doi:10.

1093/brain/awx143 BRAIN 2017: Page 1 of 12 | 1

Interictal epileptiform activity outside the

seizure onset zone impacts cognition
Hoameng Ung,1,2 Christian Cazares,1 Ameya Nanivadekar,3 Lohith Kini,1,2
Joost Wagenaar,2,4 Danielle Becker,4 Abba Krieger,5 Timothy Lucas,2,6 Brian Litt1,2,4 and
Kathryn A. Davis2,4

Cognitive deficits are common among epilepsy patients. In these patients, interictal epileptiform discharges, also termed spikes, are
seen routinely on electroencephalography and believed to be associated with transient cognitive impairments. In this study, we
investigated the effect of spikes on memory encoding and retrieval, taking into account the spatial distribution of spikes in relation
to the seizure onset zone as well as anatomical regions of the brain. Sixty-seven patients with medication refractory epilepsy
undergoing continuous intracranial electroencephalography monitoring engaged in a delayed free recall task to test short-term
memory. In this task, subjects were asked to memorize and recall lists of common nouns. We quantified the effect of each spike on
the probability of successful recall using a generalized logistic mixed model. We found that in patients with left lateralized seizure
onset zones, spikes outside the seizure onset zone impacted memory encoding, whereas those within the seizure onset zone did not.
In addition, spikes in the left inferior temporal gyrus, middle temporal gyrus, superior temporal gyrus, and fusiform gyrus during
memory encoding reduced odds of recall by as much as 15% per spike. Spikes also reduced the odds of word retrieval, an effect
that was stronger with spikes outside of the seizure onset zone. These results suggest that seizure onset regions are dysfunctional at
baseline, and support the idea that interictal spikes disrupt cognitive processes related to the underlying tissue.

1 Department of Bioengineering, University of Pennsylvania, Philadelphia, PA, USA

2 Center for Neuroengineering and Therapeutics, University of Pennsylvania, Philadelphia, PA, USA
3 Department of Bioengineering, University of Pittsburgh, Pittsburgh, PA, USA
4 Department of Neurology, University of Pennsylvania, Philadelphia, PA, USA
5 Department of Statistics, University of Pennsylvania, Philadelphia, PA, USA
6 Department of Neurosurgery, University of Pennsylvania, Philadelphia, PA, USA
Correspondence to: Hoameng Ung
301 Hayden Hall
3320 Smith Walk
Philadelphia PA 19104, USA
E-mail: [email protected]

Keywords: intracranial EEG; interictal spikes; epileptiform discharges; memory; cognition

Abbreviations: BA = Brodmann area; GLMM = generalized linear mixed model; SOZ = seizure-onset zone

some of which are not captured by traditional neuropsycho-

Introduction logical tests (Blake et al., 2000; Wang et al., 2011).
Epilepsy is a chronic seizure disorder that affects over 65 Characteristic patterns recorded on EEG such as interictal
million people worldwide (Ngugi et al., 2010). In these pa- spikes, rhythmic bursts, wave discharges, and focal slowing
tients, cognitive and memory deficits are common complaints, are not always accompanied by overt clinical symptoms but

Received November 30, 2016. Revised April 23, 2017. Accepted May 7, 2017.
ß The Author (2017). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved.
For Permissions, please email: [email protected]
2 | BRAIN 2017: Page 2 of 12 H. Ung et al.

can still be detrimental to the patient’s psychosocial function- In this study, we focus on interictal epileptiform dis-
ing and quality of life (Holmes and Lenck-Santini, 2006; charges, also termed spikes, recorded on iEEG. Interictal
Hoppe et al., 2007). Investigation of these electrographic pat- spikes are highly correlated with the presence of epilepsy,
terns relative to cognition date back to 1939 (Schwab, 1939) though their representation with respect to regions of struc-
and has since been associated with impairments in cognition tural and functional abnormalities, influence on patient be-
and memory (Rausch et al., 1978; Aarts et al., 1984; Holmes haviour and physiology, and relationship to ictal activity
and Lenck-Santini, 2006; Kleen et al., 2010, 2013; Gelinas has not yet been well defined (Spencer et al., 2008). EEG-
et al., 2016). Elucidating the relationship between these elec- functional MRI studies have confirmed that interictal spikes
trophysiological patterns and their functional consequences is may be separated into different populations with some re-
a crucial step in improving treatment and quality of life for sulting in modifications of metabolism well beyond the clin-
patients with epilepsy. ically identified epileptic focus (Kobayashi et al., 2006).
Parameters for functional assessment of patients with epi- This suggests that spiking activity in the epileptic foci
lepsy involve neuropsychological tests that explore cogni- may be deleterious for a larger section of the brain
tive, behavioural, linguistic or motor impairment. These (beyond the foci) and may interrupt normal electrophysi-
tests provide information on a global level such as whether ology and functioning, a concept introduced as ‘nociferous
a person can safely live by themselves, return back to work cortex’ by Penfield and Jasper (1954). Past studies have
or school, or drive. While important for patient manage- suggested a negative correlation of spiking with global
ment, these measures are limited in their ability to assess neuropsychological tests of memory and intelligence
the anatomical location of abnormalities as well as the cor- (Rausch et al., 1978; Aarts et al., 1984) and a specific
relation with EEG phenomena such as seizures and inter- effect of hippocampal spikes on cognition (Kleen et al.,
ictal spikes. For these reasons, a more precise measure of 2010, 2013; Gelinas et al., 2016). In an independent
cognitive function is necessary to understand the mechan- study recently performed by the Restoring Active
Memory collaborative research group, inferior temporal,
isms underlying the prominent memory and cognitive com-
middle temporal, and parietal spikes were shown to nega-
plaints that plague a significant portion of patients with
tively impact memory processes (Horak et al., 2016). We
epilepsy. Here, we use a controlled memory task during
aim to more specifically quantify the effect of spikes on a
intracranial EEG monitoring to evaluate the impact of
whole brain level and with respect to the seizure onset zone
interictal spikes on memory processes with high spatial
(SOZ).
and temporal resolution.
Improper memory encoding and recall in correlation
Intracranial EEG (iEEG) recordings provide more precise
with interictal spiking places emphasis on targeted therapy
recordings of brain activity than non-invasive EEG. These
for cognitive dysfunction and potential improvement in the
data provide an opportunity to study the electrophysiolo-
localization of epileptic foci (Blake et al., 2000). As spikes
gical correlates of a wide range of cognitive processes with
are helpful in localization of the SOZ only in some pa-
greater detail (Rausch et al., 1978; Lachaux et al., 2012).
tients (Marsh et al., 2010; Goncharova et al., 2013),
For example, iEEG has been used to study the neural basis
understanding the differences between populations of
of human cognition using non-epileptic regions of the spikes may better aid onset localization. We hypothesize
brain as models for normal healthy brain and as controls. that regional interictal spiking will have a detrimental
Typical electrophysiological patterns are consistently seen impact on cognition by disrupting the subsequent
across various performance tasks. In studies of memory, memory effect, suggesting that altering epileptic networks
there is evidence that electrophysiological and haemo- to reduce spiking in normal brain may positively impact
dynamic changes occur during encoding, termed the sub- cognition. Furthermore, regions underlying the SOZ are
sequent memory effect (Paller and Wagner, 2002; believed to be damaged from not only the initial insult
Sederberg et al., 2003). High frequency oscillatory activity but also from recurrent seizures, leading to neuronal loss
of neurons has been associated with specific cortical net- and cognitive decline (Sutula et al., 2003). Quantifying the
work states and has been used to quantify the electro- effect of spikes on cognition relative to the SOZ may help
physiological mechanisms of memory formation and elucidate the function (or dysfunction) of the involved
recall. Gamma oscillatory activity, for example, exhibits tissue.
anatomical, temporal and functional specificity (Jacobs The overarching hypotheses that were tested during this
and Kahana, 2010). Early studies showed that increases study are (i) interictal spikes disrupt memory processes;
in gamma activity in the hippocampus and the left tem- (ii) there is an observable functional anatomy in which
poral and frontal cortices is associated with successful these spikes lie; and (iii) epileptiform activity outside of
memory formation, consistent with previous functional the seizure onset zone disrupt memory processes whereas
MRI studies (Sederberg et al., 2007). Therefore the mag- spikes within the seizure onset zone are not deleterious to
nitude of these oscillations during encoding indicates that memory. We address these hypotheses using automated
synchrony in widespread networks of cortical regions can spike detection in iEEG during a delayed free recall
serve as a predictor for successful memory recall task and identify the brain regions implicated in incorrect
(Sederberg et al., 2003). recall.
Spikes outside seizure onset zone impacts memory BRAIN 2017: Page 3 of 12 | 3

(Murdock, 1962). During this distracter period subjects were

Materials and methods asked to solve a series of simple arithmetic problems composed
of A + B + C, where A, B, and C consisted of randomly se-
Participants lected one-digit positive integers. Participants responded by
typing their answer onto the keyboard, with feedback provided
Sixty-seven patients with drug resistant epilepsy at the Hospital through a high-pitched tone for correct answers and a low-
of the University of Pennsylvania (n = 16) and Thomas Jefferson pitched tone for incorrect answers. Each distracter period
University Hospital (n = 51) were included in this study and lasted for 1600–2000 ms and was followed by a 300 ms tone
comprised 45 males and 22 females, with an average age of concurrent with a row of asterisks that signalled the start of
35.55 [range 15–57, standard deviation (SD) = 12.17]. Verbal the recall period. The recall period lasted for 4500 ms during
intelligence quotient (IQ) was determined using the Wechsler which subjects were asked to recall words in any order pre-
Adult Intelligence Scale. Additional subject information is given sented during the encoding period and their vocal responses
in Supplementary Table 1. A total of 6144 intracranial electrodes were recorded. Words that were presented during the encoding
were implanted across all patients during the clinical monitoring period and retrieved during the recall period were considered
period and included both depth and subdural electrodes. correctly encoded and correctly recalled. Furthermore, words
Medications were reduced and or withdrawn during the moni- from prior encoding periods that were retrieved during a recall
toring period. Cognitive testing (described below) was not per- period were considered incorrectly recalled.
formed after a seizure until the patient was back to prior baseline The computer sent a pulse to an unused recording channel in
neurologic status, which varied for each subject depending upon order to synchronize the behavioural events during the memory
the seizure semiology, post-ictal symptoms, and abortive medi- task with the electrophysiological recordings. The time stamps
cations given during a seizure (e.g. lorazepam). Typically, cogni- associated with these pulses were used to annotate the iEEG
tive testing was performed a minimum of 3 h after a generalized recordings. Annotations and iEEG recordings were converted
tonic clonic seizure and 1 h after a partial seizure. All partici- to the Multiscale Electrophysiology Format (Brinkmann et al.,
pants provided informed consent with procedures approved by 2009) and uploaded to IEEG.org for analysis.
the Institutional Review Board at the University of Pennsylvania
and Thomas Jefferson University Hospital.
Preprocessing and removal of
Delayed free recall task artefact channels
A delayed free recall episodic memory task was used to inves- Sampling rates varied from 400 to 2000 Hz. Artefact channels
tigate memory encoding and recall. For each trial, subjects were were identified by calculating the line length feature across each
asked to memorize a word list composed of 15 common high channel and those that differ by more than four times the mean
frequency nouns, chosen at random and without replacement across all channels were removed (Esteller et al., 2001). This
from a pool of high-frequency English nouns (http://memory. process is intended to remove grossly artefactual channels.
psych.upenn.edu/WordPools), followed by a distractor task that
consisted of simple arithmetic problems (Sederberg et al., 2003;
Burke et al., 2013). After a tone, subjects engaged in free recall,
Spike detection
vocalizing presented words in any order. A schematic of one Despite complex methods in the literature, automatic spike
word list/trial is provided in Fig. 1. Each subject received 12–60 detection has proved to be a difficult task to perfect. Many
such word lists to encode and recall in each session, the number detectors suffer from a high false positive rate, as some arte-
of trials per subject as well as the number of sessions depended facts have spike-like morphology (Wilson and Emerson, 2002)
on the subject’s interest and availability for testing. and performance varies from dataset to dataset. Another
A computer program presented stimuli and recorded subject reason for this difficulty is reflected in the poor inter-rater re-
responses. Each trial began with a plus sign to alert the subject liability even among experts, which may range from 41–80%
to an upcoming presentation of words. The plus sign appeared (Gaspard et al., 2014). However, automatic spike detection is
for 1600 ms followed by an 800–1200 ms blank interval. Each a reasonable solution for our analysis because (i) we adopt a
word was then presented for 1600 ms followed by an 800– highly sensitive algorithm; (ii) our statistical analysis compares
1200 ms blank interstimulus interval. Subjects were asked to two groups that will be equally affected by false positives; and
read each word aloud or to themselves to ensure attentiveness. (iii) it provides an objective and quantifiable detection method.
Following the presentation of a list of words, a short distracter Thus, false positive detections are expected to be equally dis-
period was introduced between the encoding and recall periods tributed within each individual between word presentations
to reduce the recency effect often observed for free recall tasks and any differences in means will be investigated.

Figure 1 Delayed Free Recall Task schematic. Blue indicates correctly recalled words and orange indicates incorrectly recalled words.
Subjects were presented 15 words during the encoding period, followed by a distractor task that consisted of simple arithmetic problems. After a
tone, subjects engaged in free recall, vocalizing presented words in any order. Subjects were presented with 12–60 word lists in each session.
4 | BRAIN 2017: Page 4 of 12 H. Ung et al.

An algorithm published by Janca et al. (2014) was used to where bj Nð0;  2 Þ; akðjÞ Nð0;  2j Þ, ijk denotes the probability
automatically detect spikes for all patients. This algorithm of successful recall for word i of subject j, session k, bj and akðjÞ
applies a signal envelope to identify spikes by modelling back- are random effects for differences in mean recall rate as well as
ground activity and determining transient outliers. Briefly, sig- intersession (k) variability, respectively for each subject.
nals were downsampled to 200 Hz before 10–60 Hz bandpass A random effect was included for each subject nested by
and 60 Hz notch filters were applied. For each channel, the session. This allows us to model variability between subjects
signal envelope was calculated with the absolute value of the (with different baseline recall rates) as well as variability be-
Hilbert transform. Moving windows of 5 s with 4-s overlap tween different sessions within a subject. A logit link function
were used to model a log-normal statistical distribution of the permits us to model a binary outcome (recalled versus not
signal envelope. A threshold of 1 [Mode + Median] was used recalled) and to interpret the estimated coefficients probabilis-
for the initial detection of spikes, where 1 = 3.65, determined tically. The fixed effects represent the mean effect across all
empirically through cross-validation by the original authors subjects after removal of intersubject and intersession variabil-
(Janca et al., 2014). Following initial detections and to im- ity. To confirm the need for a mixed, nested model, a likeli-
prove our positive predictive rate, we added a spatial filter hood ratio test was used to test model fit before and after
to identify spikes across multiple channels. Any spikes within sequential addition of random effects and covariates
200 ms on more than one channel were combined and treated (Supplementary Tables 2 and 4). Effect sizes (odds ratios,
as one spike. A subset of candidate spikes was randomly se- OR) and confidence intervals (CI) are reported with signifi-
lected across all patients and validated by a board certified cance by Wald statistics when testing multiple parameters.
epileptologist (K.D.) to ensure adequate performance, specific- Likelihood ratios are reported when testing single parameters
ally positive predictive value. (regional analysis). Further information about the model is
The encoding phase was defined to be the period after a provided in the Supplementary material.
word stimulus prior to the subsequent word stimulus. The
number of spikes during this period was extracted for all
words. Spikes were also categorized into anatomical locations Seizure onset zone
determined by the Talairach coordinates of the electrodes fol- We also sought to determine the effect of spikes on identified
lowing co-registration of post-implant CT to pre-implant T1- SOZs. For this analysis, 57 patients with clinically localized
weighted MRI. Spikes that occurred in multiple brain regions SOZ were included. The clinical localization process varied for
or bilaterally were considered independently for each region. each patient and was directed towards identification of the
While the delayed free recall task is most suited to investi- seizure aetiology and onset region(s). This process typically
gate the encoding period, we also conducted an analysis of the includes video monitoring and patient reports to localize clin-
retrieval period based on methods from prior literature (Burke ical semiology and correlate findings from imaging as well as
et al., 2014). The recall phase was defined to be a 1-s period electrophysiology. The final SOZ is determined through expert
prior to a successful word recall. This is compared to a base- consensus. To determine the effect of spikes relative to the
line period during unsuccessful recall, which is defined as a 1-s SOZ, channels were first divided into seizure onset and non-
period in free recall trials in the same subject and session that seizure onset channels. Since spikes may occur across multiple
are at minimum 3000 ms away from any vocal utterance. This channels, if any channel within the SOZ was involved, the
1-s period is time-matched to beginning of each recall session. spike was categorized as a seizure onset spike. Similarly,
Spikes were extracted from all periods. spikes completely outside the SOZ were classified as non-seiz-
ure onset spikes. For both encoding and retrieval, the GLMM
was refitted with the addition of these two covariates as well
Statistical analysis as an interaction term for clinical SOZ lateralization.
A linear model was used to test the effect of subject level
variables (age, sex, and verbal IQ) versus mean recall rate
and mean spike rate across all patients. Welch’s t-test was
Regional analysis
used to determine differences in mean recall rates and spikes All 67 subjects were included in the regional analysis. Each
per electrode between left and right lateralized patients. To electrode was grouped into regions corresponding to level 5
model the effect of spikes on correct versus incorrect recall of the Talairach atlas, which includes Brodmann areas and the
on a word-by-word basis, we use a generalized linear mixed hippocampus (Lancaster et al., 2000). Due to variability in
model (GLMM) with a logit link function. The GLMM was fit electrode positioning across patients, independent GLMMs
to predict successful recall with the serial order of word pres- were fit on regions containing electrodes from 20 or more
entation, age, verbal IQ, and the number of spikes in the patients. For each region, a null model containing serial
encoding period as fixed effects (Equation 1). To determine word position, verbal IQ, and age was fit, and an alternative
the effect of spikes on retrieval, a similar GLMM was fit for model with the addition of regional spikes was fit. Significance
spikes within the retrieval period (Supplemental Table 4). A was determined by the likelihood ratio test and adjusted for
region of interest analysis for the encoding period was con- familywise error with the Holm-Bonferroni method (Holm,
ducted by varying the spike count to correspond with each 1979).
Brodmann area (BA).
 
ijk Code availability
log ¼ b1 agej þ b2 verbal iqj þ b3 word orderi
1  ijk ð1Þ
MATLAB was used for spike detection and extraction. R was
þb4 spike countijk þ b0 þ bj þ akðjÞ used for statistical analysis and generation of figures. Scripts
Spikes outside seizure onset zone impacts memory BRAIN 2017: Page 5 of 12 | 5

for the statistical analysis and generation of figures are avail- [2 ð1Þ = 8.64, P = 0.003]. In patients with left lateralized
able at http://github.com/hoameng/cognitive-spike-2016. SOZ, spikes outside of the SOZ had a significant effect
on recall, whereas spikes within did not. In patients with
right lateralized onset regions, spikes were not significantly
Results associated with recall performance, whether within or out-
side the SOZ (Table 1). In the majority of patients, spikes
A total of 244/6144 channels were identified as artefact were more frequent outside the SOZ [P 5 0.001,
and removed from the analysis. Over 900 detected spikes t(66) = 7.78, Supplementary Fig. 5].
were randomly selected from all patients and validated by a The top 10 regions from the regional analysis are given
board certified epileptologist (K.D.). Positive predictive rate in Table 2. The odds ratios for all regions are shown in Fig.
was 72.2%. False positive rate was 15.5%. Of all detected 3. Left temporal lobe structures such as the fusiform gyrus
spikes, 12.3% were deemed indeterminate. An example of (BA 37) and inferior temporal gyrus (BA 20) were most
a detected spike is shown in Supplementary Fig. 1. significant after multiple comparisons correction for
Subjects on average recalled at a rate of 24% (range: 4–
family-wise error with the Holm-Bonferroni method
48%, SD = 9%). Older age was significantly associated
(Holm, 1979). The middle and superior temporal gyri
with a decrease in mean recall rates [t(54) = 2.9,
(BA 21/22) were also significant after correction. A plot
P = 0.005, R2fadjg = 0.1], but not average spike rates
of the per cent change in epileptiform discharges during
(Supplementary Fig. 2A). There was no significant differ-
failed recall are shown in Fig. 4, separated by SOZ lateral-
ence in mean recall rates or mean spike rates by sex
ization and the significant regions from the above regional
(Supplementary Fig. 2B). Initial words were recalled with
analysis. These three regions are shown along with elec-
greater accuracy (Supplementary Fig. 2C). Greater verbal
trode locations across all subjects in Fig. 5.
IQ was associated with greater recall accuracy
[t(40) = 3.133, P = 0.003, R2fadjg = 0.177] (Supplementary
Fig. 3). Averaged across patients, 23.7% (SD = 11.7%) of
words that failed to be recalled occurred with no spikes. Retrieval
Patients with left lateralized onset regions (n = 23) had During the retrieval period, increased spiking correlated
lower mean recall rates than patients with right lateralized with decreased retrieval. In other words, a greater
onsets (n = 26) [t(47) = 3.1, P = 0.003, d = 0.9] (Fig. 2). In number of spikes occurred during unsuccessful retrieval
addition, a greater percentage of spikes per electrode than 1 s prior to a successful recall (Table 3). For left
occurred ipsilateral to the seizure onset region [left onset: lateralized patients, both spikes within and outside of the
t(36.3) = 4.2, P 5 0.001, d = 1.33; right onset: t(38.1) = 4.5, SOZ decreased retrieval. For right lateralized patients, only
P 5 0.001, d = 1.34] (Fig. 2). spikes outside the SOZ significantly reduced retrieval. In
either case, spikes outside the SOZ had a larger effect on
Encoding retrieval versus spikes within the SOZ. Regional analysis
was not performed due to insufficient time-matched periods
During the encoding phase, an increased number of spikes across patients and Brodmann areas.
across all electrodes were associated with decreased recall

Figure 2 Mean recall (A) and spike lateralization (B) grouped by seizure onset lateralization. Boxplots indicate median, 25th, and
75th percentiles. Significance at *P 5 0.01, ***
P 5 0.001.
6 | BRAIN 2017: Page 6 of 12 H. Ung et al.

Table 1 Estimated odds ratios of effect of spikes relative to the seizure onset zone during encoding

OR (95% CI) Z P
Right lateralized SOZ (n = 26)
Spikes within SOZ 1.005 (0.972–1.039) 1.189 0.235
Spikes outside SOZ 0.975 (0.935–1.016) 0.318 0.750#
Left lateralized SOZ (n = 23)
Spikes within SOZ 0.979 (0.925–1.036) 0.719 0.472
Spikes outside SOZ 0.925 (0.886–0.964) 3.647 0.000265*,#

A logistic GLMM model was fit to a binary response variable indicating recall success (1) or failure (0). Covariates included spikes within the SOZ, outside the SOZ, and interaction of
each with clinical SOZ lateralization. ORs are adjusted for age, serial word position, and verbal IQ. Patients with bilateral onsets were not included. OR = odds ratio, CI = confidence
interval, significance determined by Wald statistics.
*
P 5 0.05; #significant difference between left and right lateralized SOZ patients [increased odds of recall for spikes outside the SOZ if right lateralized: odds = 1.087 (1.013–1.112),
z = 3.077, P = 0.002].

Table 2 Regional effect of spikes during encoding

n L/R Region Brodmann area Odds (95% CI) 2 ð1Þ P (adjusted)

22 L Fusiform gyrus BA 37 0.847 (0.782–0.918) 44.0 50.001***
45 L Inferior temporal gyrus BA 20 0.915 (0.893–0.937) 30.2 50.001***
43 L Middle temporal gyrus BA 21 0.934 (0.908–0.962) 29.8 50.001***
22 L Superior temporal gyrus BA 22 0.909 (0.861–0.961) 27.1 0.0304*
24 L Peristriate cortex BA 19 0.906 (0.828–0.992) 20.1 0.0650
21 L Fusiform gyrus BA 36 0.955 (0.880 1.036) 15.4 0.444
28 R Motor cortex BA 4, 6 0.954 (0.909 1.001) 12.7 1.000
28 R Superior temporal gyrus BA 22 0.936 (0.902–0.972) 7.3 1.000
35 L Sup/mid temporal gyrus BA 38 0.948 (0.919–0.977) 6.4 1.000
25 R Frontal cortex BA 11 0.985 (0.958–1.012) 5.8 1.000

Number of patients with electrodes in corresponding regions is shown, along with laterality. Each region is derived from the Talairach atlas, level 5, shown with associated gyri/lobes.
The effect of each spike on the odds of recall is also given. Effect size and adjusted P-values from the likelihood ratio test are given after controlling for family-wise error with the
Holm-Bonferroni method. significant at *P = 0.05, ***P = 0.001

represented across all patients. In this group, our detections

Discussion performed with a positive predictive value of 72%.
We showed that in patients with epilepsy, interictal epilepti- Furthermore, we observed an expected association between
form spikes that occur after word presentation lead to spike and seizure onset lateralization (Fig. 2).
greater likelihood of failed recall, which suggests that During encoding, the observation that spikes outside the
spikes disrupt short term memory encoding. Spikes during SOZ affect recall greater than spikes within the SOZ sug-
the retrieval period also led to reduced recall. These effects gests that epileptic regions may have some degree of dys-
were strongest with left lateralized spikes. During the function at baseline that, in a verbal word memory task,
encoding period, spikes outside the SOZ in left lateralized manifests most prominently in left lateralized patients
patients significantly reduced recall, a finding not present in (Table 1). These observations related to the SOZ are sup-
right lateralized patients. Similarly, spikes regardless of ported neuropathologically by studies showing neuronal
SOZ significantly affected retrieval, though again the loss in more than 90% of temporal lobe epilepsies as de-
effect was greater with spikes outside the SOZ. Finally, tailed in a review by Sutula et al. (2003). In addition, elec-
we observed that spikes in specific temporal lobe structures trophysiological network studies have also shown dynamic
functionally implicated in verbal and auditory word pro- uncoupling of the seizure onset region (Warren et al., 2010;
cessing and verbal memory were most impactful. Our re- Burns et al., 2014). This finding has potential implications
gional observations support the importance of the temporal in the clinical realm. Clinical tests are routinely used in an
neocortex in memory encoding and extend it by elucidating attempt to identify regions of eloquent cortex in order to
the spatial distribution of spikes relative to the seizure onset weigh the functional consequences versus the potential
zone, the lateralization of spikes, and the quantification of benefits of surgical resection. As new focal surgical tech-
the effect of spikes in the verbal word recall task. niques, such as laser ablation, have led to improved cogni-
As our findings rest on the accuracy of our spike detec- tive outcomes by limiting damage to collateral structures
tions, we vetted a random subset of our spikes equally (Drane et al., 2015), limiting the extent of resection may
Spikes outside seizure onset zone impacts memory BRAIN 2017: Page 7 of 12 | 7

Figure 3 Estimated odds of successful recall for each region. Mean odds of successful recall per spike during the encoding period are
shown along with 95% CIs. Odds 51 indicate a decrease in the odds of successful recall per spike. Red indicates significant after multiple
comparisons correction. L/R = left/right hemisphere.

lead to improved cognitive outcome. This is supported by cortical interictal epileptiform discharges. Specifically,
the concept of a ‘nociferous cortex’, where the epilepto- that the left-sided and right-sided spikes produce verbal
genic zone may impair the function of other regions, in and spatial task impairments, respectively (Aarts et al.,
this case through spikes. Spikes have previously been 1984).
shown to provide adjunctive information in the clinical Spikes in the left inferior temporal gyrus (BA 20) and
mapping of the seizure onset region in certain populations nearby fusiform gyrus (BA 37) most significantly disrupted
(Marsh et al., 2010), and our results suggest that identify- memory encoding. These regions form the ventral spatial
ing spike populations, namely those that do not impact pathway of visual memory and are involved in visual pro-
cognition, may improve SOZ localization, surgical plan- cessing of words. Numerous studies have associated infer-
ning, and cognitive outcomes from surgical resection. ior temporal lobe function with working memory tasks in
The regional localization of spikes may also elucidate the both humans and primates (Nobre et al., 1994; Wagner
function of underlying tissue. Our regional analysis sug- et al., 1998; Kondo et al., 2005). Wagner et al. (1998)
gests that spikes in primarily left-sided structures lead to demonstrated through functional MRI that the ability to
poor recall, which agree with previous studies on cognitive remember a verbal experience is predicted by the activation
impairment that lateralize functional disruption based on of regions in the left prefrontal and temporal cortices.
8 | BRAIN 2017: Page 8 of 12 H. Ung et al.

Hamame et al. (2012) recently observed increased gamma specifically has been functionally implicated in word recog-
activity in the inferior temporal gyrus corresponding to nition and is referred to as the visual word form area in a
increased visuospatial working memory load, suggesting thorough review by McCandliss et al. (2003). Literature in
that this region acts as a ‘visual sketchpad’ during cognition implicates that a critical process in visual word
memory maintenance. This, however, was a trial of only recognition groups the letters of a word together to an
one patient, and we showed here in 67 patients that spikes integrated perceptual unit (a ‘visual word form’), a function
in this region likely interfere with memory encoding by the fusiform gyrus participates in. The visual word form
interrupting mental imagery of presented words. area within the left inferior temporal cortex has been
During word presentation, subjects were asked to read shown to have increased activity through functional MRI
each word aloud as it was presented, which likely activates during visual word recognition as well as an event related
both visual and auditory structures involved in word pro- potential 250 ms after word presentation (Posner and
cessing in addition to memory. The fusiform gyrus (BA 37) McCandliss, 1992; Cohen et al., 2002). Wagner et al.
(1998) and Hamame et al. (2012) similarly found correlated
activity in the fusiform gyrus during verbal memory tasks.
Our findings support this body of literature and extend it by
suggesting that spikes in the fusiform gyrus may impede
successful visual recognition of a word-form and mainten-
ance of verbal memory. We showed in Fig. 4 that in words
that were not recalled, there was an increased in spike rate
in BA 37 during the corresponding encoding period.
Though we were not able to determine a causal relationship
between spikes and word recall, these differences begin to
appear as more words were shown in a given word list,
indicating that the effect of spikes may manifest as a patient
finds greater difficulty in memorizing words.
There is extensive work covering the middle and superior
temporal lobe’s involvement in memory processing, specif-
ically regarding the medial structures such as the hippocam-
pus and parahippocampal gyrus (Squire and Zola-Morgan,
1991; Squire et al., 2004; Eichenbaum et al., 2007; Baxter,
Figure 4 Per cent change in spike count during failed 2009; Raslau et al., 2015). Surprisingly, we did not find
encoding for each patient. Vertical axis represents the per cent these structures to be significantly affected by spikes, with
change in spike count. The horizontal axis represents the left tem- the exception of BA 21 and 22. The relationship of these
poral lobe as well as the significant left temporal regions from our more lateral regions to memory is unclear. Several imaging
regional analysis. Boxplots represent the percent change of spikes
studies showed that both BA 21 and 22 are involved with
across all patients with the given SOZ lateralization. Only the recall
processing of auditory word-form as well as sentence gen-
of words in serial position 45 were included to account for the
primacy effect. eration (Grasby et al., 1993; Zahn et al., 2000; Ahmad
et al., 2003; Monti et al., 2009). In this case, it may play

Figure 5 Electrode coverage and significant regions. Electrodes (n = 6144; blue dots) across all patients are shown in a 3D view of the left
hemisphere (A) and an axial slice (B) on a template brain in Montreal Neurological Institute (MNI) space.
Spikes outside seizure onset zone impacts memory BRAIN 2017: Page 9 of 12 | 9

Table 3 Estimated odds ratios of effect of spikes relative to the seizure onset zone during retrieval

OR (95% CI) Z P
Right lateralized SOZ (n = 26)
Spikes within SOZ 0.995 (0.952–1.040) 0.213 0.831
Spikes outside SOZ 0.847 (0.810–0.886) 7.198 50.001*
Left lateralized SOZ (n = 22)
Spikes within SOZ 0.796 (0.738–0.860) 5.829 50.001*,#
Spikes outside SOZ 0.750 (0.705–0.798) 9.178 50.001*,#

A logistic GLMM model was fit to a binary response variable indicating recall success (1) or failure (0). Covariates included spikes within the SOZ, outside the SOZ, and interaction of
each with clinical SOZ lateralization. Odds ratios (OR) are adjusted for age and serial word position. Patients with bilateral onsets were not included. CI = confidence interval,
significance determined by Wald statistics. Significance at *P 5 0.05; #significant difference between left and right lateralized SOZ patients [increased odds of retrieval for spikes if
right lateralized; outside SOZ: odds = 1.130 (1.055–1.210), z = 3.49, P 5 0.001, within SOZ: odds = 1.250 (1.198–1.304), z = 5.248, P 5 0.001].

a role similar to the fusiform gyrus for auditory stimuli as occur more often in the SOZ (Supplementary Fig. 5).
subjects were asked to read each word aloud, but add- Furthermore, our analysis investigated whole-brain sublo-
itional work is necessary to further parse out these regions’ bar regions with greater specificity at the level of Brodmann
function in relation to memory. areas versus a priori selection of six brain regions. For ex-
Of note, a recent study of 10 patients with temporal lobe ample, our strongest effect was found in the left fusiform
epilepsy showed that right-sided hippocampal discharges gyrus, part of the inferior temporal lobe, which has associ-
significantly reduced performance in a memory retrieval ations with verbal word formation instead of with regions
task in both humans and rats, but not encoding (Kleen traditionally thought to be involved in memory. While
et al., 2010, 2013). Though we were unable to directly some of the differences between the two studies are difficult
test retrieval on a regional basis, our findings agree in to reconcile, we believe the main complementary findings
that there was no effect of hippocampal spikes on encod- are an important step to understanding the impact of spikes
ing. This is to a certain extent surprising, as the hippocam- on cognition.
pus is believed to play a role in the encoding and retrieval During retrieval, spikes reduced recall in both left and
of unconsolidated memory (Raslau et al., 2014). However, right lateralized patients (Table 3). We also observed a
it is possible that existing hippocampal pathology prevents greater effect of spikes as well as a diminished difference
the effect of spikes from manifesting, similar to the rela- between effects relative to the SOZ. The differences be-
tionship we found with the SOZ. Relatively healthy tissue, tween retrieval and encoding are difficult to explain, but
such as those surrounding the SOZ, may exhibit greater are likely related to the involvement of other functional
spike sensitivity, allowing us to observe functional conse- regions or circuits during active retrieval, which include
quences in the present study. regions involved in contextual retrieval (Friedman and
Furthermore, we found that spikes did not impede encod- Johnson, 2000). This observation has been seen previously
ing in patients with right lateralized SOZs (Table 1). Taken by Kleen et al. (2013) who demonstrated that hippocampal
into context with our previous findings showing that left- spikes reduced performance specifically during retrieval and
sided spikes are involved in memory encoding, this can be not encoding. Previous imaging studies also suggest in-
explained by the lack of spikes in the left hemisphere in volvement of right lateralized structures, such as the pre-
right lateralized patients (Fig. 2). However, even in right cuneus specifically during retrieval, which may further
lateralized patients there is increased spiking during the explain our observed effect of right hemispheric spikes out-
encoding of words that were not recalled (Fig. 4), which side the SOZ (Fletcher et al., 1995). Interestingly, this is
provides further evidence that the effect of spikes is region- supported by Kleen et al. (2013), who showed that the
ally dependent. In an attempt to further identify the regional effect of hippocampal spikes is lateralized to the right hemi-
effects, we conducted a separate analysis only on right sphere during retrieval. This may suggest that in patients
lateralized patients, but no regions (on either hemisphere) with left temporal epilepsy the language dominance has
reached significance during encoding (Supplementary shifted to the right temporal lobe, though in our case left
Table 3). hemisphere spikes still affected retrieval in patients with left
Inferior temporal lobe involvement in memory encoding lateralized SOZ. Further studies using cognitive tasks de-
is also supported in a recent finding by the Restoring Active signed to test retrieval are needed to more precisely deter-
Memory collaborative research group (Horak et al., 2016). mine the regional effects of spikes.
Interestingly, they reported an impact of spikes in parietal
lobe structures and in the SOZ, which we did not find. In
fact, our analysis suggests that spikes impair recall only if
Quantitative modelling
outside the SOZ, and in these cases, only in left lateralized Interpretation of the model estimates show that each spike
patients. This may be due to variability in the patient popu- in BA 37 reduces the odds of recall by 15%, where in BA
lation, where for the majority of patients, spikes did not 20 the odds of recall are reduced by 8% per spike.
10 | BRAIN 2017: Page 10 of 12 H. Ung et al.

Supplementary Fig. 4 show the corresponding predicted by spikes irrespective of the SOZ in left lateralized patients,
recall percentages per spike estimated from our model. and only by spikes outside the SOZ in right lateralized
This implies that increased spiking increases the probability patients. These results suggest that spikes are not benign
of failed recall and is not necessarily an all-or-nothing and can disrupt visual word recognition as well as verbal
mechanism. These findings show that spikes focally and memory processes. In addition, our findings support the use
additively impact function in a spatially dependent of surgical interventions that spare cortex outside of the
manner within a given region. Finally, it is important to SOZ. Specifically, the observation that spikes affect a par-
note that spike occurrence across BA 20, 21, 22 and 37 are ticular region of the brain suggests that this region may be
correlated (r 4 0.46), which is expected as these regions functional in the absence of spikes. Further study is war-
may play similar functional roles, though further work is ranted to determine the magnitude of this effect relative to
necessary to determine whether this contribution is known cognitive deficits in epilepsy patients.
additive.

Limitations Acknowledgments
Although automated spike detection is a difficult challenge We thank the Penn Computational Memory Lab for pro-
to perfect and validate in the field, we believe an automated viding data and supporting information.
algorithm that performs with acceptable true positive rate
allows us to still make comparisons between groups.
Furthermore, an automated algorithm allows objective de-
tection of spikes that does not suffer from differences in Funding
inter-rater reliability. This study was funded by the National Institutes of Health
We have shown that epileptiform discharges in various (NIH), the Thornton Foundation, and the Mirowski
regions of the temporal lobe affect verbal word processing Foundation grants through the University of Pennsylvania:
and memory encoding. Our ability to discern influence in NIH P20NS080181, U01NS073557, K23NS073801,
other structures is limited by electrode coverage (Fig. 5). R25GM071745, R01NS099348, and T32NS091006. The
Notably, our electrodes are primarily cortical, with the ex- International Epilepsy Electrophysiology Portal is funded
ception of medial temporal subcortical structures interro- by the NIH (U24NS063930).
gated by depth electrodes. It is also important to note
that when assessing the effect of spikes on retrieval, we
make an assumption that any time-matched period without
any vocal utterance is a period of unsuccessful retrieval. Supplementary material
Another task, such as the Sternberg task where recall per-
Supplementary material is available at Brain online.
iods are controlled, should be used to verify these findings.
There is rich literature supporting the notion that inter-
ictal epileptiform discharges are associated with cognitive
impairments, described by Aarts et al. (1984) as ‘transitory References
cognitive impairment’. Though spikes may influence
Aarts JH, Binnie CD, Smit AM, Wilkins AJ. Selective cognitive im-
memory and other cognitive processes, it remains unclear pairment during focal and generalized epileptiform EEG activity.
whether spikes should be treated in clinical management of Brain 1984; 107 (Pt 1): 293–308.
otherwise well controlled epilepsy (Binnie, 2003). In add- Ahmad Z, Balsamo LM, Sachs BC, Xu B, Gaillard WD. Auditory
ition, while we have shown that memory encoding is af- comprehension of language in young children: neural networks iden-
tified with fMRI. Neurology 2003; 60: 1598–605.
fected by spikes in a spatially distributed manner, the
Baxter MG. Involvement of medial temporal lobe structures in
magnitude of the effect is fairly small depending on the memory and perception. Neuron 2009; 61: 667–77.
region (roughly a 5% decline in probability for spikes out- Binnie CD. Cognitive impairment during epileptic form discharges: is it
side the seizure onset zone; Supplementary Fig. 4). Coupled ever justifiable to treat the EEG? Lancet Neurol 2003; 2: 725–30.
with the observation that recall percentage is limited to Blake RV, Wroe SJ, Breen EK, McCarthy RA. Accelerated forgetting
in patients with epilepsy: evidence for an impairment in memory
begin with, the consequence of spikes on overall patient
consolidation. Brain 2000; 123: 472–83.
cognition and quality of life needs to be further studied. Brinkmann BH, Bower MR, Stengel KA, Worrell GA, Stead M.
Multiscale electrophysiology format: an open-source electrophysi-
ology format using data compression, encryption, and cyclic redun-
Conclusions dancy check. In: Proceedings of the 31st Annual International
Conference of the IEEE Engineering in Medicine and Biology
Society: Engineering the Future of Biomedicine, Minneapolis, MN,
In this study we have shown that interictal epileptiform
EMBC, Sept 3–6 2009. p. 7083–6.
spikes impede memory encoding when present outside the Burke JF, Sharan AD, Sperling MR, Ramayya AG, Evans JJ, Healey
SOZ and within the left temporal lobe, specifically in the MK, et al. Theta and high-frequency activity mark spontaneous
temporal and fusiform gyri. Word retrieval is also affected recall of episodic memories. J Neurosci 2014; 34: 11355–65.
Spikes outside seizure onset zone impacts memory BRAIN 2017: Page 11 of 12 | 11

Burke JF, Zaghloul KA, Jacobs J, Williams RB, Sperling MR, Sharan Kobayashi E, Bagshaw AP, Bénar CG, Aghakhani Y, Andermann F,
AD, et al. Synchronous and asynchronous theta and gamma activity Dubeau F, et al. Temporal and extratemporal BOLD responses to
during episodic memory formation. J Neurosci 2013; 33: 292–304. temporal lobe interictal spikes. Epilepsia 2006; 47: 343–54.
Burns SP, Santaniello S, Yaffe RB, Jouny CC, Crone NE. Network Kondo Y, Suzuki M, Mugikura S, Abe N, Takahashi S, Iijima T, et al.
dynamics of the brain and influence of the epileptic seizure onset Changes in brain activation associated with use of a memory strat-
zone. Proc Natl Acad Sci USA 2014; 111: E5321–30. egy: a functional MRI study. Neuroimage 2005; 24: 1154–63.
Cohen L, Lehéricy S, Chochon F, Lemer C, Rivaud S, Dehaene S. Lachaux J-P, Axmacher N, Mormann F, Halgren E, Crone NE. High-
Language-specific tuning of visual cortex? Functional properties of frequency neural activity and human cognition: past, present and
the visual word form area. Brain 2002; 125: 1054–69. possible future of intracranial EEG research. Prog Neurobiol 2012;
Drane DL, Loring DW, Voets NL, Price M, Ojemann JG, Willie JT, 98: 279–301.
et al. Better object recognition and naming outcome with MRI- Lancaster JL, Woldorff MG, Parsons LM, Liotti M, Freitas CS, Rainey
guided stereotactic laser amygdalohippocampotomy for temporal L, et al. Automated Talairach Atlas labels for functional brain map-
lobe epilepsy. Epilepsia 2015; 56: 101–13. ping. Hum Brain Mapp 2000; 10: 120–31.
Eichenbaum H, Yonelinas AP, Ranganath C. The medial temporal Marsh ED, Peltzer B, Brown MW, Wusthoff C, Storm PB, Litt B, et al.
lobe and recognition memory. Annu Rev Neurosci 2007; 30: Interictal EEG spikes identify the region of electrographic seizure
123–52. onset in some, but not all, pediatric epilepsy patients. Epilepsia
Esteller R, Echauz J, Tcheng T, Litt B, Pless B. Line length: an efficient 2010; 51: 592–601.
feature for seizure onset detection. Proceedings of the 23rd Annual McCandliss BD, Cohen L, Dehaene S. The visual word form area:
International Conference of the IEEE Engineering in Medicine and expertise for reading in the fusiform gyrus. Trends Cogn Sci 2003;
Biology Society, Istanbul, Turkey, Oct 25–28, 2001. Vol. 2, p. 7: 293–9.
1707–10. Monti MM, Parsons LM, Osherson DN. The boundaries of language
Fletcher PC, Frith CD, Grasby PM, Shallice T, Frackowiak RSJ, Dolan and thought in deductive inference. Proc Natl Acad Sci USA 2009;
RJ. Brain systems for encoding and retrieval of auditory-verbal 106: 12554–9.
memory an in vivo study in humans. Brain 1995; 118: 401–16. Murdock BB. The serial position effect of free recall. J Exp Psychol
Friedman D, Johnson R. Event-related potential (ERP) studies of 1962; 64: 482–8.
memory encoding and retrieval: a selective review. Microsc Res Ngugi AK, Bottomley C, Kleinschmidt I, Sander JW, Newton CR.
Tech 2000; 51: 6–28. Estimation of the burden of active and life-time epilepsy: a meta-
Gaspard N, Alkawadri R, Farooque P, Goncharova II, Zaveri HP. analytic approach. Epilepsia 2010; 51: 883–90.
Automatic detection of prominent interictal spikes in intracranial Nobre AC, Allison T, McCarthy G. Word recognition in the human
EEG: validation of an algorithm and relationsip to the seizure inferior temporal lobe. Nature 1994; 372: 260–3.
onset zone. Clin Neurophysiol 2014; 125: 1095–103. Paller KA, Wagner AD. Observing the transformation of experience
Gelinas JN, Khodagholy D, Thesen T, Devinsky O, Buzsáki G. into memory. Trends Cogn Sci 2002; 6: 93–102.
Interictal epileptiform discharges induce hippocampal–cortical cou- Penfield W, Jasper H. Epilepsy and the functional anatomy of the
pling in temporal lobe epilepsy. Nat Med 2016; 22: 641–8. human brain. JAMA 1954; 155: 86.
Goncharova II, Spencer SS, Duckrow RB, Hirsch LJ, Spencer DD, Posner MI, McCandliss BD. Brain circuitry during reading. In:
Zaveri HP. Intracranially recorded interictal spikes: relation to seiz- Converging methods for understanding reading and dyslexia.
ure onset area and effect of medication and time of day. Clin 1992. The MIT Press, Cambridge, MA, p. 305–37.
Neurophysiol 2013; 124: 2119–28. Raslau FD, Klein AP, Ulmer JL, Mathews V, Mark LP. Memory part
Grasby PM, Frith CD, Friston KJ, Bench C, Frackowiak RSJ, Dolan 1: overview. AJNR Am J Neuroradiol 2014; 35: 2058–60.
RJ. Functional mapping of brain areas implicated in auditory— Raslau FD, Mark IT, Klein AP, Ulmer JL, Mathews V, Mark LP.
verbal memory function. Brain 1993; 116: 1–20. Memory part 2: the role of the medial temporal lobe. AJNR Am J
Hamame CM, Vidal JR, Ossandon T, Jerbi K, Dalal SS, Minotti L, Neuroradiol 2015; 36: 846–9.
et al. Reading the mind’s eye: online detection of visuo-spatial work- Rausch R, Lieb JP, Crandall PH. Neuropsychologic correlates of
ing memory and visual imagery in the inferior temporal lobe. depth spike activity in epileptic patients. Arch Neurol 1978; 35:
Neuroimage 2012; 59: 872–9. 699–705.
Holm S. A simple sequentially rejective multiple test procedure. Scand Schwab RS. A method of measuring consciousness in attacks of petit
J Stat 1979; 6: 65–70. mall epilepsy. Arch Neurol Psychiatry 1939; 41: 215–17.
Holmes GL, Lenck-Santini PP. Role of interictal epileptiform abnorm- Sederberg PB, Kahana MJ, Howard MW, Donner EJ, Madsen JR.
alities in cognitive impairment. Epilepsy Behav 2006; 8: 504–15. Theta and gamma oscillations during encoding predict subsequent
Hoppe C, Elger CE, Helmstaedter C. Long-term memory impair- recall. J Neurosci 2003; 23: 10809–14.
ment in patients with focal epilepsy. Epilepsia 2007; 48 (Suppl 9): Sederberg PB, Schulze-Bonhage A, Madsen JR, Bromfield EB,
26–9. McCarthy DC, Brandt A, et al. Hippocampal and neocortical
Horak P, Meisenhelter S, Song Y, Testorf M, Kahana M, Viles W, gamma oscillations predict memory formation in humans. Cereb
et al. Interictal epileptiform discharges impair word recall in multiple Cortex 2007; 17: 1190–6.
brain areas. Epilepsia 2016; 58; 373–80. Spencer SS, Goncharova II, Duckrow RB, Novotny EJ, Zaveri HP.
Jacobs J, Kahana MJ. Direct brain recordings fuel advances in cogni- Interictal spikes on intracranial recording: behavior, physiology,
tive electrophysiology. Trends Cogn Sci 2010; 14: 162–71. and implications. Epilepsia 2008; 49: 1881–92.
Janca R, Jezdik P, Cmejla R, Tomasek M, Worrell GA, Stead M, et al. Squire LR, Stark CEL, Clark RE. The medial temporal lobe. Annu Rev
Detection of interictal epileptiform discharges using signal envelope Neurosci 2004; 27: 279–306.
distribution modelling: application to epileptic and non-epileptic Squire LR, Zola-Morgan S. The medial temporal lobe memory system.
intracranial recordings. Brain Topogr 2014; 28: 172–83. Science 1991; 253: 1380–6.
Kleen JK, Scott RC, Holmes GL, Lenck-Santini PP. Hippocampal Sutula TP, Hagen J, Pitkänen A. Do epileptic seizures damage the
interictal spikes disrupt cognition in rats. Ann Neurol 2010; 67: brain? Curr Opin Neurol 2003; 16: 189–95.
250–7. Wagner AD, Schacter DL, Rotte M, Koutstaal W, Maril A, Dale AM,
Kleen JK, Scott RC, Holmes GL, Roberts DW, Rundle MM, Testorf et al. Building memories: remembering and forgetting of verbal ex-
M, et al. Hippocampal interictal epileptiform activity disrupts cog- periences as predicted by brain activity. Science 1998; 281:
nition in humans. Neurology 2013; 81: 18–24. 1188–91.
12 | BRAIN 2017: Page 12 of 12 H. Ung et al.

Wang W-H, Liou H-H, Chen C-C, Chiu M-J, Chen T-F, Cheng T-W, Wilson SB, Emerson R. Spike detection: a review and comparison of
et al. Neuropsychological performance and seizure-related risk fac- algorithms. Clin Neurophysiol 2002; 113: 1873–81.
tors in patients with temporal lobe epilepsy: a retrospective cross- Zahn R, Huber W, Drews E, Erberich S, Krings T, Willmes K, et al.
sectional study. Epilepsy Behav 2011; 22: 728–34. Hemispheric lateralization at different levels of human auditory
Warren CP, Hu S, Stead M, Brinkmann BH, Bower MR, Worrell GA. word processing: a functional magnetic resonance imaging study.
Synchrony in normal and focal epileptic brain: the seizure onset Neurosci Lett 2000; 287: 195–8.
zone is functionally disconnected. J Neurophysiol 2010; 104:
3530–9.