F3, P7-O1) and all contralateral channels. Spreading depolarizations were manifested as EEG depressions in these channels by suppressing the high-amplitude baseline such that, during maximal depression, amplitudes were similar across all channels. Accordingly, the degree of amplitude depression should positively correlate with the baseline EEG amplitude prior to depression; this was observed for 3 of 5 patients with >10 isolated-type depressions (R2 range 5 0.46–0.67, p’s < 0.001). Similar examples of this phenomenon for isolated-type depressions are shown in Figures 2A and 4B. Figure 2B further illustrates a case in which all ipsilateral channels have similar amplitude due to suppression of high-amplitude delta in Fp2-F4 and F4-C4 during a series of repetitive spreading depolarizations. When depolarizations cease, high-amplitude delta activity emerges, revealing the former state as depressed. Polymorphic delta was restricted to 1 to 2 channels in some 687
ANNALS
of Neurology
TABLE 3. EEG Abnormalities Described in Clinical Reports of Epileptologists
No.
Generalized Slowing
Focal Slowing
Epileptiform Discharges
EEG Seizures
1
Sev, Mod-Sev
L FT CPDA
L TP
2
Sev, Mod
R Hem CPDA
3
Sev
4
ECoG Seizures, No., Duration
Breach
Other
—
Yes
SIRPIDs
R FC
—
Yes
Bilat F delta
—
—
Yes
Sev
R FCT CPDA/IPDA
Freq R FCP
25 R FC
Yes
5
Sev, Mod
R Hem CPDA
—
—
Yes
6
Mod-Sev, Mod
R FCT CPDA
R FC
—
Yes
7
Sev, Mod, Mild
R Hem CPDA
—
—
Yes
8
Sev, Mod-Sev
L FT CPDA
Rare L C
—
Yes
9
Sev, Mod-Sev
PDA, R > L
Freq R FC, R CP, L FC
—
Yes
10
Sev, Mod, Mild
R FCT CPDA
Rare R CP
—
Yes
11
Sev, Mod
R FCT CPDA/IPDA
Freq R FC
Freq R F
12
Sev, Mod-Sev
L Hem CPDA
Rare L FC, CP
—
13
Mild-Mod
L Hem CPDA
Rare L C
Freq L F
14
Sev, Mild
Bilat F IPDA
—
—
Yes
15
Sev
R FCT CPDA
—
—
Yes
16
Sev
L FCP CPDA
Rare L CP
—
17
Sev
R Hem CPDA
R F, R P
—
Yes
PLEDs
18
Mod
L Hem IPDA
Int L CP
—
Yes
SIRPIDs
105, 221 minutes
BS
SIRPIDs
Yes Yes
161, 317 minutes
26, 103 minutes
Yes
PLEDs
Yes
Generalized slowing is described as severe, moderate, or mild based on the background frequency (delta, theta, or alpha) and the presence of reactivity. Focal slowing is described as CPDA and/or IPDA. Epileptiform discharges are described as frequent, intermittent, or rare. All EEG seizures were nonconvulsive, and all patients had breach rhythm. Other abnormalities described in reports are SIRPIDs, BS, and PLEDs. Number and duration of electrographic seizures in ECoG recordings are given for comparison. Bilat 5 bilateral; BS 5 burst suppression; C 5 central; CP 5 centroparietal; CPDA 5 continuous polymorphic delta activity; ECoG 5 electrocorticography; EEG 5 electroencephalography; F 5 frontal; FC 5 frontocentral; FCP 5 frontocentroparietal; FCT 5 frontocentrotemporal; Freq 5 frequent; FT 5 frontotemporal; Hem 5 hemispheric; Int 5 intermittent; IPDA 5 intermittent polymorphic delta activity; L 5 left; Mod 5 moderate; P 5 parietal; PDA 5 polymorphic delta activity; PLEDs 5 periodic lateralized epileptiform discharge; R 5 right; Sev 5 severe; SIRPID 5 stimulus-induced rhythmic, periodic, or ictal discharge; TP 5 temporoparietal.
patients, but was widespread through most of the injured hemisphere in others. EEG amplitude depressions were rarely observed in channels without high-amplitude delta activity, although high-amplitude delta was sometimes present in channels beyond those exhibiting depressions. The role of delta activity in observing spreading depolarizations in the EEG was confirmed in quantitative analysis. Baseline EEG activity prior to isolated-type depressions was dominated by delta (0.5–4.0Hz), which accounted for 89% of total power (Fig 5A). Thus, suppression of delta activity accounts for nearly all of the absolute magnitude of amplitude depression induced in 688
the EEG by spreading depolarizations. Furthermore, the relative suppression of delta activity was also greater than for higher frequency bands; delta was depressed to a mean 47% of baseline power, whereas higher bands had progressively less depression (see Fig 5B). Hemispheric Spreading Depression of EEG Amplitude In all 10 patients for whom EEG amplitude depressions were observed, these changes occurred in at least 1 of the bipolar EEG channels in closest proximity to the subdural ECoG strip (see Table 2). In 3 patients (Nos. 5, 11, 12), EEG Volume 76, No. 5
Hartings et al: Spreading Depression in EEG
FIGURE 3: Widespread propagation of spreading depolarization evidenced by spreading depression of continuous electroencephalography (EEG). (A) Schematic diagram and scout computed tomography (CT) show positions of cortical electrode strip and scalp electrodes in Patient 6 following partial resection of the right temporal lobe and evacuation of a right subdural hematoma. CT image shows diffuse subarachnoid hemorrhage and intraparenchymal hemorrhage in right frontal and temporal lobes. Electrocorticography (ECoG; black upper traces): In this 8.5-hour recording segment, 6 spreading depolarizations are observed as evidenced by slow potential changes (dashed circles) in the near–direct current (DC) ECoG recording that propagate from channel 2–3 to channel 4–5 and are accompanied by depression of 0.5 to 50Hz activity. EEG: Five bipolar EEG channels from the same hemisphere are shown. Colored traces correspond to the colored electrode positions in the head schematic diagram. Arrows indicate the clear amplitude depressions in channels Fp2-F4 and C4-P4 that occur in time-locked, 1to-1 correspondence with the ECoG-identified spreading depolarizations. Note that depressions occur first in C4-P4 and other posterior channels and are followed, after a substantial time lag of 30 to 40 minutes, by frontal Fp2-F4, indicating spreading depressions of EEG activity induced by spreading depolarization. The 2nd and 3rd depolarizations occurring in rapid succession are manifested as unique depression periods in Fp2-F4 but as a single continuous depression in C4-P4. Scale bars apply to all traces of the same type. (B) A 7.3-hour recording from Patient 17 is shown following the same conventions as in A. Electrode 1 (arrow) is seen in the CT scan following evacuation of a large right frontal intracerebral hematoma. Three spreading depolarizations are seen in ECoG recordings and each is accompanied by unique, isolated depressions in EEG channels throughout the hemisphere. Depressions begin in F4-C4 (blue), in near-synchrony with the electrode strip, and propagate posteriorly to P4-O2 (green) after delays of 57, 11, and 26 minutes. The changing delays suggest somewhat different origins for each depolarization wave.
amplitude depressions were restricted to a single channel. In the other 7 patients, spreading depolarizations induced EEG depressions in 2 to 6 channels, in some cases spanning broad expanses of the ipsilateral hemisphere. In 5 of these 7 patients, 62 of 179 isolated-type events (35%) showed a clear spread of the amplitude depression with a distinct time delay between depression onset in different channels. Examples from 2 patients are shown in Figure 3 (see also Fig 1, Fig 4A). For each depolarization, time intervals between nadirs of the amplitude depressions of the 2 channels with the clearest signals, and not sharing a common electrode, were measured. These intervals were 17 minutes (median; IQR 5 11–34) and did not significantly differ by the distance separating the electrode pairs. In 3 patients (Nos. 6, 16, 17), spreading depolarizations induced spreading EEG depressions in the most anterior and posterior channels (eg, Fp2-F4 and P4-O2), November 2014
indicating widespread propagation throughout the ipsilateral cerebral cortex. DC Potentials in EEG Shifts of the extracellular DC potential are a signature of spreading depolarization, reflecting the mirror image of sustained intracellular depolarization. In 4 patients with both spreading depolarizations and DC-coupled EEG recordings, monopolar EEG channels were examined for the presence of DC shifts accompanying depressions of the conventional frequency band (0.5–50Hz). In the majority of cases, DC shifts could not be identified. However, clear instances were found in 3 of 4 patients (Nos. 15, 17, 18) with peak-to-peak amplitudes of 708 6 242lV (n 5 37), 333 6 60lV (n 5 6), and 432lV (n 5 2), respectively. In Patient 15, DC shifts were 689
ANNALS
of Neurology
FIGURE 4: Further examples of EEG depression patterns. A and B show 5.25- and 3.5-hour epochs from Patients 17 and 16, respectively, during which no electrocorticographic recordings were obtained. In A, 5 recurring amplitude depressions are observed in channels Fp2-F4, F4-C4, C4-P4, and P4-O2 in a fashion similar to the example of Figure 3B from the same patient. In contrast to Figure 3B, depressions recur more frequently at 53- to 75-minute intervals and do not spread as far posteriorly. Spread is clearly observed as a march in depression from frontal to central to parietal locations (arrows). In B, 8 cycling amplitude depressions occur at 25- to 30-minute intervals in C3-P3 and, to a lesser extent, P3-O1. This represents a rare case in which depolarizations recurring at short intervals each produced an isolated-type depression instead of a single continuous depression. Unlike in A, no spread is evident in this case. Integrals are shown below.
observed during a cluster of 42 repetitive spreading depolarizations that spanned a total of 27 hours. During this time, high-frequency EEG activity was continuously depressed and the DC shifts appeared as a highly unstable, continuously fluctuating baseline potential (see Fig 2B). The DC shifts propagated between electrodes Fp2 and F4 with a delay of 9.3 minutes (mean; standard deviation 5 2.3, n 5 25). The cessation of depolarizations was then signaled by a marked EEG state change; the DC potential became steady and the amplitude of 0.5 to 50Hz activity increased with the onset of polymorphic delta activity. A similar fluctuating DC potential was observed in Patient 17 during a cluster of 11 depolarizations spanning 7 hours of continuous depression, although the DC shifts corresponding to individual depolarizations could not be deciphered.
Discussion Here we have shown that the vast majority of spreading depolarizations (>80%) recorded from the surface of cerebral cortex by ECoG are manifested as depressions of high-amplitude delta activity in clinically standard, noninvasive EEG recordings. When intervals between depolarizations are long enough, they induce unique 690
suppressions of this high-amplitude baseline such that amplitudes are normalized to levels in other head regions. These depressions, attributable to individual depolarization waves, develop slowly and may persist up to 1 hour before the baseline EEG amplitude is fully restored. When depolarizations occur repetitively at short intervals, the suppression of high-amplitude delta activity is maintained for longer periods, from hours to >1 day. We have further shown that Le~ao’s “spreading depression of activity”5 can be observed in scalp EEG, because many spreading depolarizations produced depressions that spread with a temporal delay between EEG channels. In some cases, affected channels were located far from the ECoG recording strip, thus providing the first evidence for widespread propagation of spreading depolarizations across the injured hemisphere. These results suggest a new application for continuous EEG by which a neuronal pathophysiological mechanism can be monitored as a possible etiology of neurologic condition and marker of ongoing secondary injury.
Spreading Depression in Scalp EEG From the time of Le~ao’s discovery in 1944, there was no evidence that spreading depolarizations could be observed Volume 76, No. 5
Hartings et al: Spreading Depression in EEG
area. Similarly, continuous depression periods are explained by the finding that multiple depolarizations, when recurring at short intervals, may occur simultaneously in the same recording area and thus maintain a continuous cycle of partial depression. The failure to detect Le~ao’s spreading depression through decades of EEG research is thus likely attributable to the long durations of development, persistence, and spread; recognition requires viewing of long recording epochs on highly compressed time scales (4 to 9cm/h) that historically have not been used in clinical neurophysiology. Routine 30-minute EEG examinations reviewed at 3cm/s preclude observation of events on this temporal scale.
FIGURE 5: Electroencephalographic (EEG) depressions are suppressions of delta activity. (A) Proportions of total power of baseline activity contributed by activity in the 4 frequency bands. Assessments were made prior to depolarization-induced EEG depression for 3 events in each of 7 patients. Error bars show standard deviations. (B) Degrees of depression as percentages of baseline power for each frequency band.
noninvasively from scalp EEG until the recent report of Drenckhahn et al.20 In part, this was due to the belief that the spatial wavelength of spreading depression across the cortex was too short to influence EEG signals recorded from a much larger volume of tissue. However, it is estimated that scalp potentials reflect the synchronous activity of at least 6 to 10cm2 across the brain’s surface, which may include twice that area of cerebral cortex.21,22 Because spreading depolarization travels at 1 to 8mm/min and depresses cortical activity for an average duration of 8 minutes,11,13 the typical spatial wavelength of depression is estimated at 0.8 to 6.4cm. Assuming a wavefront width of at least 1cm, and given that depressions can be arbitrarily long (hours), it is clear that depolarizations can simultaneously depress the requisite volume of cortex to be observed in EEG. Compared to intracranial electrodes, the larger recording area of scalp electrodes further explains the lesser degree of maximal EEG depression, the slower onset and recovery of depressions, and the longer total duration; only a subset of the total recording area is depressed at a given instance and more time is required for the wave to invade maximally and traverse the larger November 2014
Clinical Use of Continuous EEG As a measure of the brain’s essential electrophysiologic function, EEG has long offered unrealized promise as a tool for monitoring the progression of intracranial pathology.23,24 The advantages of EEG over other neuromonitoring techniques are that it is noninvasive and continuous and assesses both regional and global cerebral status. Furthermore, quantitative analysis can facilitate interpretation of the large volumes of data generated.25 Long-term trending of alpha variability and other metrics, for instance, have been investigated as predictors of delayed cerebral ischemia in SAH patients.26,27 Spreading depolarizations have now emerged as an important mechanism of lesion development and a target for EEG monitoring. In focal ischemia in the rat, a delayed phase of repetitive depolarization activity begins 6 to 10 hours after stroke onset and coincides with the period of delayed infarct maturation.28 Importantly, these depolarization clusters did not produce unique depressions of spontaneous EEG activity, but rather were manifested as a continuous suppression of high-amplitude polymorphic delta activity.29 The occurrence of continuous depression periods in EEG recordings during temporal clusters of spreading depolarizations was since confirmed in SAH patients.20 As in animals, such clusters have a delayed onset and high positive and negative predictive value for delayed cerebral ischemia.13,14 Here we have further shown in patients that both isolated and continuous depressions of EEG amplitude induced by spreading depolarizations represent suppressions of pathologic high-amplitude delta activity, a common feature of all types of acute brain injury. Increases in delta power result from preserved but abnormal synaptic activity due to structural damage or reduction of cerebral blood flow to a range of 15 to 22ml/100g/min. This increase is lost with suppression of synaptic activity during mass depolarization of cortex, which may indicate worsening ischemia because anoxic terminal spreading 691
ANNALS
of Neurology
depolarization develops at a threshold of 5 to 10ml/ 100g/min. However, spreading depolarization can invade cortex with normal or oligemic blood flow. In either case, the suppression of delta activity caused by depolarization paradoxically does not indicate resolution of abnormalities, but rather the occurrence of more severe electrophysiologic dysfunction. These translational studies thus raise the strong possibility that EEG monitoring of depolarizations could be used as a surveillance tool for progression of intracranial pathologies. In recent years, continuous EEG has gained traction and is performed as standard of care at an increasing number of hospitals. The rationale for continuous, multiday monitoring is primarily the evaluation of subclinical seizures as a possible etiology for altered mental status or coma, as well as belief that suppression of all early seizures, whether convulsive or nonconvulsive, may improve outcomes.30,31 Seizures and spreading depolarizations often occur with interacting patterns in the same tissue, and depolarizations may enhance epileptic activity by suppression of inhibitory neurotransmission.32–34 However, compared to seizures, spreading depolarizations have a 3- to 7-fold higher incidence in patients.19,35 The present findings thus suggest a new and complementary application for continuous EEG in acute brain injury. Implications for EEG Methodology In several respects, our results are similar to those reported by Drenckhahn et al in SAH patients.20 In that study, 47% of ECoG-identified spreading depolarizations induced isolated depressions in the scalp EEG, compared to 39% found here, and depressions reached 53% of baseline amplitude compared to 57% in the present study. In both studies, the remainder of depolarizations occurred at short intervals and did not induce unique depressions but rather maintained continuous depressions throughout temporal clusters of events, as noted above. Thus, these results are reproducible across different centers and diseases and represent general features of acute brain injury. Results differed somewhat in other respects. First, DC shifts were observed in scalp EEG for a majority (70%) of spreading depolarizations recorded by Drenckhahn et al, but were rarely observed in the 4 patients with DCcoupled EEG in the present study. Second, we found that both EEG depressions and DC shifts spread between scalp electrodes with a temporal delay, whereas all changes occurred synchronously in SAH patients. The wide spatial extent of spreading depression observed here was surprising and contrasts with the limited extent (
Spreading Depression in Continuous Electroencephalography of Brain Trauma Jed A. Hartings, PhD,1,2,3 J. Adam Wilson, PhD,1 Jason M. Hinzman, PhD,1 Sebastian Pollandt, MD,4 Jens P. Dreier, MD, PhD,5 Vince DiNapoli, MD, PhD,1,3 David M. Ficker, MD,4 Lori A. Shutter, MD,1,2,3,4 and Norberto Andaluz, MD1,2,3 Objective: Cortical spreading depolarizations are a pathophysiological mechanism and candidate target for advanced monitoring in acute brain injury. Here we investigated manifestations of spreading depolarization in continuous electroencephalography (EEG) as a broadly applicable, noninvasive method for neuromonitoring. Methods: Eighteen patients requiring surgical treatment of traumatic brain injury were monitored by invasive electrocorticography (ECoG; subdural electrodes) and noninvasive scalp EEG during intensive care. Spreading depolarizations were first identified in subdural recordings, and EEG was then examined visually and quantitatively to identify correlates. Results: A total of 455 spreading depolarizations occurred during 65.9 days of simultaneous ECoG/EEG monitoring. For 179 of 455 events (39%), depolarizations caused temporally isolated, transient depressions of spontaneous EEG amplitudes to 57% (median) of baseline power. Depressions lasted 21 minutes (median) and occurred as suppressions of high-amplitude delta activity present as a baseline pattern in the injured hemisphere. For 62 of 179 (35%) events, isolated depressions showed a clear spread of depression between EEG channels with delays of 17 minutes (median), sometimes spanning the entire hemisphere. A further 188 of 455 (41%) depolarizations were associated with continuous EEG depression that lasted hours to days due to ongoing depolarizations. Depolarizations were also evidenced in EEG as shifts in direct current potentials. Interpretation: Le~ ao’s spreading depression can be observed in clinically standard, continuous scalp EEG, and underlying depolarizations can spread widely across the injured cerebral hemisphere. These results open the possibility of monitoring noninvasively a neuronal pathophysiological mechanism in a wide range of disorders including ischemic stroke, subarachnoid hemorrhage, and brain trauma, and suggest a novel application for continuous EEG. ANN NEUROL 2014;76:681–694
C
linical treatment of stroke and traumatic brain injury (TBI) is limited by the lack of methods to monitor pathologic mechanisms of secondary injury. Presently, intracranial pressure is the only cerebral variable that is widely monitored in clinical practice, but remains controversial and is used mainly in TBI. Other monitoring modalities are available but are not widely used due to their invasive nature and uncertain clinical value. Nonetheless, advances in neuromonitoring are required to identify pathophysiological mechanisms that are active in individual patients who might benefit from neuroprotective therapies. In epilepsy and cardiology, for
instance, continuous electrophysiologic monitoring allows diagnosis of disease subtype, pathologic localization, and evaluation of treatment effects. In stroke, the concept of the ischemic penumbra was originally developed on the basis of electrophysiologic monitoring in animals.1,2 As currently understood, the core ischemic region experiences a loss of spontaneous electrical activity followed by mass terminal depolarization, which develops in a spreading fashion within minutes of arterial occlusion. The penumbral tissue surrounding the core may also experience electrical silencing and loss of function, but here cellular membrane
View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.24256 Received Jun 5, 2014, and in revised form Aug 15, 2014. Accepted for publication Aug 19, 2014. Address correspondence to Dr Hartings, 231 Albert Sabin Way, ML0517, Department of Neurosurgery, University of Cincinnati, Cincinnati, OH 45267. E-mail: [email protected] From the 1Department of Neurosurgery, University of Cincinnati College of Medicine, Cincinnati, OH; 2Neurotrauma Center at University of Cincinnati Neuroscience Institute, Cincinnati, OH; 3Mayfield Clinic, Cincinnati, OH; 4Department of Neurology, University of Cincinnati College of Medicine, e University Medicine Berlin, Berlin, Cincinnati, OH; and 5Center for Stroke Research, Departments of Experimental Neurology and Neurology, Charit Germany.
C 2014 American Neurological Association V 681
ANNALS
of Neurology
polarization is initially preserved. Over time, transient mass depolarizations arise spontaneously and periodically and spread throughout the penumbra and surrounding, normally perfused tissue. These spreading depolarizations eventually become prolonged and evolve to terminal depolarizations, recruiting the vulnerable penumbra into the expanding ischemic core.3,4 Thus, electrophysiologic characterization of time-dependent lesion growth yielded the concept of secondary injury and raised the possibility of neuroprotective interventions. The phenomena of transient and terminal spreading depolarizations of cerebral cortex were discovered in the 1940s by A. P. P. Le~ao.5,6 When a strong stimulus was applied to rabbit cortex, he described a “spreading depression of activity in the cerebral cortex,” indicating loss of function, accompanied by a negative shift of the direct current (DC) potential, indicating mass tissue depolarization. It is now understood that the spreading depression of activity is a consequence of the spreading depolarization. Since 2002, these same electrophysiological methods have been applied clinically to investigate spreading depolarizations in the pathophysiology of acute brain injury in humans.7,8 In patients requiring craniotomies, electrode strips are placed on the cortical surface for subsequent electrophysiologic monitoring during intensive care.9 As in animals, frequent spreading depolarizations are observed as propagating shifts of DC potential and spreading depression of spontaneous cortical activity.10–12 In patients with aneurysmal subarachnoid hemorrhage (SAH), clusters of repetitive spreading depolarizations are associated with delayed cerebral ischemia, characterized by new neurologic deficits and cortical infarcts.13,14 In TBI, the occurrence of spreading depolarizations is an independent predictor of poor clinical outcome.11,15 Accumulating evidence thus suggests that monitoring of spreading depolarizations could have similar value in acute brain injury as electrophysiological tools have in epilepsy and cardiology, because real-time diagnosis of secondary injury becomes possible. However, the present requirement for invasive procedures has limited the application of such monitoring to a small minority of patients. Therefore, here we combined invasive monitoring with scalp electroencephalography (EEG) to determine whether spreading depolarizations could be detected noninvasively in TBI patients. Our results suggest new applications and methods for continuous EEG.
Patients and Methods Eighteen TBI patients were enrolled at the University of Cincinnati Medical Center in an observational study registered at ClinicalTrials.gov (number NCT00803036). Inclusion criteria
682
were the clinical decision for craniotomy and age 18 years. Patients with fixed, dilated pupils were excluded. The study was approved by the institutional review board, and written informed consent for research was obtained from patients’ legally authorized representatives. At the conclusion of surgery for lesion evacuation or decompression, an electrode strip was placed on the surface of viable perilesional cortex judged to be at highest risk for secondary injury.9,10,15 Placement thus varied by injury location. After surgery, patients were transferred to the intensive care unit, where continuous electrocorticography (ECoG) and simultaneous scalp EEG were performed. All patients were managed in accordance with the local standard of care for TBI and the Guidelines for Management of Severe Traumatic Brain Injury.16 Prophylactic antiseizure drugs were administered for 7 days after injury, with levetiracetam as the first choice. Midazolam was preferred for long-term sedation in patients requiring mechanical ventilation. A complete list of drugs is provided in Table 1. Intracranial pressure (ICP) was monitored and recorded hourly in 17 patients. ECoG recordings were terminated and electrode strips were removed at the bedside by gentle traction when invasive neuromonitoring was no longer clinically required or after a maximum of 7 days.
Electrophysiology Monitoring The subdural ECoG strip contained 6 platinum electrodes with 10mm spacing between contacts (Wyler; Ad-Tech Medical, Racine, WI). For 13 patients, electrodes were connected in a sequential bipolar fashion to an alternating current (AC)coupled amplifier (0.01Hz high-pass cutoff, GT205; Guger Technologies, Graz, Austria) and data were digitized and recorded with Powerlab 16/SP and LabChart software (ADInstruments, Bella Vista, Australia). In 5 patients, monopolar signals were recorded with a DC-coupled amplifier (g.USBamp, Guger Technologies) using a platinum subdermal needle (Natus Neurology–Grass Products, Warwick, RI; Astro-Med, West Warwick, RI) at the mastoid as reference. Data were displayed and acquired using a custom multimodal monitoring program based on BCI2000, a brain–computer interface tool written in C11 to support a variety of data acquisition systems.17,18 Scalp EEG recordings were obtained in the intensive care unit as clinical standard of care for management of severe TBI. Sixteen electrodes (Ag/AgCl-impregnated plastic; Ives EEG Solutions, Newburyport, MA) were placed bilaterally according to the international 10–20 electrode system, and data were acquired in a double-banana bipolar montage with a Grass (Natus Neurology) amplifier system. The high-pass cutoff was 0.5Hz, which allowed assessment of EEG signals only in the conventional frequency ranges. In 5 patients, EEG leads were also jumped to the g.USBamp DC-coupled amplifier for simultaneous data acquisition with ECoG, which also permitted assessment of full-band EEG activity.
Data Processing and Analysis Spreading depolarizations were identified in ECoG recordings according to methods previously described.10,15 Briefly, the
Volume 76, No. 5
November 2014
21
23
74
44
36
38
57
62
46
65
37
29
27
29
30
74
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Fall
MVA
Assault
MVA-P
Fall
Fall
Fall
MC
0
2
2
7
15
3
7
15
9
3
8
13
10
6
6
2
2
1
2
2
2
2
2
0
2
1
0
2
N/A 0
3
3
Unknown 6
Fall
GSW
MVA
Fall
Fall
MVA
MVA
Assault
R
R
R
R
L, R
R
L T Cont, SDH
R Fr ICH
L SDH
Bilat Fr Cont (L > R)
Bilat Cont
L Fr Cont, SDH
L SDH, Cont
R Fr Cont, SDH
R SDH
R Fr Cont, ICH
L
R
L
Bif
Bif
L
L
R
R
R
L MCA/ACA infarct L
R SDH
Bilat SDH (R > L)
R Fr Cont, SDH
EDH, SAH
Bilat Fr Cont
Bilat Fr Cont (R > L)
L
No
No
No
No
No
No
No
Yes
No
No
No
No
No
No
No
No
No
No
1.3
0.5
1.4
1.7
0.8
2.4
0.9
1.0
1.0
1.1
2.8
0.8
0.3
0.9
0.6
3.3
0.6
0.6
4.9
6.2
4.9
6.5
0.9
2.0
3.4
4.4
2.8
3.9
3.0
4.9
2.4
2.2
5.3
2.8
1.9
3.7 Ph, Lev
Ph, Lev
Pr, Fe
Pr, Mi, Fe
Pr, Fe, Lor
Fe, Lor
Pr, Mi, Fe
Fe
Mi, Fe
Pr, Fe
Mi, Fe
Pr, Mi, Fe
Pr, Mi, Fe
Pr, Mi, Fe
Pr, Mi, Fe
Fe
Pr, Mi, Lor
1
71
0
134
26
1
54
12
4
126
59
0
Lev
7
Ph, Lev, Lor 122
Lev
Lev
Ph, Lev, Lor 0
Lev
Lev
Lev
Lev, Lor
Lev
Lev
Lev
Ph, Lev
34
1
19
Sev
Sev
Good
Veg
Mod
Sev
Sev
Sev
Veg
Sev
Dead
Sev
Sev
Dead
Sev
Veg
Mod
Good
Antiseizure SDs 6-Month in Outcome ECoG by GOS
Pr, Mi, Fe, Lor Ph, Lev
Mi, Fe
Pr, Mi, Fe
Analgesia/ Craniotomy Bone Flap EEG/ EEG/ Location Replacement ECoG ECoG Sedation Start, Duration, Days Days
GCS and pupillary reactivity were assessed at admission to the study hospital following resuscitation. For pupils, 2 5 both reactive, 1 5 1 reactive, 0 5 neither. The number of SDs observed throughout the whole duration of ECoG recordings is given. ACA 5 anterior cerebral artery; Bif 5 bifrontal; Bilat 5 bilateral; Cont 5 contusion; ECoG 5 electrocorticography; EDH 5 epidural hematoma; EEG 5 electroencephalography; F 5 female; Fe 5 fentanyl and its analogues; Fr 5 frontal; FT 5 frontotemporal; GCS 5 Glasgow Coma Scale; GOS 5 Glasgow Outcome Score; GSW 5 gunshot wound; ICH 5 intracerebral hematoma; L 5 left; Lev 5 levetiracetam; Lor 5 lorazepam; M 5 male; Mi 5 midazolam; MC 5 motorcycle; MCA 5 middle cerebral artery; Mod 5 moderate; MVA 5 motor vehicle accident; MVA-P 5 pedestrian involved in motor vehicle accident; N/A 5 not applicable; P 5 parietal; Pr 5 propofol; Ph 5 phenytoin or fosphenytoin; R 5 right; SAH 5 subarachnoid hemorrhage; SD 5 spreading depolarization; SDH 5 subdural hematoma; Sev 5 severe disability; T 5 temporal; TBI 5 traumatic brain injury; Veg 5 vegetative state.
F
F
M
M
M
F
F
M
M
M
M
F
F
F
F
M
M
10
43
L FT Cont, SDH
2
2
8
MVA-P
23
1
F
GCS Pupil Lesions Reactive
No. Age, Sex Cause yr of TBI
TABLE 1. Demographics and Clinical Characteristics
Hartings et al: Spreading Depression in EEG
683
ANNALS
of Neurology
signature of spreading depolarization is a negative shift of the DC potential, approximately 5 to 15mV in amplitude in recordings with platinum electrodes and DC amplifiers.14,17,19,20 With AC amplifiers using 0.01Hz high-pass cutoff, the DC shift appears as a slow-potential change of 1 to 5mV peak-to-peak amplitude in the near-DC (6 hours. This pattern of continuous, prolonged EEG depression during repetitive spreading depolarizations was also evidenced in 3 patients (Nos. 8, 11, 15) from the onset of EEG recordings. In these cases, depolarizations were ongoing at the start of EEG recordings and therefore the onset of EEG depression with the first depolarization was not witnessed. However, in each case, the EEG Volume 76, No. 5
November 2014
R FP
RT
R FP
R Sup P
LP
R FP
R FB
LF
RF
L FP
R FP
LO
2
4
5
6
8
9
11
12
15
16
17
18
T7-P7, P7-O1
Fp2-F4, F4-C4, C4-P4, P4-O2, Fp2-F8, F8-T8, T8-P8
Fp1-F3, F3-C3, C3-P3, P3-O1
Fp2-F4, F4-C4
Fp1-F3
Fp2-F4
Fp2-F4, F4-C4
F3-C3, C3-P3, P3-O1, F7-T7, T7-P7
Fp2-F4, F4-C4, C4-P4, P4-O2, T8-P8, P8-O2
Fp2-F8
—
—
EEG Channels with Depressions
5
115
Yes (28)
Yes (2)
125
45
1
16
83
24
36
1
3
1
ECoG SDs, No.b
Yes (7)
No
No
No
No
Yes (6)
Yes (15)
No
—
—
Spread in EEG (No.)a
2
66
45
3
1
5
19
10
27
1
—
—
Isolated
—
11
5
42
—
11
14
14
—
—
—
—
Continuous
—
21
24
—
—
—
45
—
1
—
—
—
Intermediate
EEGc
—
16
49
—
—
—
3
—
8
—
—
—
Artifact
3
1
2
—
—
—
2
—
—
—
3
1
No Change
Indicates whether depression of EEG activity was observed as spreading between 2 or more channels for any ECoG-identified spreading depolarizations; the number is given in parentheses. b Number of ECoG-identified spreading depolarizations during period of simultaneous EEG/ECoG monitoring. For each patient, the number of depolarizations with each type of EEG manifestation is given. c Isolated types induced EEG depression and were scored for depression depth and duration. Continuous types exhibited sustained depression from a prior depolarization. Intermediate types were associated with only partial recovery from prior depolarizations and no clear further depression of amplitudes. Recordings were sometimes contaminated with artifact, precluding assessment. For some depolarizations, no EEG changes were observed. ECoG 5 electrocorticography; EEG 5 electroencephalography; F 5 frontal; FB 5 frontobasal; FP 5 frontoparietal; L 5 left; O 5 occipital; P 5 parietal; R 5 right; SD 5 spreading depolarization; Sup 5 superior; T 5 temporal.
a
ECoG Strip Location
Patient No.
TABLE 2. Summary of EEG Manifestations of Spreading Depolarizations
Hartings et al: Spreading Depression in EEG
685
ANNALS
of Neurology
FIGURE 1: Isolated, continuous, and intermediate depressions of electroencephalographic (EEG) amplitude induced by spreading depolarizations. (A) Eight hours of time-compressed recordings from Patient 8 illustrate the 3 types of EEG depressions for spreading depolarizations labeled 1 to 7. Depressions are seen in raw traces C3-P3 and T7-P7, but not in other channels of the ipsilateral hemisphere or corresponding contralateral channels. Depolarizations #6 and #7 each produce unique, isolated depressions in EEG channels C3-P3 and T7-P7 followed by full amplitude recovery before the next depolarization. Depolarization #2 illustrates a continuous-type depression, because the EEG depression initiated by depolarization #1 is continued when the second occurs 29 minutes later. Depolarizations #4 and #5 illustrate intermediate cases in which EEG amplitude recovery from prior depolarizations is only partial and a relative depression is maintained. Power integrals for channels C3-P3 and T7-P7 are shown, illustrating the depression periods and spreading nature (arrows) of depolarizations #6 and #7. (B) Representative traces from time points marked with red dashed lines in A, shown on an expanded time scale (10 seconds each). Note that the 2 EEG channels in which depressions are observed have high baseline amplitudes with prominent polymorphic delta activity. In the depressed state, these waves are suppressed such that amplitudes are similar to other EEG channels. ECoG 5 electrocorticogram. (C) Pie chart shows the relative proportion of all 455 ECoG-identified spreading depolarizations with EEG manifestations characterized as the 3 types illustrated in A.
amplitude recovered when repetitive depolarizations stopped. An example is shown in Figure 2B. In total, 97 of 455 ECoG-identified depolarizations (21%) were associated with continuous EEG depressions sustained by multiple depolarizations. Intervals between depolarizations with continuous-type EEG depressions (33 minutes; IQR 5 28–38) were significantly shorter than intervals preceding depolarizations with isolated-type EEG depressions (53 minutes; IQR 5 34–90; p < 0.001, Mann–Whitney U test [M-W]). The third category was intermediate between isolatedtype and continuous-type EEG depressions. In these instances, which accounted for 91 of 455 ECoG-identified depolarizations (20%), recovery of EEG amplitude from a prior depolarization was only partial. Therefore, maximal depression was not considered to be continuing, but distinct EEG depressions could not be identified as clearly associated 686
with individual depolarizations (see Fig 1A, examples 4, 5). In some cases, modulations of EEG amplitude were evident but too poorly defined to be scored. These depolarizations with intermediate EEG depression occurred at intervals (median 5 33 minutes, IQR 5 22–45) similar to continuous-type events but significantly shorter than the isolated type (p < 0.001, M-W). For 76 depolarizations, the EEG signal, and therefore the power integral, was contaminated by highamplitude artifact from patient movement, nursing procedures, or changes in electrode impedance. This noise precluded observation and/or scoring of the events. In a further 12 cases, no EEG depressions were observed to be induced by the depolarizations or to be continuing from previous ones. To determine whether depressions of EEG amplitude might be better identified in monopolar recordings, Volume 76, No. 5
Hartings et al: Spreading Depression in EEG
FIGURE 2: Slow potential changes accompany amplitude depressions as electroencephalographic (EEG) signatures of spreading depolarizations. (A) A 3.3-hour epoch from Patient 15 shows isolated depressions and direct current (DC) shifts in scalp EEG traces for 2 spreading depolarizations. Electrocorticography (ECoG): ECoG traces show spreading shifts of the DC potential (circles) and depressions of high-frequency (HF; 0.5–50Hz) activity. EEG: Full-band (DC) referential recordings from EEG electrode F4 show corresponding DC shifts (circles) and high-pass filtered (0.5Hz) traces show isolated amplitude depressions in Fp2-F4 and F4-C4, but not other ipsilateral channels. (B) DC shifts are also observed in EEG during continuous depression. From the start of EEG recordings in the same patient, 42 spreading depolarizations were identified in ECoG occurring at intervals of 34 minutes (median; interquartile range 5 29–38) over a period of 27 hours. B shows 7.5 hours spanning the end of this period, marked by the asterisk; channels and conventions are the same as in A. During repetitive ECoG-identified spreading depolarizations, clear EEG amplitude depressions could not be identified, but EEG amplitudes increased in channels Fp2-F4 and F4-C4 with the onset of polymorphic delta activity when depolarizations stopped. Correspondingly, DC recordings from Fp2 (not shown) and F4 showed an unstable baseline with slow potential fluctuations during ECoG-identified depolarizations, with a marked transition to a stable baseline when depolarizations stopped and high-frequency amplitudes increased. The epileptologist report on continuous EEG findings states that “1–4Hz frequency delta background is seen. At times, it appears to be relatively suppressed.” These suppressions are clearly the isolated (A) and continuous (B) depressions induced by spreading depolarizations. Scale bars apply to all traces of the same type.
depression magnitudes were compared between bipolar and monopolar (referential) channel derivations in the 7 patients with at least 3 isolated-type depolarizations. Amplitudes were significantly more depressed in bipolar recordings, reaching 47% of baseline values, compared to monopolar recordings with depression to only 55% of baseline (p < 0.01, paired t test, n 5 3 per patient). EEG Depressions Are Suppressions of HighAmplitude Delta Activity Table 3 summarizes the EEG abnormalities noted in clinical reports, including epileptiform discharges and seizures. All patients exhibited focal slowing consisting of high-amplitude polymorphic delta activity. Depolarization-induced amplitude depressions of all 3 types were caused primarily by suppression of this highamplitude delta activity, rather than by suppression of a normal-amplitude baseline activity. In Figure 1, for instance, polymorphic delta activity elevates baseline amplitudes in several channels of the injured hemisphere November 2014
(eg, C3-P3, T7-P7), in contrast to lower amplitude signals of other ipsilateral (eg, Fp1-F3, P7-O1) and all contralateral channels. Spreading depolarizations were manifested as EEG depressions in these channels by suppressing the high-amplitude baseline such that, during maximal depression, amplitudes were similar across all channels. Accordingly, the degree of amplitude depression should positively correlate with the baseline EEG amplitude prior to depression; this was observed for 3 of 5 patients with >10 isolated-type depressions (R2 range 5 0.46–0.67, p’s < 0.001). Similar examples of this phenomenon for isolated-type depressions are shown in Figures 2A and 4B. Figure 2B further illustrates a case in which all ipsilateral channels have similar amplitude due to suppression of high-amplitude delta in Fp2-F4 and F4-C4 during a series of repetitive spreading depolarizations. When depolarizations cease, high-amplitude delta activity emerges, revealing the former state as depressed. Polymorphic delta was restricted to 1 to 2 channels in some 687
ANNALS
of Neurology
TABLE 3. EEG Abnormalities Described in Clinical Reports of Epileptologists
No.
Generalized Slowing
Focal Slowing
Epileptiform Discharges
EEG Seizures
1
Sev, Mod-Sev
L FT CPDA
L TP
2
Sev, Mod
R Hem CPDA
3
Sev
4
ECoG Seizures, No., Duration
Breach
Other
—
Yes
SIRPIDs
R FC
—
Yes
Bilat F delta
—
—
Yes
Sev
R FCT CPDA/IPDA
Freq R FCP
25 R FC
Yes
5
Sev, Mod
R Hem CPDA
—
—
Yes
6
Mod-Sev, Mod
R FCT CPDA
R FC
—
Yes
7
Sev, Mod, Mild
R Hem CPDA
—
—
Yes
8
Sev, Mod-Sev
L FT CPDA
Rare L C
—
Yes
9
Sev, Mod-Sev
PDA, R > L
Freq R FC, R CP, L FC
—
Yes
10
Sev, Mod, Mild
R FCT CPDA
Rare R CP
—
Yes
11
Sev, Mod
R FCT CPDA/IPDA
Freq R FC
Freq R F
12
Sev, Mod-Sev
L Hem CPDA
Rare L FC, CP
—
13
Mild-Mod
L Hem CPDA
Rare L C
Freq L F
14
Sev, Mild
Bilat F IPDA
—
—
Yes
15
Sev
R FCT CPDA
—
—
Yes
16
Sev
L FCP CPDA
Rare L CP
—
17
Sev
R Hem CPDA
R F, R P
—
Yes
PLEDs
18
Mod
L Hem IPDA
Int L CP
—
Yes
SIRPIDs
105, 221 minutes
BS
SIRPIDs
Yes Yes
161, 317 minutes
26, 103 minutes
Yes
PLEDs
Yes
Generalized slowing is described as severe, moderate, or mild based on the background frequency (delta, theta, or alpha) and the presence of reactivity. Focal slowing is described as CPDA and/or IPDA. Epileptiform discharges are described as frequent, intermittent, or rare. All EEG seizures were nonconvulsive, and all patients had breach rhythm. Other abnormalities described in reports are SIRPIDs, BS, and PLEDs. Number and duration of electrographic seizures in ECoG recordings are given for comparison. Bilat 5 bilateral; BS 5 burst suppression; C 5 central; CP 5 centroparietal; CPDA 5 continuous polymorphic delta activity; ECoG 5 electrocorticography; EEG 5 electroencephalography; F 5 frontal; FC 5 frontocentral; FCP 5 frontocentroparietal; FCT 5 frontocentrotemporal; Freq 5 frequent; FT 5 frontotemporal; Hem 5 hemispheric; Int 5 intermittent; IPDA 5 intermittent polymorphic delta activity; L 5 left; Mod 5 moderate; P 5 parietal; PDA 5 polymorphic delta activity; PLEDs 5 periodic lateralized epileptiform discharge; R 5 right; Sev 5 severe; SIRPID 5 stimulus-induced rhythmic, periodic, or ictal discharge; TP 5 temporoparietal.
patients, but was widespread through most of the injured hemisphere in others. EEG amplitude depressions were rarely observed in channels without high-amplitude delta activity, although high-amplitude delta was sometimes present in channels beyond those exhibiting depressions. The role of delta activity in observing spreading depolarizations in the EEG was confirmed in quantitative analysis. Baseline EEG activity prior to isolated-type depressions was dominated by delta (0.5–4.0Hz), which accounted for 89% of total power (Fig 5A). Thus, suppression of delta activity accounts for nearly all of the absolute magnitude of amplitude depression induced in 688
the EEG by spreading depolarizations. Furthermore, the relative suppression of delta activity was also greater than for higher frequency bands; delta was depressed to a mean 47% of baseline power, whereas higher bands had progressively less depression (see Fig 5B). Hemispheric Spreading Depression of EEG Amplitude In all 10 patients for whom EEG amplitude depressions were observed, these changes occurred in at least 1 of the bipolar EEG channels in closest proximity to the subdural ECoG strip (see Table 2). In 3 patients (Nos. 5, 11, 12), EEG Volume 76, No. 5
Hartings et al: Spreading Depression in EEG
FIGURE 3: Widespread propagation of spreading depolarization evidenced by spreading depression of continuous electroencephalography (EEG). (A) Schematic diagram and scout computed tomography (CT) show positions of cortical electrode strip and scalp electrodes in Patient 6 following partial resection of the right temporal lobe and evacuation of a right subdural hematoma. CT image shows diffuse subarachnoid hemorrhage and intraparenchymal hemorrhage in right frontal and temporal lobes. Electrocorticography (ECoG; black upper traces): In this 8.5-hour recording segment, 6 spreading depolarizations are observed as evidenced by slow potential changes (dashed circles) in the near–direct current (DC) ECoG recording that propagate from channel 2–3 to channel 4–5 and are accompanied by depression of 0.5 to 50Hz activity. EEG: Five bipolar EEG channels from the same hemisphere are shown. Colored traces correspond to the colored electrode positions in the head schematic diagram. Arrows indicate the clear amplitude depressions in channels Fp2-F4 and C4-P4 that occur in time-locked, 1to-1 correspondence with the ECoG-identified spreading depolarizations. Note that depressions occur first in C4-P4 and other posterior channels and are followed, after a substantial time lag of 30 to 40 minutes, by frontal Fp2-F4, indicating spreading depressions of EEG activity induced by spreading depolarization. The 2nd and 3rd depolarizations occurring in rapid succession are manifested as unique depression periods in Fp2-F4 but as a single continuous depression in C4-P4. Scale bars apply to all traces of the same type. (B) A 7.3-hour recording from Patient 17 is shown following the same conventions as in A. Electrode 1 (arrow) is seen in the CT scan following evacuation of a large right frontal intracerebral hematoma. Three spreading depolarizations are seen in ECoG recordings and each is accompanied by unique, isolated depressions in EEG channels throughout the hemisphere. Depressions begin in F4-C4 (blue), in near-synchrony with the electrode strip, and propagate posteriorly to P4-O2 (green) after delays of 57, 11, and 26 minutes. The changing delays suggest somewhat different origins for each depolarization wave.
amplitude depressions were restricted to a single channel. In the other 7 patients, spreading depolarizations induced EEG depressions in 2 to 6 channels, in some cases spanning broad expanses of the ipsilateral hemisphere. In 5 of these 7 patients, 62 of 179 isolated-type events (35%) showed a clear spread of the amplitude depression with a distinct time delay between depression onset in different channels. Examples from 2 patients are shown in Figure 3 (see also Fig 1, Fig 4A). For each depolarization, time intervals between nadirs of the amplitude depressions of the 2 channels with the clearest signals, and not sharing a common electrode, were measured. These intervals were 17 minutes (median; IQR 5 11–34) and did not significantly differ by the distance separating the electrode pairs. In 3 patients (Nos. 6, 16, 17), spreading depolarizations induced spreading EEG depressions in the most anterior and posterior channels (eg, Fp2-F4 and P4-O2), November 2014
indicating widespread propagation throughout the ipsilateral cerebral cortex. DC Potentials in EEG Shifts of the extracellular DC potential are a signature of spreading depolarization, reflecting the mirror image of sustained intracellular depolarization. In 4 patients with both spreading depolarizations and DC-coupled EEG recordings, monopolar EEG channels were examined for the presence of DC shifts accompanying depressions of the conventional frequency band (0.5–50Hz). In the majority of cases, DC shifts could not be identified. However, clear instances were found in 3 of 4 patients (Nos. 15, 17, 18) with peak-to-peak amplitudes of 708 6 242lV (n 5 37), 333 6 60lV (n 5 6), and 432lV (n 5 2), respectively. In Patient 15, DC shifts were 689
ANNALS
of Neurology
FIGURE 4: Further examples of EEG depression patterns. A and B show 5.25- and 3.5-hour epochs from Patients 17 and 16, respectively, during which no electrocorticographic recordings were obtained. In A, 5 recurring amplitude depressions are observed in channels Fp2-F4, F4-C4, C4-P4, and P4-O2 in a fashion similar to the example of Figure 3B from the same patient. In contrast to Figure 3B, depressions recur more frequently at 53- to 75-minute intervals and do not spread as far posteriorly. Spread is clearly observed as a march in depression from frontal to central to parietal locations (arrows). In B, 8 cycling amplitude depressions occur at 25- to 30-minute intervals in C3-P3 and, to a lesser extent, P3-O1. This represents a rare case in which depolarizations recurring at short intervals each produced an isolated-type depression instead of a single continuous depression. Unlike in A, no spread is evident in this case. Integrals are shown below.
observed during a cluster of 42 repetitive spreading depolarizations that spanned a total of 27 hours. During this time, high-frequency EEG activity was continuously depressed and the DC shifts appeared as a highly unstable, continuously fluctuating baseline potential (see Fig 2B). The DC shifts propagated between electrodes Fp2 and F4 with a delay of 9.3 minutes (mean; standard deviation 5 2.3, n 5 25). The cessation of depolarizations was then signaled by a marked EEG state change; the DC potential became steady and the amplitude of 0.5 to 50Hz activity increased with the onset of polymorphic delta activity. A similar fluctuating DC potential was observed in Patient 17 during a cluster of 11 depolarizations spanning 7 hours of continuous depression, although the DC shifts corresponding to individual depolarizations could not be deciphered.
Discussion Here we have shown that the vast majority of spreading depolarizations (>80%) recorded from the surface of cerebral cortex by ECoG are manifested as depressions of high-amplitude delta activity in clinically standard, noninvasive EEG recordings. When intervals between depolarizations are long enough, they induce unique 690
suppressions of this high-amplitude baseline such that amplitudes are normalized to levels in other head regions. These depressions, attributable to individual depolarization waves, develop slowly and may persist up to 1 hour before the baseline EEG amplitude is fully restored. When depolarizations occur repetitively at short intervals, the suppression of high-amplitude delta activity is maintained for longer periods, from hours to >1 day. We have further shown that Le~ao’s “spreading depression of activity”5 can be observed in scalp EEG, because many spreading depolarizations produced depressions that spread with a temporal delay between EEG channels. In some cases, affected channels were located far from the ECoG recording strip, thus providing the first evidence for widespread propagation of spreading depolarizations across the injured hemisphere. These results suggest a new application for continuous EEG by which a neuronal pathophysiological mechanism can be monitored as a possible etiology of neurologic condition and marker of ongoing secondary injury.
Spreading Depression in Scalp EEG From the time of Le~ao’s discovery in 1944, there was no evidence that spreading depolarizations could be observed Volume 76, No. 5
Hartings et al: Spreading Depression in EEG
area. Similarly, continuous depression periods are explained by the finding that multiple depolarizations, when recurring at short intervals, may occur simultaneously in the same recording area and thus maintain a continuous cycle of partial depression. The failure to detect Le~ao’s spreading depression through decades of EEG research is thus likely attributable to the long durations of development, persistence, and spread; recognition requires viewing of long recording epochs on highly compressed time scales (4 to 9cm/h) that historically have not been used in clinical neurophysiology. Routine 30-minute EEG examinations reviewed at 3cm/s preclude observation of events on this temporal scale.
FIGURE 5: Electroencephalographic (EEG) depressions are suppressions of delta activity. (A) Proportions of total power of baseline activity contributed by activity in the 4 frequency bands. Assessments were made prior to depolarization-induced EEG depression for 3 events in each of 7 patients. Error bars show standard deviations. (B) Degrees of depression as percentages of baseline power for each frequency band.
noninvasively from scalp EEG until the recent report of Drenckhahn et al.20 In part, this was due to the belief that the spatial wavelength of spreading depression across the cortex was too short to influence EEG signals recorded from a much larger volume of tissue. However, it is estimated that scalp potentials reflect the synchronous activity of at least 6 to 10cm2 across the brain’s surface, which may include twice that area of cerebral cortex.21,22 Because spreading depolarization travels at 1 to 8mm/min and depresses cortical activity for an average duration of 8 minutes,11,13 the typical spatial wavelength of depression is estimated at 0.8 to 6.4cm. Assuming a wavefront width of at least 1cm, and given that depressions can be arbitrarily long (hours), it is clear that depolarizations can simultaneously depress the requisite volume of cortex to be observed in EEG. Compared to intracranial electrodes, the larger recording area of scalp electrodes further explains the lesser degree of maximal EEG depression, the slower onset and recovery of depressions, and the longer total duration; only a subset of the total recording area is depressed at a given instance and more time is required for the wave to invade maximally and traverse the larger November 2014
Clinical Use of Continuous EEG As a measure of the brain’s essential electrophysiologic function, EEG has long offered unrealized promise as a tool for monitoring the progression of intracranial pathology.23,24 The advantages of EEG over other neuromonitoring techniques are that it is noninvasive and continuous and assesses both regional and global cerebral status. Furthermore, quantitative analysis can facilitate interpretation of the large volumes of data generated.25 Long-term trending of alpha variability and other metrics, for instance, have been investigated as predictors of delayed cerebral ischemia in SAH patients.26,27 Spreading depolarizations have now emerged as an important mechanism of lesion development and a target for EEG monitoring. In focal ischemia in the rat, a delayed phase of repetitive depolarization activity begins 6 to 10 hours after stroke onset and coincides with the period of delayed infarct maturation.28 Importantly, these depolarization clusters did not produce unique depressions of spontaneous EEG activity, but rather were manifested as a continuous suppression of high-amplitude polymorphic delta activity.29 The occurrence of continuous depression periods in EEG recordings during temporal clusters of spreading depolarizations was since confirmed in SAH patients.20 As in animals, such clusters have a delayed onset and high positive and negative predictive value for delayed cerebral ischemia.13,14 Here we have further shown in patients that both isolated and continuous depressions of EEG amplitude induced by spreading depolarizations represent suppressions of pathologic high-amplitude delta activity, a common feature of all types of acute brain injury. Increases in delta power result from preserved but abnormal synaptic activity due to structural damage or reduction of cerebral blood flow to a range of 15 to 22ml/100g/min. This increase is lost with suppression of synaptic activity during mass depolarization of cortex, which may indicate worsening ischemia because anoxic terminal spreading 691
ANNALS
of Neurology
depolarization develops at a threshold of 5 to 10ml/ 100g/min. However, spreading depolarization can invade cortex with normal or oligemic blood flow. In either case, the suppression of delta activity caused by depolarization paradoxically does not indicate resolution of abnormalities, but rather the occurrence of more severe electrophysiologic dysfunction. These translational studies thus raise the strong possibility that EEG monitoring of depolarizations could be used as a surveillance tool for progression of intracranial pathologies. In recent years, continuous EEG has gained traction and is performed as standard of care at an increasing number of hospitals. The rationale for continuous, multiday monitoring is primarily the evaluation of subclinical seizures as a possible etiology for altered mental status or coma, as well as belief that suppression of all early seizures, whether convulsive or nonconvulsive, may improve outcomes.30,31 Seizures and spreading depolarizations often occur with interacting patterns in the same tissue, and depolarizations may enhance epileptic activity by suppression of inhibitory neurotransmission.32–34 However, compared to seizures, spreading depolarizations have a 3- to 7-fold higher incidence in patients.19,35 The present findings thus suggest a new and complementary application for continuous EEG in acute brain injury. Implications for EEG Methodology In several respects, our results are similar to those reported by Drenckhahn et al in SAH patients.20 In that study, 47% of ECoG-identified spreading depolarizations induced isolated depressions in the scalp EEG, compared to 39% found here, and depressions reached 53% of baseline amplitude compared to 57% in the present study. In both studies, the remainder of depolarizations occurred at short intervals and did not induce unique depressions but rather maintained continuous depressions throughout temporal clusters of events, as noted above. Thus, these results are reproducible across different centers and diseases and represent general features of acute brain injury. Results differed somewhat in other respects. First, DC shifts were observed in scalp EEG for a majority (70%) of spreading depolarizations recorded by Drenckhahn et al, but were rarely observed in the 4 patients with DCcoupled EEG in the present study. Second, we found that both EEG depressions and DC shifts spread between scalp electrodes with a temporal delay, whereas all changes occurred synchronously in SAH patients. The wide spatial extent of spreading depression observed here was surprising and contrasts with the limited extent (
