YEBEH-05957; No of Pages 7

Epilepsy & Behavior xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Epilepsy & Behavior

journal homepage: www.elsevier.com/locate/yebeh

Afterdischarges elicited by cortical electric stimulation in humans: When

do they occur and what do they mean?☆
Stephanie Gollwitzer a,⁎, Rüdiger Hopfengärtner a, Karl Rössler b, Tamara Müller a, David Gerhard Olmes a,
Johannes Lang a, Julia Köhn a, Müjgan Dogan Onugoren a, Jana Heyne a, Stefan Schwab a, Hajo Martinus Hamer a
a
Epilepsy Center, Department of Neurology, University Hospital Erlangen, Erlangen, Germany
b
Department of Neurosurgery, University Hospital Erlangen, Erlangen, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Introduction: Afterdischarges (ADs) are a common and unwanted byproduct of direct cortical stimulation during
Received 1 August 2018 invasive electroencephalography (EEG) recordings. Brief pulse stimulation (BPS) can sometimes terminate ADs.
Accepted 9 September 2018 This study investigated AD characteristics and their relevance for emergence of stimulation seizures. In addition,
Available online xxxx AD response to BPS was analyzed.
Material and methods: Invasive EEG recordings including mapping with direct cortical stimulation in patients
Keywords:
with refractory epilepsy at the Erlangen Epilepsy Center were retrospectively reviewed. Afterdischarge defined
Cortical stimulation
Invasive EEG
as stimulation-induced rhythmic epileptiform discharges of more than a two-second duration were analyzed re-
Afterdischarges garding incidence, localization, duration, propagation pattern, morphology, and seizure emergence. In addition,
Focal epilepsy the influence of AD characteristics and stimulation settings on BPS success rate was studied.
Brief pulse stimulation Results: A number of 4261 stimulation trials in 20 patients were investigated. Afterdischarge occurred in 518 tri-
als (14.2%) and lasted 12.4 s (standard deviation [SD]: 8.6 s) on average. We elicited ADs in the seizure onset zone
(SOZ) (n = 64; 19.4%), the irritative zone (n = 105, 20.0%), and outside the irritative area (n = 222, 12.5%).
Rhythmic spikes (30.5%) and spike–wave complexes (30.3%) represented predominant morphologies.
Afterdischarge morphology in the SOZ and hippocampus differed from other areas with polyspikes and sequen-
tial spikes being the most common types there (p = 0.0005; p b 0.0001 respectively). Hippocampal ADs were
particularly frequent (n = 50, 38.2%) and long-lasting (mean: 16.6, SD: 8.3 s). Brief pulse stimulation was applied
in 18.1% of the AD trials (n = 94) and was successful in 37.4% (n = 40). Success rates were highest when BPS was
delivered within 9.5 s (p = 0.0048) and in ADs of spike–wave morphology (p = 0.0004). Fifteen clinical seizures
emerged from ADs (3.55%), mostly evolving from sequential spikes. Afterdischarges in patients with stimulation
seizures appeared more widespread (p b 0.0001) and lasted longer (mean duration 7.0 s) than in those without
(mean duration 21.0 s, p = 0.0054).
Conclusion: Afterdischarges appear in the epileptogenic and nonepileptogenic cortex. Duration and propagation
patterns can help to quantify the risk of stimulation seizures, with sequential spikes being most susceptible to
seizure elucidation. The hippocampus is highly sensitive to AD release. Brief pulse stimulation is a safe and effi-
cacious way to terminate ADs, especially when delivered quickly after AD onset.
© 2018 Elsevier Inc. All rights reserved.

1. Introduction Underlying mechanisms of ADs are still poorly understood as is the

relevance of ADs for delineation of the epileptic network. While ADs
Since the early 20th century cortical electrical stimulation (ES) has can be elicited in all brain regions, the propensity of different cortical
been used for the mapping of eloquent cortex in patients with intracta- areas to generate ADs is highly variable [5]. Single neuron studies hy-
ble focal epilepsy undergoing intracranial electroencephalography pothesized that the epileptic cortex is more easily driven to generate
(EEG) recordings for presurgical assessment [1,2]. Stimulation- ADs than cortical areas outside the epileptic zone [6]. However, the like-
induced rhythmic epileptiform discharges referred to as afterdischarges lihood of AD occurrence can fluctuate from one stimulation trial to
(ADs) are frequent unwanted byproducts of ES [3,4]. another depending on the functional state of the stimulated network
at the time of stimulation [5]. That could explain why even stimulating
☆ Funding: This research did not receive any specific grant from funding agencies in the
the same pair of electrodes repeatedly may elicit ADs in some but
public, commercial, or not-for-profit sectors.
⁎ Corresponding author at: Epilepsy Center, Department of Neurology, University
not all trains of stimulation [5]. Also, stimulation thresholds exhibit con-
Hospital Erlangen, Schwabachanlage 6, 91054 Erlangen, Germany. siderable inter- and intraindividual variability [7]. It is furthermore un-
E-mail address: [email protected] (S. Gollwitzer). clear why certain ADs propagate and evolve to clinical seizures, while

https://doi.org/10.1016/j.yebeh.2018.09.007
1525-5050/© 2018 Elsevier Inc. All rights reserved.

Please cite this article as: Gollwitzer S, et al, Afterdischarges elicited by cortical electric stimulation in humans: When do they occur and what do
they mean?, Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.09.007
2 S. Gollwitzer et al. / Epilepsy & Behavior xxx (xxxx) xxx–xxx

others remain limited to stimulated electrodes and terminate spontane- emission tomography [FDG-PET]) and semiology. The implantation de-
ously [8]. sign was limited to the left hemisphere in 9, and to the right in the ad-
Afterdischarges tend to adopt a set of stereotypic morphologies that ditional 9 patients; 2 patients had bilateral coverage. The frontal lobe
were reproducible in different patient cohorts [4,8]. The underlying was the main target in 10, the parietal lobe in 3, the temporal lobe in
neuronal process and pathophysiological significance of these distinct 7 patients. All temporal implantations comprised hippocampal elec-
waveform patterns, that can appear in the same individual and even trode contacts. On MRI, no clearcut lesion was seen in 9 patients;
in the same electrode-pair in consecutive trains of stimulation, are malformations of cortical development were assumed in 9 patients;
unknown [9]. postoperative/posttraumatic lesions were found in 2 patients (Table 1).
Brief pulse stimulation (BPS) can terminate ADs in some cases while All patients provided written informed consent and approved the
no effect is seen in others [10,11]. Various characteristics have been pro- usage of EEG data for scientific research.
posed to differentiate refractory ADs from those responding to BPS, but
safe prediction of BPS-success based on AD properties is still impossible 2.2. Intracranial EEG and cortical stimulation
[10,12].
Therefore, we aimed to shed more light on incidence, localization, Prior to implantation, computed tomography (CT) and MRI scans in-
duration, propagation patterns and morphology of ADs. Furthermore, cluding magnetic resonance angiography were performed, and imaging
we addressed the question whether AD characteristics can help to iden- data were transferred to neuronavigation [14]. Trajectories were then
tify patients at risk of stimulation-induced seizures, and we examined in planned considering target structures and vessels. After implantation,
which scenarios BPS is most likely to abort ADs. CT and MRI scans were repeated, and electrode maps were generated
on 3D brain surface rendering allowing for localization of electrodes
2. Material and methods with respect to anatomical structures.
Electroencephalographic data were recorded with a standard clinical
2.1. Patients video-EEG system (Micromed) sampled at 2048 Hz per channel, with
bandpass filtering between 0.15 Hz and 500 Hz.
We retrospectively analyzed intracranial EEG recordings of 20 pa- Cortical stimulation for functional mapping was performed at the
tients with drug-refractory focal epilepsy who underwent invasive end of the recording. Antiepileptic drugs (AEDs) were stopped tran-
EEG (iEEG) monitoring with either stereotactically implanted depth siently during the monitoring in order to elicit seizures, but reinstituted
electrodes or subdural grid electrodes at the Erlangen Epilepsy Center more than 24 h before cortical stimulation in 18 patients (90%). Eleven
between 2016 and 2018 (Table 1). In addition, cortical stimulation map- patients (55%) received benzodiazepines prior to stimulation in order to
ping as part of their presurgical assessment had to be performed. prevent stimulation seizures.
The implantation scheme was based solely on the hypothesis of Cortex was mapped in two to three stimulation sessions lasting be-
the epileptic zone derived from noninvasive presurgical diagnostics tween 1 h and 3 h. Adjacent electrodes were stimulated in a bipolar
(noninvasive EEG monitoring, 3-T magnetic resonance imaging (MRI), fashion using biphasic pulses of 0.3 ms at a frequency of 50 Hz, each
magnetoencephalography [MEG], ictal [single-photon emission com- train lasting for up to 5 s. Current was gradually increased in 1- to
puted tomography] SPECT, and/or fluorodeoxyglucose-positron 2-mA steps from 1 mA to a maximum of 15 mA or until functional re-
sponses occurred.
Table 1
If ADs appeared, stimulation was only resumed after their termina-
Patients' characteristics. tion. Brief pulse stimulation was applied when ADs were regarded
as persistent based on clinical judgment of the stimulating physician.
Number of patients 20
Age, years, mean ± SD 32.5 ± 10.9
Current strength of BPS usually resembled the initial stimulus causing
Female, N (%) 11 (55%) ADs. Brief pulse stimulation trains lasted between 500 ms and 5 s.
Duration of epilepsy, years, mean ± SD 18.4 ± 9.3
Seizure frequency per month, median (range) 2.3. EEG analysis
Focal 30 (2–100)
Generalized 0 (0–20)
Localization of epileptogenic zone Electroencephalography during stimulation was analyzed offline by
Left/right/bilateral, N 9/9/2 a board-certified electroencephalographer (SG). In addition, a second
Frontal/temporal/parietal, N 10/7/3 board-certified electroencephalographer (HMH) additionally reviewed
MRI pathology, N (%) unclear cases, which were resolved in discussion between electroen-
Nonlesional 9 (45%)
Lesional 11 (55%)
cephalographers. All ADs of ≥2 s of duration were marked.
Number of electrodes, median (range) 94.5 (60–131) In terms of localization, AD electrodes were assigned to the follow-
Localization of electrodes, N (%) ing regions of interest (ROI): SOZ, defined as the zone displaying first
Frontal 978 (51%) ictal changes in intracranial EEG signal within the context of patients'
Temporal 473 (25%)
spontaneous seizures [15]; propagation zone (PZ), including cortical
Hippocampal 89 (5%)
Cingulum 59 (3%) areas involved in early seizure spread [16]; irritative zone (IZ), referred
Insula 218 (11%) to as the area generating interictal epileptic discharges [17]; neocortex;
Parietal 79 (4%) hippocampus.
Occipital 11 (1%) Afterdischarge evolution was divided into three groups: 1. ADs
Resection/no resection, N (%) 12 (60%)/8 (40%)
limited to stimulated electrodes; 2. ADs evolving to adjacent electrodes;
Outcome class 1 7 (58%)
Outcome class 2 0 3. ADs involving in electrodes remote from the stimulated site.
Outcome class 3 1 (8%) Afterdischarges evolving to clinical seizures were specifically marked.
Outcome class 4 3 (25%) In addition, AD duration was noted as well as the minimal current
Outcome class 5 1 (8%)
eliciting ADs in a given electrode pair. Afterdischarge morphology was
Pathology, N (%)
Focal cortical dysplasia 9 (75%) classified according to established criteria [4]: 1. Rhythmic waves, i.e.,
Polymicrogyria 1 (8%) regular 5- to 9-Hz rhythmic activity (RW) [18]; 2. Rhythmic waves
Hippocampal sclerosis 1 (8%) evolving into rhythmic spike waves of medium frequency (4–9 Hz;
Nonspecific 1 (8%) RS); 3. Bursts of polyspikes at a frequency of 1–4 Hz (PS); 4. Repetitive
⁎Surgical outcome according to ILAE criteria [13]. spike waves at 1–4 Hz (SW); 5. Sequential spikes, i.e., runs of rapid

Please cite this article as: Gollwitzer S, et al, Afterdischarges elicited by cortical electric stimulation in humans: When do they occur and what do
they mean?, Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.09.007
 S. Gollwitzer et al. / Epilepsy & Behavior xxx (xxxx) xxx–xxx 3

spikes at N 9 Hz (SS) (Fig. 1). Whenever ADs at onset were masked by ar- Because of the Bonferoni correction, the significance level was set to
tifacts, for example, because of capacitor saturation, the earliest analyz- 0.01. To evaluate the association between time delay and BPS success,
able pattern was used for categorization. When one waveform evolved we used Mann–Whitney-U tests and calculated the optimal cutoff
into another over time, decision was based on the earliest pattern time applying receiver operator characteristic. Afterdischarge morphol-
persisting stably over ≥2 s. ogies in trials successfully terminated by BPS compared with that in re-
In addition, we reviewed patient scalp-EEG monitoring data col- fractory ones were also compared using Chi-square testing. Properties
lected during a previous telemetry in order to assess baseline frequency of hippocampal ADs as compared with that of ADs in other regions
of interictal epileptiform discharges (IEDs). A one-hour sample of sleep were analyzed by Fisher's exact tests.
EEG assigned to sleep stage I or II, sampled during the first 24 h of re-
cording, prior to any seizure provoking measures, e.g., drug reduction
or sleep deprivation, was analyzed. Interictal epileptiform discharges 3. Results
(spikes, sharp waves, and polyspikes) were identified based on gener-
ally accepted criteria [19]. Quantification of IEDs in intracranial EEG In 20 patients, a total of 1905 electrodes were implanted. Of those,
was not performed, as continuous spiking in some electrodes was the 1366 electrodes were electrically stimulated in 4261 stimulation trials.
predominant pattern in the majority of patients. Fifteen patients generated ADs during stimulation. Patients who devel-
oped ADs did not differ from those remaining free of ADs regarding age
2.4. Statistical analysis (Mann–Whitney test: p = 0.39), duration of epilepsy (p = 0.83), and
median seizure frequency (p = 0.38). However, we found a higher
We applied Mann–Whitney tests to test for differences between pa- IED frequency in baseline surface EEG in patients with AD (median
tients with and without ADs with reference to IED frequency on surface 40 per hour, range 1–300) as compared with that in patients free from
EEG. Percentage of AD trials and IED frequency were correlated using AD (median 10 per hour, range 0–26; Mann–Whitney test p = 0.016;
nonparametric correlation test (Spearman). Chi-square tests were cal- receiver operator characteristics (ROCs) p = 0.015, area under the
culated to compare AD-propagation patterns in patients who developed ROC curve (AUC): 0.87). The total amount of IED per hour did not corre-
at least one clinical seizure as a consequence of cortical stimulation and late with percentage of AD trials referred to all trials per patient
patients who remained seizure-free throughout the testing sessions. (Spearman's rho 0.26, p = 0.13).
We also applied Chi-square tests to assess AD morphologies in different Overall, ADs appeared in 14.2% (n = 518) of stimulation trials and in
ROIs as well as in patients with and without stimulation-seizures. 32.2% (n = 419) of stimulated electrodes. Afterdischarges lasted on

Fig. 1. Examples of AD morphologies. A. Rhythmic waves (RW), frequency range 5–9 Hz; B. Rhythmic spikes (RS), frequency range 4–9 Hz; C. Spike waves (SW), frequency range 1–4 Hz; D.
Polyspikes (PS), frequency range 1–4 Hz; E. Sequential spikes (SS), frequency range N 9 Hz.

Please cite this article as: Gollwitzer S, et al, Afterdischarges elicited by cortical electric stimulation in humans: When do they occur and what do
they mean?, Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.09.007
4 S. Gollwitzer et al. / Epilepsy & Behavior xxx (xxxx) xxx–xxx

Fig. 2. AD characteristics in patients with and without stimulation seizures A. Patients suffering stimulation seizures generated less circumscribed AD, evolving to adjacent electrodes (AE);
in seizure-free patients, AD were restricted to stimulated electrodes (SE). Remote electrodes (RE) were involved in both groups likewise. B. ADs lasted significantly longer (p = 0.0054) in
patients with than in patients without stimulation seizures.

average 12.4 s (SD: 8.3 s); 81.9% (n = 422) of AD stopped spontane- waves being far more common outside epileptogenic areas, while
ously; 3.55% of AD (n = 15) evolved to clinical seizures. polyspikes appeared more frequently in the SOZ (Table 2).
Of the stimulations, 19.4% in the SOZ and 20.0% in the IZ led to AD, Brief pulse stimulation was used in 18.1% of AD trials. A mean time of
while stimulation in nonepileptogenic cortical regions caused ADs in 13.80 s (SD: 4.70 s, range: 3–50 s) elapsed from the beginning of AD
12.5% of the cases. The difference was not significant (Kruskal–Wallis until start of BPS. The BPS lasted 1.89 s (SD: 1.61 s) on average. In
test, p = 0.97). 37.4% of the patients, BPS terminated ADs (Fig. 3). Brief pulse stimula-
In consecutive trains of stimulation with increasing current, ADs tion was significantly more likely to terminate AD when applied earlier
first appeared at a median current of 7.5 mA (range 3–15 mA) aver- (mean time delay in successful trials 10.1 s (SD: 4.8 s); mean time delay
aged across all electrode pairs. The cutoff was slightly lower in the in failed trials 19.0 s (SD: 4.7 s); Mann–Whitney test p = 0.0048;
SOZ (median 6 mA, range 3–13 mA) and IZ (median 5 mA, range Table 2, Fig. 3 A,B). According to the ROCs (area under the curve 0.75,
3–13 mA) when compared to nonepileptogenic regions (median 8 mA, p b 0.001), BPS was most promising when started within a time window
range 3–15 mA); the difference, however, was not significant (Kruskal– of 9.5 s (sensitivity 71.1%, specificity 72.5%) (Fig. 4). The BPS duration
Wallis test, p = 0.2). had no impact on success rate (Mann–Whitney test p = 0.62).
Afterdischarges remained restricted to stimulated electrodes in The morphology of AD sensitive to BPS differed from AD refractory
24.8% (mean; SD: 18.7%; n = 128) of stimulation trials, involved adja- to BPS, spike waves being predominant in the first, polyspikes occurring
cent electrodes in 21.2% (mean; SD: 23.7%; n = 110) and propagated more commonly in the second group (Chi-square test, p = 0.004;
to remote nonadjacent electrode contacts in 53.0% (mean; SD: 27.1%, Table 2, Fig. 3 C,D).
n = 280). Circumscribed ADs appeared more frequently in patients In five patients, hippocampal stimulation elicited AD. In two of
without stimulation seizures (ADs in stimulated electrodes only: these, the AD generating hippocampus represented the assumed epi-
34.1%; AD involving adjacent electrodes: 10.5%; AD propagating to re- leptogenic zone. In three patients, however, ADs were elucidated in a
mote contacts: 55.4%) as compared with that in those who developed supposedly healthy hippocampus. Overall, the hippocampus generated
one or more stimulation-induced seizures (AD in stimulated electrodes AD more frequently than other brain regions (38.2% vs. 14.0%; Chi-
only: 14.1%%; AD involving adjacent electrodes: 35.6%%; AD propagating square test p = 0.0001) and at lower current applied (median 5 mA,
to remote contacts: 50.3%; Chi-square test: p b 0.001). range 3–11 mA vs. median 7.5 mA, range 3–15 mA). Moreover, AD in
Patients with stimulation seizures generated longer-lasting AD the hippocampus lasted longer (mean: 16.6 s, SD: 16.7 s) than elsewhere
(mean: 7.0 s, SD: 2.12 s) than patients without stimulation seizures (mean: 12.4 s, SD: 8.3 s). Therefore, hippocampal AD were more likely to
(mean: 21.0 s, SD: 9.21 s; Mann–Whitney test p = 0.0054; ROC be judged as persistent, and BPS was applied more regularly (32.0%)
p = 0.0045, AUC = 0.94; cutoff 9.8 s) (Fig. 2). as compared to other areas (18.1%). However, BPS success rate was
Most ADs consisted of rhythmic spikes (RS) and spike waves (SW) low in the hippocampus (6.25% vs. 37.4% elsewhere; Chi-square test
(30.5% and 30.3%, respectively). Less frequently, AD appeared as p b 0.0001). Regarding AD morphology, polyspikes outweighed other
sequential spikes (SS; 21.8%), polyspikes (PS; 10.2%), and rhythmic waveforms while spike waves were rarely seen (Chi-square test for
waves (RW; 7.3%). Afterdischarge morphologies differed between trials comparison to other regions: p b 0.0001; Table 2).
within electrodes. All patients exhibited different morphologies. Seven patients (35%) suffered from a total of 15 clinical seizures as
Patients with and without stimulation-induced seizures did not vary a consequence of cortical stimulation. In nine cases (60%), seizures
in terms of distribution of AD morphologies (Chi-square test, p = 0.07). emerged from AD consisting of sequential spikes. Five ictal patterns
However, AD morphologies elicited in the SOZ differed significantly (33%) evolved from spike–wave ADs, one (7%) from rhythmic spikes.
from those evoked in the nonepileptic cortex (p = 0.0005), spike Seizures could be induced by stimulation in patients' SOZ (n = 7) or

Table 2
Morphological patterns of afterdischarges.

Rhythmic waves Rhythmic spikes Polyspikes Spike waves Sequential spikes Chi-square test, p

SOZ 4.3% 32.9% 24.3% 17.1% 21.4%

Nonepileptogenic cortex 6.3% 29.3% 5.3% 36.0% 23.0% 0.0005
Total AD 7.3% 30.5% 10.2% 30.3% 21.8%
Pt. with stimulation seizures 7.1% 26.2% 6.0% 42.9% 17.9%
Pt. without stimulation-induced seizures 7.3% 32.6% 12.3% 24.0% 23.8% 0.066
All areas 7.3% 30.5% 10.2% 30.3% 21.8%
Hippocampus 0% 34.0% 34.0% 4.0% 28.0% b0.0001
AD terminated by BPS 0% 17.5% 5.0% 50.0% 27.5%
AD resistant to BPS 5.2% 20.8% 12.5% 37.5% 24.0% 0.0027

SOZ: seizure onset zone; BPS: brief pulse stimulation; Pt: patient; Level of significance after Bonferoni correction: p = 0.01. Bold values indicate significant results.

Please cite this article as: Gollwitzer S, et al, Afterdischarges elicited by cortical electric stimulation in humans: When do they occur and what do
they mean?, Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.09.007
 S. Gollwitzer et al. / Epilepsy & Behavior xxx (xxxx) xxx–xxx 5

Fig. 3. Impact of brief pulse stimulation (BPS) on ADs. A. Rhythmic spike ADs terminated by BPS applied after 8 s; B. sequential spike ADs refractory to BPS applied after 14 s; C. polyspike
ADs refractory to BPS; spike–wave ADs sensitive to BPS.

IZ (n = 1), and in nonepileptogenic tissue (n = 7) alike. A median cur- reliable functional responses and to avoid ADs at the same time can
rent of 9 mA (range 3–15 mA) set off seizures. In 13 cases, stimulation be challenging. In our study, 14.2% of stimulation trials elicited ADs.
seizures resembled patients' habitual seizures in terms of semiology; This is in line with previous research [4]. Afterdischarges appeared in
only two seizures were rated atypical. The ictal pattern was localized 75% of the patients, the remaining 25% did not generate ADs although
within the SOZ in 10 cases, although three of those resulted from identical stimulation parameters were applied. As we used bipolar
stimulation of electrodes remote from the SOZ. Five ictal patterns that stimulation in most cases, the results do not contribute to the discus-
were induced distant from the SOZ remained localized in the stimulated sion about potential differences between bipolar and referential stim-
electrodes. Seizures patterns did not propagate to the cortex that was ulation [22].
neither stimulated nor intrinsically epileptogenic. Brief pulse stimula- We could not confirm the findings of Lesser et al. indicating that
tion was applied without any success in seven evolving seizures after stimulation becomes more likely to induce ADs with increasing patient
a mean time delay of 13.1 s. age and duration of epilepsy [5]. There was, however, an association
between the appearance of AD and the prevalence of interictal epileptic
4. Discussion discharges on noninvasive surface EEG. Thus, an active IZ indicates an
increased risk to induce ADs with cortical stimulation. We evaluated
Direct cortical stimulation is still regarded as gold standard for lo- the baseline spike frequency in surface rather than intracranial EEG as
calization of eloquent cortical areas in neurosurgical resection plan- the generally high amount and variability of IEDs in invasive EEG did
ning [20,21]. Performing stimulation with sufficient current to obtain not allow for a differentiated analysis in this sample size. Intracranial

Fig. 4. Association between time delay and success rate of BPS. The likelihood to abolish ADs by BPS was significantly higher, when BPS was applied early after AD onset. A. mean time delay
between AD onset and BPS initiation in successful and nonsuccessful BPS trials (Mann–Whitney test, p = 0.0048). B. In receiver operator characteristics, chances to terminate ADs were
highest when BPS was delivered within 9.5 s (*) (AUC 0.75, p b 0.001).

Please cite this article as: Gollwitzer S, et al, Afterdischarges elicited by cortical electric stimulation in humans: When do they occur and what do
they mean?, Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.09.007
6 S. Gollwitzer et al. / Epilepsy & Behavior xxx (xxxx) xxx–xxx

spikes are often numerous and widespread. Therefore, their contri- Brief pulse stimulation is used in most centers performing cortical
bution to the exact delineation of the epileptogenic zone remains stimulation mapping to abort ADs or stimulation-induced seizures.
controversial. Spikes, which recruit sufficient surrounding cortex to The decision when and how to apply BPS is made instantaneously
be detectable on the scalp, are, therefore, considered to be of higher based on the investigators impression that the ADs might be self-
significance [23]. sustaining or progressing into seizures. In consequence, BPS trials are
The informative value of AD for localization of the SOZ has been con- subject to a certain variability in terms of latency and pulse duration.
troversially discussed. In our cohort, ADs appeared in epileptogenic Taking this into consideration, we could still identify favorable circum-
and nonepileptogenic areas alike. While a high concordance between stances for AD termination by BPS. First and foremost, BPS should
spontaneous and induced clinical seizures and auras has been observed, be applied quickly. The longer the latency between AD initiation and
several studies failed to prove the congruence between the epilepto- BPS deliverance is, the smaller the impact. In our cohort, BPS was
genic zone and areas generating ADs [4,8,24]. This appears plausible most effective when started within a time window of around 10 s. In
from a pathophysiological point of view as ADs result from increased contrast to other reports, the duration of the BPS train did not influ-
synchronization of local field potentials in adjacent cortical columns ence success rate [11]. As instructions of the investigator concerning
[9,25]. Thus, by causing an excitatory–inhibitory imbalance, electrical start and stop of BPS were given spontaneously and thus arbitrarily,
stimulation can obviously drive both the epileptic and healthy cortex most BPS trains had similar lengths, and the majority lasted between
to generate ADs [26,27]. This conclusion is substantiated by our find- 1 and 2 s. A structured investigation with BPS trains of predefined du-
ing that AD current threshold did not differ in epileptogenic and rations would be helpful to study how long BPS should last to be most
nonepileptogenic areas. effective.
The hippocampus, however, represented a particularly sensitive en- Regarding morphology, spike–wave ADs were by far most sensi-
vironment in terms of AD elucidation. Here, ADs were more frequent tive to BPS. Again, membrane repolarization and hyperpolarization,
and longer-lasting and were elicited at lower stimulation currents. expressed by the slow wave, represent a target to intrinsic or iatrogenic
This finding is backed up by several studies [8,28]. In the diseased hip- AD termination. In any case, BPS did not worsen ADs or promote seizure
pocampus, increased excitability is expectable [29]. Interestingly, in emergence, thus the application appeared to be safe.
three patients of our cohort, ADs resulted from stimulation of assumedly To conclude, ADs are a frequent byproduct of cortical stimulation
healthy mesial temporal structures. If bitemporal pathology as a conse- that bears a substantial risk of stimulation seizures. We identified spe-
quence of secondary epileptogenesis in temporal lobe epilepsy may cific features of individual ADs, such as their duration and propagation
account for this finding or if irritability is generally increased in the hip- pattern, as warning signals of an increased seizure risk. AD patterns of
pocampus remains unclear [30,31]. lower frequency containing slow wave components can be considered
In keeping with previously applied classifications, we assigned ADs as more benign than fast spike rhythms. Brief pulse stimulation is a
to five distinct morphological categories [4,8]. Overall, ADs consisting safe and effective way to terminate AD, especially when delivered
of rhythmic spikes and spike–wave patterns were most common in within 10 s after AD onset.
our cohort. In two specific areas, however, namely, the SOZ and the hip-
pocampus, spike waves were rarely induced. Sequential spikes and Conflicts of interest
polyspikes were the predominant pattern in these structures.
Also, most clinical seizures developed from sequential spike pat- Hajo M. Hamer has served on the scientific advisory board of
terns. Other patterns, especially when comprising slow wave compo- Cerbomed, Desitin, Eisai, GlaxoSmithKline, Pfizer, and UCB Pharma.
nents, evolved to seizures less frequently. Slow waves following He served on the speakers' bureau of or received unrestricted grants
spikes are assumed to counteract evolution of epileptic discharges by from Ad-Tech, Cyberonics, Desitin, Eisai, GlaxoSmithKline, Ingelheim
inhibition and disfacilitation of neuronal firing following the initial Boehringer, Nihon Kohden, Novartis, Pfizer, and UCB Pharma.
high frequency neuronal excitation causing the “spike” [32]. During Stephanie Gollwitzer has served on the scientific advisory board or
cortical stimulation, one should, therefore, be aware of an increased on the speaker's bureau or received grants from of Desitin, Eisai, and
risk of seizures associated with fast activity ADs lacking slow wave UCB Pharma.
components. Müjgan Dogan Onugoren received a travel grant from Fresenius
Moreover, specific AD characteristics point to patients at risk of stim- Medical Care (Bad Homburg, Germany) and obtained honoraria for a
ulation seizures. For instance, ADs lasted significantly longer in patients talk from Eisai (Frankfurt, Germany). All other authors report no
suffering stimulation seizures compared with those remaining seizure- disclosures.
free throughout the session, the cutoff being 9.8 s. Afterdischarges in
seizure patients were often widespread and propagated to electrode
References
contacts in the vicinity of the stimulation site while AD remained
circumscribed and restricted to stimulated electrodes in seizure-free [1] Penfield W. Somatic motor and sensory representation in the cerebral cortex of
patients. The inability to limit AD to a focal region could indicate a pre- mam as studied by electrical stimulation. Brain 1937;60:389–443.
[2] Foerster O. The motor cortex in men in the light of Hughlings Jackson's doctrines.
existent widespread network of pathologic circuits promoting seizure Brain 1936;59.
emergence [29]. In both patient groups, however, ADs could involve re- [3] Adrian ED. The spread of activity in the cerebral cortex. J Physiol 1936;88:127–61.
mote electrodes. As cortical function is imagined to be a property of [4] Blume WT, Jones DC, Pathak P. Properties of after-discharges from cortical electrical
stimulation in focal epilepsies. Clin Neurophysiol 2004;115:982–9.
distributed adaptable networks within the cortex, stimulation could [5] Lesser RP, Lee HW, Webber WR, Prince B, Crone NE, Miglioretti DL. Short-term
activate pathways within these functional networks [5]. This is in line variations in response distribution to cortical stimulation. Brain 2008;131:1528–39.
with insights gained by single-pulse electrical stimulation inducing [6] Wyler AR, Ward Jr AA. Neurons in human epileptic cortex. Response to direct corti-
cal stimulation. J Neurosurg 1981;55:904–8.
corticocortical evoked potentials in order to explore interareal connec-
[7] Lesser RP, Luders H, Klem G, Dinner DS, Morris HH, Hahn J. Cortical afterdischarge
tivity in vivo [33]. and functional response thresholds: results of extraoperative testing. Epilepsia
Stimulation-induced seizures tended to resemble patients' habitual 1984;25:615–21.
[8] Liu C, Yu T, Ren ZW, Xu CP, Wang XY, Qiao L, et al. Properties of afterdischarges from
seizures clinically and electrographically. Even when stimulated distant
electrical stimulation in patients with epilepsy. Epilepsy Res 2017;137:39–44.
from the SOZ, ictal patterns mostly evolved in known tracks of elec- [9] Kalamangalam GP, Tandon N, Slater JD. Dynamic mechanisms underlying
trodes typically bearing onset activity of the patient's habitual seizures. afterdischarge: a human subdural recording study. Clin Neurophysiol 2014;125:
Alternatively, ictal patterns remained in stimulated electrodes. The cor- 1324–38.
[10] Motamedi GK, Lesser RP, Miglioretti DL, Mizuno-Matsumoto Y, Gordon B, Webber
tex neither stimulated nor endogenously epileptogenic did not produce WR, et al. Optimizing parameters for terminating cortical afterdischarges with
ictal discharges. pulse stimulation. Epilepsia 2002;43:836–46.

Please cite this article as: Gollwitzer S, et al, Afterdischarges elicited by cortical electric stimulation in humans: When do they occur and what do
they mean?, Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.09.007
 S. Gollwitzer et al. / Epilepsy & Behavior xxx (xxxx) xxx–xxx 7

[11] Lesser RP, Kim SH, Beyderman L, Miglioretti DL, Webber WR, Bare M, et al. Brief [22] Kovac S, Scott CA, Maglajlija V, Toms N, Rodionov R, Miserocchi A, et al. Comparison
bursts of pulse stimulation terminate afterdischarges caused by cortical stimulation. of bipolar versus monopolar extraoperative electrical cortical stimulation mapping
Neurology 1999;53:2073–81. in patients with focal epilepsy. Clin Neurophysiol 2014;125:667–74.
[12] Ren ZW, Li YJ, Yu T, Ni DY, Zhang GJ, Du W, et al. High-frequency and brief-pulse [23] Tao JX, Ray A, Hawes-Ebersole S, Ebersole JS. Intracranial EEG substrates of scalp EEG
stimulation pulses terminate cortical electrical stimulation-induced afterdischarges. interictal spikes. Epilepsia 2005;46:669–76.
Neural Regen Res 2017;12:938–44. [24] Bernier GP, Richer F, Giard N, Bouvier G, Mercier M, Turmel A, et al. Electrical
[13] Wieser HG, Blume WT, Fish D, Goldensohn E, Hufnagel A, King D, et al. ILAE commis- stimulation of the human brain in epilepsy. Epilepsia 1990;31:513–20.
sion report. Proposal for a new classification of outcome with respect to epileptic [25] Luczak A, Bartho P, Marguet SL, Buzsaki G, Harris KD. Sequential structure of neocor-
seizures following epilepsy surgery. Epilepsia 2001;42:282–6. tical spontaneous activity in vivo. Proc Natl Acad Sci U S A 2007;104:347–52.
[14] Roessler K, Sommer B, Merkel A, Rampp S, Gollwitzer S, Hamer HM, et al. A frameless [26] Li T, Tian C, Scalmani P, Frassoni C, Mantegazza M, Wang Y, et al. Action potential
stereotactic implantation technique for depth electrodes in refractory epilepsy using initiation in neocortical inhibitory interneurons. PLoS Biol 2014;12:e1001944.
intraoperative magnetic resonance imaging. World Neurosurg 2016;94:206–10. [27] Ross ST, Soltesz I. Selective depolarization of interneurons in the early posttraumatic
[15] Lagarde S, Bonini F, McGonigal A, Chauvel P, Gavaret M, Scavarda D, et al. Seizure-onset dentate gyrus: involvement of the Na(+)/K(+)-ATPase. J Neurophysiol 2000;83:
patterns in focal cortical dysplasia and neurodevelopmental tumors: relationship with 2916–30.
surgical prognosis and neuropathologic subtypes. Epilepsia 2016;57:1426–35. [28] Shigeto H, Boongird A, Baker K, Kellinghaus C, Najm I, Luders H. Systematic study of
[16] Talairach J, Bancaud J, Bonis A, Szikla G, Trottier S, Vignal JP, et al. Surgical therapy for the effects of stimulus parameters and stimulus location on afterdischarges elicited
frontal epilepsies. Adv Neurol 1992;57:707–32. by electrical stimulation in the rat. Epilepsy Res 2013;104:17–25.
[17] Bartolomei F, Nica A, Valenti-Hirsch MP, Adam C, Denuelle M. Interpretation of SEEG [29] Sloviter RS. The functional organization of the hippocampal dentate gyrus and its
recordings. Neurophysiol Clin 2018 Feb;48(1):53–7. relevance to the pathogenesis of temporal lobe epilepsy. Ann Neurol 1994;35:
[18] Giagante B, Oddo S, Silva W, Consalvo D, Centurion E, D'Alessio L, et al. Clinical- 640–54.
electroencephalogram patterns at seizure onset in patients with hippocampal [30] Gollwitzer S, Scott CA, Farrell F, Bell GS, de Tisi J, Walker MC, et al. The long-term
sclerosis. Clin Neurophysiol 2003;114:2286–93. course of temporal lobe epilepsy: from unilateral to bilateral interictal epileptiform
[19] Cascino GD, Trenerry MR, So EL, Sharbrough FW, Shin C, Lagerlund TD, et al. Routine discharges in repeated video-EEG monitorings. Epilepsy Behav 2017;68:17–21.
EEG and temporal lobe epilepsy: relation to long-term EEG monitoring, quantitative [31] Ikrar T, Guo N, He K, Besnard A, Levinson S, Hill A, et al. Adult neurogenesis modifies
MRI, and operative outcome. Epilepsia 1996;37:651–6. excitability of the dentate gyrus. Front Neural Circ 2013;7:204.
[20] Luders H, Lesser RP, Dinner DS, Morris HH, Wyllie E, Godoy J. Localization of cortical [32] Sato Y, Doesburg SM, Wong SM, Okanishi T, Anderson R, Nita DA, et al. Dynamic
function: new information from extraoperative monitoring of patients with changes of interictal post-spike slow waves toward seizure onset in focal cortical
epilepsy. Epilepsia 1988;29(Suppl. 2):S56–65. dysplasia type II. Clin Neurophysiol 2015;126:1670–6.
[21] Diehl B, Piao Z, Tkach J, Busch RM, Lapresto E, Najm I, et al. Cortical stimulation for [33] Matsumoto R, Kunieda T, Nair D. Single pulse electrical stimulation to probe func-
language mapping in focal epilepsy: correlations with tractography of the arcuate tional and pathological connectivity in epilepsy. Seizure 2017;44:27–36.
fasciculus. Epilepsia 2010;51:639–46.

Please cite this article as: Gollwitzer S, et al, Afterdischarges elicited by cortical electric stimulation in humans: When do they occur and what do
they mean?, Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.09.007