Clinical Neurophysiology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph

Is ictal cognitive dysfunction predictable? Ivan Osorio ⇑ Department of Neurology, University of Kansas Medical Center, 3901 Rainbow Blvd, Kansas City, KS 66160, United States

a r t i c l e

i n f o

Article history: Accepted 11 February 2014 Available online xxxx Keywords: Seizures Clinical Sub-clinical Cognitive Inter-ictal Pre-ictal Ictal Prediction Warning Quality of life

h i g h l i g h t s  Accurate prediction of electrographic seizure onset is unlikely to improve quality of life, unless,

patients know if the impending seizure would be clinical or sub-clinical.  The results of automated administration of a reaction time indicate cognitive dysfunction is abrupt,

suggesting there is no discernible (for forecasting purposes) cognitive pre-ictal state.  If reproduced in other studies, these findings would mean that ictal cognitive dysfunction is

unpredictable.

a b s t r a c t Objective: Accurate prediction of electrographic seizure onset may reduce injuries and improve quality of life in pharmaco-resistant epileptics. However, because sub-clinical, far out-number clinical seizures, indiscriminate issuance of warnings may have a paralyzing effect on these patients. This study investigates the predictability of ictal cognitive dysfunction. Methods: Latency and percentage of correct responses to a reaction time test triggered by automated seizure detections were compared to those obtained inter-ictally in 14 subjects undergoing surgery evaluation. Since accurate prediction of seizures is elusive, early detection was used, as it indirectly but reliably investigates for the existence of a cognitive pre-ictal state. Results: Significant differences between ictal and inter-ictal cognitive performance were not uncovered until late into the temporal evolution of ‘‘focal’’ seizures. Conclusions: These observations suggest that cognitive dysfunction is unpredictable in seizures originating from discrete cortical regions, as the transition into unawareness seems abrupt. Significance: Prediction of electrographic seizure onsets with worthwhile accuracy would likely result in large numbers of daily warnings, the great majority for sub-clinical seizures. This outcome would considerably increase, without safety justification, patients’ psychological burden inherent to each forecast, thus further diminishing quality of life. Ó 2014 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

1. Introduction Prediction of the electrographic onset of seizures remains one of the chief priorities of modern epileptology, due to its potential for preventing injuries, and allowing subjects with uncontrollable seizures to re-enter mainstream society. To fully reap these benefits, not only should prediction accuracy be high (e.g., >80%), but warnings should be limited to impending seizures with clinical manifestations, foremost among them, unawareness and/or loss ⇑ Tel.: +1 913 5884529; fax: +1 913 5884585.

of consciousness. Detailed analyses of prolonged ECoGs of patients undergoing evaluation for epilepsy surgery support the common place observations of epileptologists that sub-clinical seizures (defined as ictal electrographic activity of which patients and observers are unaware), far outnumber clinical seizures (Osorio et al., 2009, 2010; Feldwisch-Drentrup et al., 2011a). While these observations are made under conditions intended to facilitate the emergence of seizures (dose reduction of anti-seizure drugs), the expected decrease in seizure frequency brought about by antiseizure therapies may not sufficiently reduce them to minimize the negative psychological impact caused by frequent warnings, the majority of which would not be accompanied by clinical

E-mail address: [email protected] http://dx.doi.org/10.1016/j.clinph.2014.02.022 1388-2457/Ó 2014 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

Please cite this article in press as: Osorio I. Is ictal cognitive dysfunction predictable?. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/ j.clinph.2014.02.022

2

I. Osorio / Clinical Neurophysiology xxx (2014) xxx–xxx

manifestations. Forecasting that an impending seizure would be devoid of clinical manifestations may obviate the need for warnings and their attendant psychological toll (Arthurs et al., 2010). To date, there are no published studies of attempts to predictively classify seizures as clinical or sub-clinical using cognitive tests. The feasibility of predictively classifying seizures as clinical or subclinical may be assessed by studying the cognitive/behavioral transition from the inter-ictal to the ictal state. Drawing an analogy between cognitive functions and what is known about interictal-to-ictal transitions of electrical cortical signals, if impairment of cognition is gradual as purported to be the case for the electrographic transition into partial seizures (Lopes da Silva et al., 2003a,b; Suffczynski et al., 2004, 2006), predictive classification of the impending seizure will be challenging, but manageable. However, if the transition from awareness into unawareness is sudden, as may be the case with the onset of electrographic ictal activity in spike-wave epilepsies (Lopes da Silva et al., 2003a,b; Suffczynski et al., 2004, 2006), cognitive status will likely be unpredictable. An investigation of the time course of cognitive changes (e.g., from awareness to unawareness) was undertaken in order to assess the feasibility of automated real-time warnings of the onset of clinical seizures in subjects with pharmaco-resistant localization-related epilepsies (Osorio and Frei, 2010). However, given the lack of algorithms that reliably predict electrographic seizure onset (Stacey et al., 2011) a validated seizure detection algorithm (Osorio et al., 1998, 2002) was used instead. To this end, performance on a complex reaction time test (Zajdel and Nowak, 2007) that involves: (a) stimuli registration; (b) choice between two simultaneously presented stimuli; and (c) construction of a decision and enactment of a response was evaluated during the period beginning with the automated detection of electrographic onset. The latency of responses to each test presentation and the percent of correct responses were compared between the inter-ictal and

ictal states in subjects enrolled in this study (Osorio and Frei, 2010) and the results were utilized to indirectly determine if cognitive impairment is predictable. 2. Methods This study was conducted at the University of Kansas Medical Center with approval from the Human Subjects Committee, in subjects with pharmaco-resistant localization-related epilepsies undergoing invasive surgical evaluation in accordance with this institution’s protocol that includes discontinuation of anti-seizure drugs. Subjects which served as their own control were enrolled in the order of admission to the monitoring unit. Inclusion criteria were: (a) good candidate for invasive epilepsy surgery evaluation; (b) normal motor function; and (c) normal vision with or without correction. Exclusion criteria were (a) mental retardation; and (b) occurrence of any of the following during the evaluation prior to collection of an adequate sample of test presentations and responses: (1) status epilepticus; (2) use of rescue, psychoactive, or CNS-depressant drugs; (3) medical or neurological complications. To evaluate cognitive status a complex reaction time test (CRT) (Zajdel and Nowak, 2007) was administered to each subject under two conditions: (a) during seizures; and (b) inter-ictally (i.e., inbetween seizures). Testing began no earlier than 24 h. after electrode implantation to allow for recovery from general anesthesia and lessening of immediate postoperative pain/discomfort. 2.1. Test description Each subject had to successfully complete a training session prior to the start of the study. The CRT consisted of the serial presentation of 2 visual ‘‘stimuli’’ (the letter A and a square h), displayed simultaneously on a 1500 computer monitor (Fig. 1). The

Fig. 1. Experimental setup. The EEG acquisition system receives cortical signals from intracranially implanted electrodes, and outputs these data to a computer running the seizure detection algorithm. Seizure detection by the algorithm triggers administration of the neuropsychological test. The patient responds to the test by pressing either the right or the left buttons of mouse attached to the bed rail. The test is also administered inter-ictally at random times.

Please cite this article in press as: Osorio I. Is ictal cognitive dysfunction predictable?. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/ j.clinph.2014.02.022

I. Osorio / Clinical Neurophysiology xxx (2014) xxx–xxx

position of the letter on either half of the screen (i.e., left ‘‘A h ’’, or right ‘‘ h A’’) (Fig. 1), was randomly chosen for each presentation and the subject was instructed to immediately press, upon appearance of the stimuli, the mouse button (right or left) ipsilateral to the side of the screen (right or left) on which the letter A was displayed. As soon as the subject pressed either button, or after a maximal presentation time (1 s.) had elapsed without a response, each stimulus presentation was removed until the next programmed stimuli. At the end of each presentation, a random timer was set, the expiration of which would trigger the next presentation. A total of 36 stimuli were presented in each testing session. Inter-trial presentation time intervals were randomly chosen from the set {0.5, 1.0, 1.5, 2.0 s} to minimize adaptation. 2.2. Timing of complex reaction time (CRT) Tests Administration The CRTs were administered between 08:00 and 20:00 throughout the surgical evaluation and were triggered by: 1. the earlier of either, automated seizure detection (Osorio et al., 1998, 2002) or of event button presses by the subject or an observer, or 2. Interictally, at random times; the timer that triggered inter-ictal CRTs was set for 6 presentations/day, uniformly distributed throughout the test period, with the additional constraints that no inter-ictal test would be administered within the 15 min period after a seizure or a randomly triggered test. To minimize fatigue, no more than 30 tests (inter-ictal plus seizure-triggered) could be administered over any 12 h period. Whenever a CRT was triggered, a sound file consisting of a voice saying ‘‘Begin Test’’ was automatically played to summon the subject to take the test. The subject was instructed to, upon hearing the summon, press either of the mouse buttons to activate the test. The summon was replayed every 5 s with increasing loudness with each repetition until the subject either activated the test or after it was delivered 6 times (30 s). If the subject did not activate the test, this information was logged and the system went dormant until the next test session. Seizure intensity and duration were quantified off-line with the same algorithm used for their detection; most automated detections occurred within 5 s after electrographic onset as marked visually by independent experts. The classification of CRTs as ictal or inter-ictal was validated off-line via expert visual analysis of the ECoG associated with each test. CRTs triggered by false positive detections or correlated in time with false negative detections were reclassified accordingly. 2.3. Complex reaction time data recording and processing The following were recorded (with millisecond precision) and saved to the computer’s memory: (a) Test condition (seizure vs. inter-ictal); (b) The time of each test summon; (c) Latency of responses to the summons; (d) Stimuli presentation times and side of the screen (left vs. right) where the letter ‘‘A’’ appeared; and (e) Times and sides (left vs. right) of all mouse button presses. These data were processed to derive the following measures: I. Compliance, defined as the fraction of presented stimuli within each testing session for which the subject pressed the button regardless of correctness (Compliance Score = Number of responses/Number of stimuli presentations); II. Percentage of Correct Responses defined as those for which the subject pressed only once the button ipsilateral to the side (left or right) of the screen where the letter ‘‘A’’ appeared, (% Correct Responses = Number of Correct Responses  100/Number of Stimuli presentations). Responses were classified as Incorrect if: (a) The mouse button contralateral to side of the screen where the letter ‘‘A’’ appeared was pressed; (b) Both mouse buttons were pressed simultaneously or sequentially; (c) The correct button was pressed more than once

3

per stimulus presentation; or (d) No button was pressed; III. Time to Impaired Responsiveness (TIR): The time (in sec.) elapsed between each test summon and the last correct response prior to the first test failure as defined below. The time of the last correct response is used for this purpose, instead of the time of subsequent failure, as it represents the last moment at which the subject was behaving without measurable impairment. While it is possible that subjects remain unimpaired for an even longer period of time thereafter, this time point was used to conservatively underestimate TIR, which is the preferred approach for safety-related considerations. The mean, range and SD of TIR were computed for 3 different definitions of test failure, listed in order from most to least stringent (Fig. 2): (1). A correct response but with latency exceeding the 90th percentile of those for inter-ictal tests (TIR-A); speed of reaction is in certain situations as important as correctness of response; (2). Any incorrect response as defined above in II (TIR-B); (3). Three consecutive incorrect responses, regardless of their latencies, or lack of responses (TIR-C), a definition that attempts to account for the fact that subjects intermittently err even during inter-ictal tests. If no failure occurred in a test, the interval ends with the time of the correct response to the last stimuli. Subjects’ data were included in the final analyses only if: (1). The ECoG tracings were of sufficiently good quality to allow visual ascertainment of the presence (for seizure-triggered) or absence (for inter-ictal tests) of ictal activity; and (2). The subject took at least 2 inter-ictal and 1 seizure-triggered test. 2.4. Seizure classification Seizures were classified as either clinical (e.g., complex partial) or sub-clinical. Sub-clinical seizures were operationally defined as paroxysmal electrographic activity not associated with event button presses or behavioral changes; bursts of epileptiform discharges were not covered this definition. 2.5. ECoG recording ECoG was recorded using commercially available depth or grid electrodes electrodes (Ad-Tech, Racine, WI). These signals were fed into commercially available systems (Nicolet, Madison, WI), filtered (0.5–70 Hz), digitized (240 Hz, 10 bits of precision, 0.59 lV/ bit) and analyzed using a validated algorithm (Osorio et al., 1998, 2002) implemented into a custom bedside system (Peters et al., 2001). 2.6. Statistical data analysis The data were analyzed for each as well as for all (pooled) subjects, where appropriate, using the paired-t- and the Kolmogorov–Smirnoff test. 3. Results Twenty subjects who met the inclusion criteria were enrolled in this study but data from 6 were excluded from analyses as they did not take the minimum required number of seizure-triggered tests. Demographic/clinical data, type, numbers of electrodes, sites implanted, and localization of the epileptogenic zone(s)) are shown in Table 1. In 12/14 subjects electrographic preceded clinical seizure onset; in subjects 16 & 20 clinical preceded electrographic onset, but their data was included in the analyses as it could uncover differences between early and late ictal cognitive functioning. Sites of seizure origin were: mesial temporal in 12/14 and frontal in 2/14 subjects. The total number of seizures/subject, seizure class (clinical or

Please cite this article in press as: Osorio I. Is ictal cognitive dysfunction predictable?. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/ j.clinph.2014.02.022

4

I. Osorio / Clinical Neurophysiology xxx (2014) xxx–xxx

Fig. 2. Top panel (a): ECoG recorded bilaterally from: (i) amygdalae and hippocampi in a rostro-caudal orientation; (ii) orbito-frontal cortices (y-axis: Electrode labels and contact numbers. L = Left; R = Right; A = Anterior; M = Middle; P = Posterior; T = Temporal; F = Frontal); Vertical dash line at 11 s.: Time of automated detection; Middle panel (b): Ictal intensity curve (y-axis) as function of time (x-axis). R(t) is the change in the ictal content of ECoG in a 2 s. window. Automated detections are issued when R(t) P 22 for at least 0.84 s.; Bottom Panel (c): Graph of CRT test performance; Vertical dash line: Subject summoned to take the test; Vertical dot lines: Times at which each stimulus was presented; y-axis: Response latency (in sec.); x-axis: time (in sec.). Symbols: +: Correct response; O: No response; TIR-A: correct response but with latency exceeding the 90th percentile of inter-ictal tests; TIR-B: incorrect response or no response; 3. TIR-C: Three consecutive incorrect responses, regardless of latency or no responses.

sub-clinical) and number of seizures by class are provided in Table 2. Simple partial seizures (‘‘auras’’) manifested with epigastric discomfort in some patients, a sensation of ‘‘hotness’’ in others and in a few, the symptoms were indescribable. Complex partial seizures were characterized in the majority of patients by motionless, followed by reactive or de-novo automatisms depending on their duration (the longer the seizure the more likely for automatisms to ensue).

A total of 856 CRT tests were administered: 649 (76%) were inter-ictal with subjects responding to 520 (80%) and 207 (19%) were seizure-triggered (all true positive detections) with subjects responding to 73 (35%). The mean and SD of the average compliance scores for the 14 subjects analyzed were: Inter-ictal tests: 0.91 ± 0.12 vs. Seizure tests: 0.82 ± 0.27, differences that were not statistically significant (paired t-test: p  0.14). The mean and SD of the percentage of correct responses for the 14 subjects

Please cite this article in press as: Osorio I. Is ictal cognitive dysfunction predictable?. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/ j.clinph.2014.02.022

5

I. Osorio / Clinical Neurophysiology xxx (2014) xxx–xxx

Table 1 Relevant demographic, clinical and recording electrode information. CP = Complex partial; Sec. Gral. = Secondarily generalized; N/A = Not available; Rt. = Right; Lt. = Left; B.D. = Bilateral Depth; Lat. = Lateral. Subject number

Age

Years w/epilepsy

Gender

Etiology

Class.

Frequency (Sz/mo.)

Epileptogenic zone(s)

Electrode type(s)

1 2 3 5 8 10 11 14 15 16 17 18 19 20

32 30 34 11 23 26 48 34 32 22 19 21 25 30

27 29 22 4 23 26 38 22 19 21 11 17 16 11

F F M M F M M F M F M M F M

Cryptogenic Cryptogenic Infection Cryptogenic Cryptogenic Infection Cryptogenic Cryptogenic Trauma Cryptogenic Cryptogenic Trauma Cryptogenic Cryptogenic

CP CP Sec. CP CP CP CP CP Sec. Sec. CP CP Sec. CP

10 4 5 16 N/A 15 75 3 3 2 3 4 N/A 16

Rt. Amygdala Lt. Amygdala & Hippocampus Rt. Fronto-Temporal Neocortex Rt. Amygdala Lt. Hippocamp. Left Amygdala & Hippocampus Rt Amygdala Lt. Hippocamp; Rt. Amygdala Rt. Frontal Neocortex Rt. & Lt. Hippocampus; bi-Fronto-Polar Lt. Hippocamp.; Rt. Amygdala Lt. Hippocamp. Lt. Hippocamp. Rt. Amygdala & Hippocampus

B.D. B.D. Grid B.D. B.D. B.D. B.D. B.D. Grid B.D., Strips B.D.; Lat. B.D. B.D.; Lat. B.D.

Gral.

Gral. Gral.

Gral.

Table 2 Total number and class of seizures. SP = Simple partial seizure; CP = Complex partial; Sec. Gralized = Secondarily generalized seizure; sC = Sub-clinical. Subject

Total number of seizures

1 2 3 5 8 10 11 14 15 16 17 18 19 20

1 (SP) 19 (7 CP; 2 SPS; 10 sC) 2 (SP) 2 (1 SP; 1SP) 121 (8 CP; 11 SP; 102 sC) 7 (5 CP; 2 sC) 2 (CP) 10 (5 CP; 2 SPS; 3 sC) 10 (6 CP; 4sC) 1 (Sec. Gralized) 6 (4 CPS; 2 sC) 5 (4 CP; 1 sC) 2 (1 CP; 1sC) 1 (CP)

Table 3 TIR is the Time to Impairment of Responsiveness, with failure defined in decreasing order of stringency as: A. A correct response with latency exceeding the 90th percentile of inter-ictal test response latencies; B. Any incorrect response; or C. Three consecutive incorrect responses. Inter-ictal tests

TIR-A TIR-B TIR-C

Seizure tests

Mean (s)

SD (s)

Range (s)

Mean (s)

SD (s)

Range (s)

22.0 37.2 55.5

17.3 25.0 24.8

0.2–83.5 0.2–92.0 2.0–103.0

27.1 42.8 56.1

19.8 24.2 23.9

0.6–75.0 0.6–82.2 0.6–88.7

were: Inter-ictal tests: 85 ± 14% vs. Seizure tests: 76 ± 30%, differences that were not statistically significant (paired t-test: p  0.15). The Kolmogorov–Smirnoff test did not reveal significant differences between the seizure and inter-ictal distributions for either compliance or percentage of correct responses. The mean time (measured from when patients were summoned to take the test) to cognitive impairment (Table 3) using the most stringent criteria (a correct response but with latency exceeding the 90th percentile of those for random tests) was 27.1 s (SD: 19.8 s) and with the least stringent (three consecutive incorrect responses, regardless of latency) was 56.1 s (SD: 23.9 s). Summon time always followed automated detection of electrographic onset which in turn followed electrographic onset as visually determined by an epileptologist. The paired-t test showed differences in means between seizure vs. inter-ictal tests that were significant for TIR-A (p  0.02) and TIR-B (p  0.04), but not for TIR-C (p  0.4), but the Kolmogorov–Smirnoff test did not uncover significant differences among their distributions. The occurrence of a delayed (but correct) response, of an incorrect response or of no response was

abrupt/sudden (Fig. 2) in all abnormal tests, in all subjects, independent of site of seizure origin, intensity or duration; the time between a correct and incorrect response ranged between 0.5–2.0 s, reflecting the inter-stimulus delay presentation. There was no impairment in reaction time tests during simple partial seizures (‘‘auras’’); the transition from simple (awareness) to complex partial (unawareness) was sudden. No differences in complex reaction time test performance based on side of seizure origin and hemispheric dominance for language were uncovered in this study. 4. Discussion The lack of significant differences (indicative of cognitive impairment) in latency and percent correct responses between inter-ictal and ictal tests at electrographic onset, and for some time thereafter, insinuates that cognitive dysfunction and by extension, seizure class (clinical vs. sub-clinical) may be unpredictable, due to the abruptness with which it occurs, which in turn suggests the lack of a pre-ictal cognitive state. While the paired-t test revealed significant differences in means between seizure vs. inter-ictal tests for TIR-A (p  0.02) and TIR-B (p  0.04), these were due to paradoxically longer latencies to impairment of responsiveness for seizures than for inter-ictal tests. Explicitly, unless and until cognitive dysfunction ensued, performance was better during the ictal than the inter-ictal state (Osorio and Frei, 2010; Osorio, 2011; Hermann, 2011). Moreover, when: (a) lengthening of response latencies beyond the 90th percentile of those recorded inter-ictally; (b) issuance of an incorrect response; or (c) absence of responses occurred, they were not preceded by specific harbingers of dysfunction. Fig. 2 illustrates the abruptness of cognitive impairment during a partial seizure: The delayed (but correct) response at approximately 20 s (x-axis) is flanked by normal latency responses; the absence of a response 5 s later is again preceded and followed by timely and correct responses and finally the failure to respond to all stimuli is sudden. While there is the appearance of a ‘‘trend’’ of increasing latencies prior to the failure to respond to all stimuli, the responses were correct, had normal latencies, and would have been preceded by two false positive warnings. It should be pointed out that while the dose of anti-seizure medications was reduced in all patients, this action is unlikely to confound interpretation of the results. The seemingly sudden transition between awareness and unawareness may reflect, in part, the low sensitivity or lack of discriminating power of the test administered to these subjects. It is thus plausible that if other cognitive tests had been administered, some form or degree of cognitive dysfunction may have been uncovered pre-ictally. However, for cognitive tests to

Please cite this article in press as: Osorio I. Is ictal cognitive dysfunction predictable?. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/ j.clinph.2014.02.022

6

I. Osorio / Clinical Neurophysiology xxx (2014) xxx–xxx

have extensive clinical applicability predicting whether an impending seizure will be associated with cognitive impairment, they must be (a) short in duration and implementable in real-time; (b) administrable under any conditions; (c) repeatable ‘‘ad infinitum;’’ and (d) have negligible practice effect. Should a cognitive pre-ictal state be discernible and detectable (in real-time) in localization-related epilepsies, cortical electrical signals alone may be relied upon to predict cognitive impairment. But, this may require extensive spatial sampling and sophisticated implantable/portable signal processing capabilities to record and analyze the enormous data stream, caveats that may make this approach impractical. Should these results be reproduced in larger studies, their implication would be that there is no ‘‘pre-ictal’’ cognitive state, ostensibly in contrast to the electrographic domain where claims of its existence have been put forth (Stacey et al., 2011; Feldwisch-Drentrup et al., 2011b). Lack of a pre-ictal cognitive state that by definition, must differ in at least one property or characteristic from the inter-ictal, ictal and post-ictal states, is likely to translate into unpredictability (Sornette and Osorio, 2011) of ictally-associated cognitive changes. Said unpredictability would degrade quality of life benefits for those reliant upon prediction of seizures to lead normal lives, since sub-clinical outnumber clinical seizures (Osorio et al., 2009, 2010; Feldwisch-Drentrup et al., 2011a), making the great majority of issued warnings superfluous or irrelevant from a safety perspective. Moreover, that sub-clinical seizures may be more reliably predictable than clinical seizures (Feldwisch-Drentrup et al., 2011a) further diminish the potential benefits of prediction of electrographic onset of partial seizures as the ratio of superfluous to relevant warnings will be even higher. The prediction of seizure onset using neuronal electrical signals has been deemed, ‘‘a priori’’, of great clinical value to pharmacoresistant epileptics, but the ostensive unpredictability of cognitive impairment cast doubts about the practicality and value of this research endeavor. Acknowledgements The author has no conflicts of interest to report. M.G. Frei, Ph.D., processed the data, performed statistical analyses and generated the figures.

References Arthurs S, Zaveri H, Frei MG, Osorio I. Patient and caregiver perspectives on seizure prediction. Epilepsy Behav 2010;19:474–7. Feldwisch-Drentrup H, Ihle M, Quyen Mle V, Teixeira C, Dourado A, Timmer J, et al. Anticipating the unobserved: prediction of sub-clinical seizures. Epilepsy Behav 2011a;22(Suppl. 1):S119–26. Feldwisch-Drentrup H, Staniek M, Schulze-Bonhage A, Timmer J, Dickten H, Elger CE, et al. Identification of pre-seizure states in epilepsy: a data-driven approach for multichannel EEG recordings. Front Comput Neurosci 2011b;5:32–9. Hermann B. Cognition in epilepsy and its transient impairment. Epilepsy Behav 2011;22:419. Lopes da Silva F, Blanes W, Kalitzin SN, Parra J, Suffczynski P, Velis DN. Epilepsies as dynamical diseases of brain systems: basic models of the transition between normal and epileptic activity. Epilepsia 2003a;44(Suppl. 12):72–83. Lopes da Silva FH, Blanes W, Kalitzin SN, Parra J, Suffczynski P, Velis DN. Dynamical diseases of brain systems: different routes to epileptic seizures. IEEE Trans Biomed Eng 2003b;50:540–8. Osorio I, Frei M. Feasibility of automated warning in subjects with localizationrelated epilepsies. Epilepsy Behav 2010;19:602–7. Osorio I, Frei MG, Wilkinson SB. Real-time automated detection and quantitative analysis of seizures and short-term prediction of clinical onset. Epilepsia 1998;39:615–27. Osorio I, Frei MG, Giftakis J, Peters T, Ingram J, Turnbull M, et al. Performance reassessment of a real-time seizure-detection algorithm on long ECoG series. Epilepsia 2002;43:1522–35. Osorio I, Frei MG, Sornette D, Milton J. Pharmaco-resistant seizures: self-triggering capacity, scale-free properties and predictability? Eur J Neurosci 2009;30:1554–8. Osorio I, Frei MG, Sornette D, Milton J, Lai YC. Seizures: quakes of the brain? Phys Rev E 2010;82:1–13. Osorio I. Mesial temporal seizures may enhance complex reaction time test performance. Epilepsy Behav 2011;22:490–4. Peters TE, Bhavaraju NC, Frei MG, Osorio I. Network system for automated seizure detection and contingent delivery of therapy. J Clin Neurophysiol 2001;18:545–9. Sornette D, Osorio I. Prediction. In: Osorio I, Zaveri HP, Frei MG, Arthurs S, editors. Epilepsy: The Intersection of Neurosciences, Biology, Mathematics, Engineering and Physics. Boca Raton: CRC Press; 2011. p. 203–37. Stacey W, Le Van Quyen M, Mormann F, Schulze-Bonhage A. What is the presentday EEG evidence for a preictal state? Epilepsy Res 2011;97:243–51. Suffczynski P, Kalitzin S, Lopes Da Silva FH. Dynamics of non-convulsive epileptic phenomena modeled by a bistable neuronal network. Neuroscience 2004;126:467–8. Suffczynski P, Lopes da Silva FH, Parra J, Velis DN, Bouwman BM, van Rijn CM, et al. Dynamics of epileptic phenomena determined from statistics of ictal transitions. IEEE Trans Biomed Eng 2006;53:524–32. Zajdel R, Nowak D. Simple and complex reaction time measurement. Comput Biol Med 2007;37:1724–30.

Please cite this article in press as: Osorio I. Is ictal cognitive dysfunction predictable?. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/ j.clinph.2014.02.022

Is ictal cognitive dysfunction predictable?

Accurate prediction of electrographic seizure onset may reduce injuries and improve quality of life in pharmaco-resistant epileptics. However, because...
1MB Sizes 0 Downloads 3 Views

Recommend Documents


Preexisting cognitive impairment is associated with postoperative cognitive dysfunction after hip joint replacement surgery.
This study investigated the prevalence of cognitive impairment in elderly noncardiac surgery patients and any association between preoperative cognitive impairment and postoperative cognitive dysfunction (POCD). Additionally, the incidence of cogniti

Post-Treatment Edema after Meningioma Radiosurgery is a Predictable Complication.
Symptomatic post-treatment edema (PTE) causing seizures, focal deficits, and intracranial hypertension is a rather common complication of meningioma radiosurgery. Factors associated to the occurrence of PTE still needs to be clarified. We retrospecti

How predictable is the erectile function of patients with epilepsy?
Erectile dysfunction (ED) is often reported by patients with epilepsy and may be related to endocrine system abnormalities, side effects of antiepileptic drugs, psychiatric comorbidities, and family or social difficulties.

Outcome of surgery for benign prostatic hyperplasia-is it predictable?
The study was done to evaluate the pre-operative and intra operative factors which influence the post-operative outcome in patients undergoing surgery for Benign Prostatic Hyperplasia (BPH).

Epileptic monocular nystagmus and ictal diplopia as cortical and subcortical dysfunction.
We present the case of a patient with ictal monocular nystagmus and ictal diplopia who became seizure-free after resection of a right frontal focal cortical dysplasia (FCD), type 2B. Interictal neuroophthalmological examination showed several beats o

Ictal asystole and ictal syncope: insights into clinical management.
Ictal asystole is a rare, serious, and often treatable cause of syncope. There are currently limited data to guide management. Characterization of ictal syncope predictors may aid in the selection of high-risk patients for treatments such as pacemake

Heart failure and cognitive dysfunction.
It has been estimated that 5.1million Americans suffer from heart failure. Cognitive impairment has been described as a consequence of heart failure in numerous studies spanning the last three decades. This systematic review helps differentiate "cogn

Diabetes, dysglycemia and cognitive dysfunction.
Evidence from the last decade supports the hypothesis that diabetes may be viewed as a disease of accelerated cognitive ageing. It is a risk factor for the development of dementia, for an accelerated rate of cognitive decline and for cognitive dysfun

Interleukin 6 and cognitive dysfunction.
The interleukin-6 (IL-6) is a pleiotropic cytokine that plays a key role in interaction between immune and nervous system. Although IL-6 has neurotrophic properties and beneficial effects in the CNS, its overexpression is generally detrimental, addin