- Email: [email protected]
Cessation of gamma activity in the dorsomedial nucleus associated with loss of consciousness during focal seizures
Epilepsy & Behavior 51 (2015) 215–220
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Cessation of gamma activity in the dorsomedial nucleus associated with loss of consciousness during focal seizures B.A. Leeman-Markowski a,⁎, O.L. Smart b, R.E. Faught a, R.E. Gross b, K.J. Meador c a b c
Department of Neurology, Emory University, 101 Woodruff Circle, Suite 6000, Atlanta, GA 30322, USA Department of Neurosurgery, Emory University, 101 Woodruff Circle, Suite 6000, Atlanta, GA 30322, USA Department of Neurology and Neurological Sciences, Stanford University, 300 Pasteur Drive (Room A343), Stanford, CA 94305-5235, USA
a r t i c l e
i n f o
Article history: Received 22 April 2015 Revised 20 June 2015 Accepted 20 July 2015 Available online 24 August 2015 Keywords: EEG Intracranial electrodes Complex partial seizures Thalamus Dorsomedial nucleus Consciousness Postictal state
a b s t r a c t Rationale: Impaired consciousness during seizures may be mediated by ictal propagation to the thalamus. Functions of individual thalamic nuclei with respect to consciousness, however, are largely unknown. The dorsomedial (DM) nucleus of the thalamus likely plays a role in arousal and cognition. We propose that alterations of firing patterns within the DM nucleus contribute to impaired arousal during focal seizures. Methods: Electroencephalograph data were collected from electrodes within the left DM thalamus and midcingulate cortex (MCC) in a patient undergoing seizure monitoring. Spectral power was computed across ictal states (preictal, ictal, and postictal) and level of consciousness (stupor/sleep vs. awake) in the DM nucleus and MCC. Results: Eighty-seven seizures of multifocal left frontal and temporal onsets were analyzed, characterized by loss of consciousness. At baseline, the left DM nucleus demonstrated rhythmic bursts of gamma activity, most frequently and with greatest amplitude during wakefulness. This activity ceased as ictal discharges spread to the MCC, and consciousness was impaired, and it recurred at the end of each seizure as awareness was regained. The analysis of gamma (30–40 Hz) power demonstrated that when seizures occurred during wakefulness, there was lower DM ictal power (p b 0.0001) and higher DM postictal power (p b 0.0001) relative to the preictal epoch. This spectral pattern was not evident within the MCC or when seizures occurred during sleep. Conclusions: Data revealed a characteristic pattern of DM gamma bursts during wakefulness, which disappeared during partial seizures associated with impaired consciousness. The findings are consistent with studies suggesting that the DM nucleus participates in cognition and arousal. © 2015 Elsevier Inc. All rights reserved.
1. Introduction
2. Methods
Focal seizures may impair consciousness, likely mediated, in part, by the spread of ictal activity to the thalamus [1,2]. The precise thalamic regions involved and their roles with respect to the maintenance of consciousness, however, are unclear. This report presents unique intracranial EEG data from a patient with localization-related epilepsy, using depth electrode contacts within the left dorsomedial (DM) nucleus of the thalamus and left midcingulate cortex (MCC). The DM nucleus demonstrated rhythmic bursts of gamma activity during wakefulness, which were abolished when ictal activity spread to the MCC and consciousness was impaired. We propose that taking the DM nucleus “offline” during ictal activity contributes to alteration of consciousness.
2.1. Subject
⁎ Corresponding author at: 2407 Avalon Pines Dr., Coram, NY 11727, USA. Tel.: +1 248 444 1302. E-mail addresses: [email protected] (B.A. Leeman-Markowski), [email protected] (O.L. Smart), [email protected] (R.E. Faught), [email protected] (R.E. Gross), [email protected] (K.J. Meador).
http://dx.doi.org/10.1016/j.yebeh.2015.07.027 1525-5050/© 2015 Elsevier Inc. All rights reserved.
Data were obtained from a single subject undergoing intracranial EEG recording for clinical indications.
2.2. EEG acquisition Bilateral 8-contact thalamic depth electrodes were implanted using stereotactic coordinates. Left-hemisphere Contact #2 was placed in the DM nucleus, while the corresponding right-hemisphere contact was placed more laterally, in the internal medullary lamina or nucleus ventrooralis internus. The most superficial contacts (#7–8) were located in the MCC bilaterally (Fig. 1D). Thalamic targets were selected to document possible subcortical spread of ictal activity, perhaps leading to an indication for future deep brain stimulation. The implantation also included a 10-contact left frontal–temporal strip and a 12-contact left frontal–hippocampal depth electrode.
216
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
Fig. 1. Magnetic resonance imaging. Precontrast T1-weighted MRI demonstrated the prior (A) left temporal lobectomy and (B) left frontotemporoparietal lesion. (C) Postcontrast T1-weighted MRI showed enhancement of the lesion. Biopsy revealed gliosis. (D) Depth electrodes were placed in the following regions of the left hemisphere: nucleus parafascicularis (Contact #1), DM nucleus (Contact #2), nucleus anteroprincipalis/anteroventralis (Contact #3), lateral ventricle (Contacts #4–5), corpus callosum (Contact #6), and MCC (Contacts #7–8). (E) A coronal T1-weighted MRI image further demonstrates the placement of the electrode contacts. (R = right, L = left, S = superior, I = inferior. Line drawings outline the caudate and thalamus.)
2.3. Data analysis Intracranial EEG was visually inspected to detect ictal events. Spectral power was computed across ictal states (preictal, ictal, and postictal) and level of consciousness (stupor/sleep vs. awake) in the DM nucleus and MCC using spectrograms (1-second window, 75% overlap, Goertzel Discrete Fourier Transform algorithm) of differential electrode signals. All signals had referential recording, in which the reference was a “flipped up” or “upside down” intracranial EEG electrode placed on the dura, prior to computing the differential signal (electrode 1–electrode 2). The differential signal was calculated from anatomically adjacent contacts, as seen in Fig. 1, using MATLAB digital re-referencing. Preictal and postictal were defined as periods of 30-second duration preseizure onset and postseizure offset, respectively. We compared gamma (30– 40 Hz) power for combinations of ictal state, level of consciousness, and anatomic location using a Kruskal–Wallis ANOVA and Mann– Whitney–Wilcoxon tests. A Bonferroni correction was used to adjust the significance level (PKW b 0.0250 and PMWW b 0.0017) for multiple comparisons.
3. Results The patient was a 45-year-old left-handed woman with medically refractory localization-related epilepsy since 15 years of age. Seizures were thought to relate to a possible history of encephalitis. The patient underwent a previous left temporal lobectomy and vagal nerve stimulator placement, but with continued seizures. The postoperative seizures were described primarily as sudden loss of consciousness with falls, occurring 10 times per day.
Long-term video scalp EEG monitoring captured 15 typical events, corresponding to rhythmic alpha and beta activity lateralized to the left hemisphere but difficult to further localize. The interictal EEG showed frequent left temporal epileptiform discharges, focal left temporal slowing, and a left temporal breach. Imaging included MRI demonstrating a left frontotemporoparietal enhancing lesion (Fig. 1A–C). Repeated neuropsychological testing was notable for variable deficits, including language impairment, verbal memory loss, and executive dysfunction. Given the need for better localization, a decision was made to proceed with intracranial monitoring. A total of 250 seizures were captured over five days. Eighty-seven seizures had associated video available. Semiology consisted of an atonic slump of the head to the left, left arm flexion, and loss of consciousness. Eleven seizures had no clear clinical correlate; the vast majority of these electrographic seizures occurred during sleep or stuporous states, with her head already positioned to the left such that a behavioral change would be difficult to detect. Seizures had multifocal onsets from the left temporal and frontal regions, although multiple seizures were nonlocalizable. Ictal activity invariably spread to the left MCC, often involving the right MCC as well. At baseline, the left DM nucleus demonstrated rhythmic bursts of gamma activity, evident more often and with greater amplitude during wakefulness. This activity ceased as ictal discharges spread to the MCC and consciousness was lost, and it recurred with the termination of each seizure as the patient regained awareness (Fig. 2). Analysis of the spectrograms demonstrated statistical significance of this DM pattern (Fig. 3). When seizures occurred during wakefulness (n = 43), there was lower DM ictal power (p b 0.0001) and higher DM postictal power (p b 0.0001) relative to the preictal epoch. As the seizures spread to the MCC, there was higher MCC ictal power (NS,
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
217
Fig. 2. Cessation of DM gamma (30–40 Hz) bursts during seizure propagation to the MCC. An example of a peri-ictal EEG recording from left thalamic/MCC depth electrodes (LTD 1–8) is shown. Contact #2 is in the DM nucleus, and Contacts #7–8 are in the MCC. Distal EEG onset refers to onset in cortical electrodes. The images reflect different filter settings applied to the same seizure. The conventional view (right) was obtained with a low frequency filter at 1 Hz, high frequency filter at 70 Hz, and notch filter at 60 Hz. The tailored filter settings (left) included a low frequency filter at 30 Hz, the high frequency filter set to “off”, and the notch filter at 60 Hz.
30–40 Hz; p b 0.05, 40–50 Hz) with lower MCC postictal power relative to the preictal epoch (p b 0.0001). This pattern of activity is also evident in the power spectral density plots, averaged across all seizures during wakefulness (Fig. 4). In contrast, this pattern was not clearly seen on the right, where the thalamic contact was placed more laterally, and the spread of ictal activity to the MCC was less robust. Nor was this spectral pattern evident when
seizures occurred during sleep (n = 44). While there were occasional sinusoidal bursts of gamma activity in the pre- and postictal periods during sleep (as seen in Fig. 5), their presence was variable with clearly less gamma activity than what was seen during wakefulness. Furthermore, no rebound of gamma activity was evident in the DM nucleus during the postictal period as had been seen after seizures occurring during wakefulness. When the subject was asleep, there was a higher ictal DM
Fig. 3. Spectrogram demonstrating loss of gamma power in the DM nucleus during ictal activity within the MCC. The spectrograms correspond to the seizure in Fig. 2. (A) Gamma power within the DM nucleus during pre- and postictal periods is not seen during MCC ictal activity. (B) An increase in gamma power during propagation of the seizure is evident in the MCC.
218
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
Fig. 4. Power spectral density (PSD) graphs of seizures during wakefulness. The PSD plots demonstrate the power for frequencies from 0 to 50 Hz, during the preictal (dark blue), ictal (red), and postictal (light blue) periods. The graph represents the mean power-frequency distribution for seizures occurring during the awake state (n = 43). (A) The PSD plot demonstrates lower ictal and higher postictal 30- to 40-Hz power relative to the preictal epoch within the DM nucleus (p b 0.0001 for both comparisons). (B) The PSD plot demonstrates lower postictal 30- to 40-Hz power relative to the preictal epoch within the MCC (p b 0.0001). The increase in ictal 30- to 40-Hz power compared to the preictal period was not statistically significant, while the increase in ictal 40- to 50-Hz power compared to the preictal period was significant at p b 0.05.
gamma power relative to the preictal epoch (p b 0.0001), but no significant difference between pre- and postictal 30- to 40-Hz activity. In the MCC, there was higher ictal and lower postictal power relative to the preictal epoch (p b 0.0001, both pairings). 4. Discussion Rhythmic bursts of 30- to 40-Hz gamma activity were evident within the DM nucleus at baseline, particularly during the awake state. The DM bursts terminated when ictal activity spread to the MCC and the patient lost consciousness, and recurred at the end of each seizure as consciousness was regained. This bursting pattern appeared to be characteristic of the DM nucleus, not seen in the other brain regions recorded. We posit that the seizures originated from the frontal and temporal lesions in this patient, and that frontal and limbic inputs to the MCC served as pathways for propagation of seizure activity. While our planned analysis focused on the 30- to 40-Hz range, an increase in 40to 50-Hz power within the MCC during seizures was evident in the power spectral density plot (Fig. 4). This 40- to 50-Hz activity likely represented ictal activity and can be seen in the raw EEG (Fig. 2). The MCC is believed to have reciprocal connectivity with the DM nucleus based upon nonhuman primate data [3,4]; hence, ictal activity within the MCC may have disrupted DM function. Alteration of DM activity may, in part, account for the impairment in cognition and consciousness seen with complex partial seizures, consistent with functional imaging, lesion, and connectivity studies suggesting that the DM nucleus participates in aspects of attention, executive function, memory, orientation, and arousal [5–7]. This finding is also consistent with previous research suggesting that consciousness is impaired in temporal lobe seizures due to diminished subcortical arousal [8]. By extension, altered interictal DM function (e.g., via interictal epileptiform activity or enduring effects of seizures) may contribute to the cognitive dysfunction evident on baseline neuropsychological testing. The functional significance of greater postictal DM gamma activity is unknown but may represent a rebound phenomenon following seizure termination. A role of DM gamma bursts during wakefulness in the maintenance of cognition and consciousness may also explain the relative absence of this gamma bursting pattern during sleep. Rodent data suggest that the DM nucleus participates in primary limbic seizure networks, and that the DM nucleus may act as a “control point” for seizure duration, propagation, and behavioral manifestations [9,10]. In humans with temporal lobe epilepsy, studies demonstrated preferential neuronal loss within the DM nucleus on postmortem immunohistochemical analysis [11], reduced glucose metabolism and benzodiazepine receptor binding in the ipsilateral DM nucleus on PET imaging
[12], and ipsilateral medial thalamic atrophy on MRI correlating with frontal and temporal cortical thinning and duration of epilepsy [13]. If the DM nucleus lies within the epileptogenic network, perhaps it could serve as an effective target for seizure therapy. While DM stimulation in a rat model of limbic epilepsy was ineffective [14], lesions of the DM nucleus prevented the pharmacologic induction of spike–wave discharges [15]. In humans receiving stimulation of the anterior nucleus of the thalamus for localization-related epilepsy, the deepest 1–2 electrode contacts may be located within the DM nucleus [16]. Although not stimulated directly, the current may spread to the DM nucleus. Whether this spread relates to the efficacy of stimulation is unknown. Stimulation of the thalamus has also been explored with respect to restoration of consciousness in vegetative or minimally conscious states [17]. Stimulation of the central lateral nucleus, for example, has been shown to decrease EEG slowing, increase cortical multiunit activity, and restore exploratory behavior when administered in the postictal period following secondarily generalized seizures in a rat model [18]. Effects of DM stimulation on arousal, however, are largely unknown. Westmoreland et al. [19] reported depth electrode placement within the left dorsal and right medial thalamus of a patient with right temporal seizures. Stimulation across these electrodes during sleep elicited behavioral and electrographic wakefulness. Electrode placement could not be verified by modern imaging techniques; however, the results suggest that the region of the DM nucleus plays a role in consciousness and that arousal can be produced by electrical stimulation of this area. To our knowledge, effects of selective DM stimulation in patients with epilepsy have not been studied and serve as a direction for future research. Alternative interpretations of the data should be considered. It is possible that the DM rhythmic bursting resulted from either the epileptogenic process or instrumentation and does not reflect normal activity. Seizure activity spreading into the MCC may impair cognitive processing and arousal, and cessation of the DM bursting could represent an epiphenomenon due to DM connectivity with the MCC or other regions. Disruption of consciousness may be due to propagation of the seizure to other brain regions that underlie arousal, or the seizures may have spread to other regions that led indirectly to depressed activity within the DM nucleus and/or impaired consciousness (“network inhibition hypothesis”), which may or may not have been captured with the available electrode contacts [8]. Finally, thalamic slowing late in the seizure, as seen in Fig. 2 (conventional view, right panel), may have contributed to an ongoing disruption of consciousness. This slow activity was unrelated to the onset of unresponsiveness, however, as it occurred well after the initial head slump and loss of awareness. Nevertheless, better understanding of the neural mechanisms underlying consciousness may help us develop new treatments that, if they do not abort seizures,
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
219
Fig. 5. Minimal pre- and postictal gamma (30–40 Hz) activity in the DM nucleus during sleep. An example of a peri-ictal EEG recording from left thalamic/MCC depth electrodes (LTD 1–8) during sleep is shown. Contact #2 is in the DM nucleus, and Contacts #7–8 are in the MCC. The images reflect different filter settings applied to the same seizure. (A) The tailored filter settings included a low frequency filter at 30 Hz, the high frequency filter set to “off”, and the notch filter at 60 Hz. (B) The conventional view was obtained with a low frequency filter at 1 Hz, high frequency filter at 100 Hz, and notch filter at 60 Hz. While there were occasional left DM sinusoidal bursts of gamma activity in the pre- and postictal periods during sleep, as evident in this example, their presence was variable with clearly less gamma activity than seen during wakefulness. In this seizure, high frequency activity was evident in the left DM nucleus as well as in the bilateral MCC (LTD #7–8 [left], RTD #7–8 [right]) during the ictal periods, likely representing the spread of seizure activity. No rebound of gamma activity was evident in the left DM nucleus during the postictal period as had been seen in seizures occurring during wakefulness.
may at least allow patients to retain awareness and lessen the impact of intractable epilepsy. Acknowledgment The authors wish to thank Klaus Mewes, PhD, for his assistance with electrode localization.
Disclosures Dr. Leeman-Markowski received a Clinical Research Training Fellowship from the American Brain Foundation. Dr. Smart reports no disclosures. Dr. Faught reports no disclosures. Dr. Gross reports no disclosures.
220
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
Dr. Meador is currently funded by NIH grants 2 U01 NS038455-13 and 3 U01 NS038455-13S1, and receives additional research support from NIH grant 1 R01 NS076665-01, PCORI 527, and the Human Epilepsy Project. Prior support was provided by UCB, Pfizer, and the Epilepsy Therapy Project. References [1] Lee KH, Meador KJ, Park YD, et al. Pathophysiology of altered consciousness during seizures: subtraction SPECT study. Neurology 2002;59:841–6. [2] Blumenfeld H. Impaired consciousness in epilepsy. Lancet Neurol 2012;11:814–26. [3] Giguere M, Goldman-Rakic PS. Mediodorsal nucleus: areal, laminar and tangential distribution of afferents and efferents in the frontal lobe of rhesus monkeys. J Comp Neurol 1988;277:195–213. [4] Russchen FT, Amaral DG, Price JL. The afferent input to the magnocellular division of the mediodorsal thalamic nucleus in the monkey, Macaca fascicularis. J Comp Neurol 1987;256:175–210. [5] Fernández-Espejo D, Junque C, Bernabeu M, Roig-Rovira T, Vendrell P, Mercader JM. Reductions of thalamic volume and regional shape changes in the vegetative and the minimally conscious states. J Neurotrauma 2010;27:1187–93. [6] Kumral E, Gulluoglu H, Dramali B. Thalamic chronotaraxis: isolated time disorientation. J Neurol Neurosurg Psychiatry 2007;78:880–2. [7] Schmahmann JD. Vascular syndromes of the thalamus. Stroke 2003;34:2264–78. [8] Motelow JE, Wei Li, Zhan Q, et al. Decreased subcortical cholinergic arousal in focal seizures. Neuron 2015;85:561–72. [9] Cassidy RM, Gale K. Mediodorsal thalamus plays a critical role in the development of limbic motor seizures. J Neurosci 1998;18(21):9002–9.
[10] Bertram EH, Zhang D, Williamson JM. Multiple roles of midline dorsal thalamic nuclei in induction and spread of limbic seizures. Epilepsia 2008;49:256–68. [11] Sinjab B, Martinian L, Sisodiya SM, Thom M. Regional thalamic neuropathology in patients with hippocampal sclerosis and epilepsy: a postmortem study. Epilepsia 2013;54:2125–33. [12] Juhász C, Nagy F, Watson C, et al. Glucose and [11C]flumazenil positron emission tomography abnormalities of thalamic nuclei in temporal lobe epilepsy. Neurology 1999;53:2037–45. [13] Bernhardt BC, Bernasconi N, Kim H, Bernasconi A. Mapping thalamocortical network pathology in temporal lobe epilepsy. Neurology 2012;78:129–36. [14] Wang S, Wu DC, Fan XN, et al. Mediodorsal thalamic stimulation is not protective against seizures induced by amygdaloid kindling in rats. Neurosci Lett 2010;481: 97–101. [15] Banerjee PK, Snead III OC. Thalamic mediodorsal and intralaminar nuclear lesions disrupt the generation of experimentally induced generalized absence-like seizures in rats. Epilepsy Res 1994;17:193–205. [16] Zumsteg D, Lozano AM, Wennberg RA. Rhythmic cortical EEG synchronization with low frequency stimulation of the anterior and medial thalamus for epilepsy. Clin Neurophysiol 2006;117:2272–8. [17] Gummadavelli A, Kundishora AJ, Willie JT, et al. Neurostimulation to improve level of consciousness in patients with epilepsy. Neurosurg Focus 2015;38(6):E10. [18] Gummadavelli A, Motelow JE, Smith N, Zhan Q, Schiff ND, Blumenfeld H. Thalamic stimulation to improve level of consciousness after seizures: evaluation of electrophysiology and behavior. Epilepsia 2015;56(1):114–24. [19] Westmoreland BF, Groover RV, Klass DW. Spontaneous sleep and induced arousal. A depth-electroencephalographic study. J Neurol Sci 1976;28:353–60.
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Cessation of gamma activity in the dorsomedial nucleus associated with loss of consciousness during focal seizures B.A. Leeman-Markowski a,⁎, O.L. Smart b, R.E. Faught a, R.E. Gross b, K.J. Meador c a b c
Department of Neurology, Emory University, 101 Woodruff Circle, Suite 6000, Atlanta, GA 30322, USA Department of Neurosurgery, Emory University, 101 Woodruff Circle, Suite 6000, Atlanta, GA 30322, USA Department of Neurology and Neurological Sciences, Stanford University, 300 Pasteur Drive (Room A343), Stanford, CA 94305-5235, USA
a r t i c l e
i n f o
Article history: Received 22 April 2015 Revised 20 June 2015 Accepted 20 July 2015 Available online 24 August 2015 Keywords: EEG Intracranial electrodes Complex partial seizures Thalamus Dorsomedial nucleus Consciousness Postictal state
a b s t r a c t Rationale: Impaired consciousness during seizures may be mediated by ictal propagation to the thalamus. Functions of individual thalamic nuclei with respect to consciousness, however, are largely unknown. The dorsomedial (DM) nucleus of the thalamus likely plays a role in arousal and cognition. We propose that alterations of firing patterns within the DM nucleus contribute to impaired arousal during focal seizures. Methods: Electroencephalograph data were collected from electrodes within the left DM thalamus and midcingulate cortex (MCC) in a patient undergoing seizure monitoring. Spectral power was computed across ictal states (preictal, ictal, and postictal) and level of consciousness (stupor/sleep vs. awake) in the DM nucleus and MCC. Results: Eighty-seven seizures of multifocal left frontal and temporal onsets were analyzed, characterized by loss of consciousness. At baseline, the left DM nucleus demonstrated rhythmic bursts of gamma activity, most frequently and with greatest amplitude during wakefulness. This activity ceased as ictal discharges spread to the MCC, and consciousness was impaired, and it recurred at the end of each seizure as awareness was regained. The analysis of gamma (30–40 Hz) power demonstrated that when seizures occurred during wakefulness, there was lower DM ictal power (p b 0.0001) and higher DM postictal power (p b 0.0001) relative to the preictal epoch. This spectral pattern was not evident within the MCC or when seizures occurred during sleep. Conclusions: Data revealed a characteristic pattern of DM gamma bursts during wakefulness, which disappeared during partial seizures associated with impaired consciousness. The findings are consistent with studies suggesting that the DM nucleus participates in cognition and arousal. © 2015 Elsevier Inc. All rights reserved.
1. Introduction
2. Methods
Focal seizures may impair consciousness, likely mediated, in part, by the spread of ictal activity to the thalamus [1,2]. The precise thalamic regions involved and their roles with respect to the maintenance of consciousness, however, are unclear. This report presents unique intracranial EEG data from a patient with localization-related epilepsy, using depth electrode contacts within the left dorsomedial (DM) nucleus of the thalamus and left midcingulate cortex (MCC). The DM nucleus demonstrated rhythmic bursts of gamma activity during wakefulness, which were abolished when ictal activity spread to the MCC and consciousness was impaired. We propose that taking the DM nucleus “offline” during ictal activity contributes to alteration of consciousness.
2.1. Subject
⁎ Corresponding author at: 2407 Avalon Pines Dr., Coram, NY 11727, USA. Tel.: +1 248 444 1302. E-mail addresses: [email protected] (B.A. Leeman-Markowski), [email protected] (O.L. Smart), [email protected] (R.E. Faught), [email protected] (R.E. Gross), [email protected] (K.J. Meador).
http://dx.doi.org/10.1016/j.yebeh.2015.07.027 1525-5050/© 2015 Elsevier Inc. All rights reserved.
Data were obtained from a single subject undergoing intracranial EEG recording for clinical indications.
2.2. EEG acquisition Bilateral 8-contact thalamic depth electrodes were implanted using stereotactic coordinates. Left-hemisphere Contact #2 was placed in the DM nucleus, while the corresponding right-hemisphere contact was placed more laterally, in the internal medullary lamina or nucleus ventrooralis internus. The most superficial contacts (#7–8) were located in the MCC bilaterally (Fig. 1D). Thalamic targets were selected to document possible subcortical spread of ictal activity, perhaps leading to an indication for future deep brain stimulation. The implantation also included a 10-contact left frontal–temporal strip and a 12-contact left frontal–hippocampal depth electrode.
216
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
Fig. 1. Magnetic resonance imaging. Precontrast T1-weighted MRI demonstrated the prior (A) left temporal lobectomy and (B) left frontotemporoparietal lesion. (C) Postcontrast T1-weighted MRI showed enhancement of the lesion. Biopsy revealed gliosis. (D) Depth electrodes were placed in the following regions of the left hemisphere: nucleus parafascicularis (Contact #1), DM nucleus (Contact #2), nucleus anteroprincipalis/anteroventralis (Contact #3), lateral ventricle (Contacts #4–5), corpus callosum (Contact #6), and MCC (Contacts #7–8). (E) A coronal T1-weighted MRI image further demonstrates the placement of the electrode contacts. (R = right, L = left, S = superior, I = inferior. Line drawings outline the caudate and thalamus.)
2.3. Data analysis Intracranial EEG was visually inspected to detect ictal events. Spectral power was computed across ictal states (preictal, ictal, and postictal) and level of consciousness (stupor/sleep vs. awake) in the DM nucleus and MCC using spectrograms (1-second window, 75% overlap, Goertzel Discrete Fourier Transform algorithm) of differential electrode signals. All signals had referential recording, in which the reference was a “flipped up” or “upside down” intracranial EEG electrode placed on the dura, prior to computing the differential signal (electrode 1–electrode 2). The differential signal was calculated from anatomically adjacent contacts, as seen in Fig. 1, using MATLAB digital re-referencing. Preictal and postictal were defined as periods of 30-second duration preseizure onset and postseizure offset, respectively. We compared gamma (30– 40 Hz) power for combinations of ictal state, level of consciousness, and anatomic location using a Kruskal–Wallis ANOVA and Mann– Whitney–Wilcoxon tests. A Bonferroni correction was used to adjust the significance level (PKW b 0.0250 and PMWW b 0.0017) for multiple comparisons.
3. Results The patient was a 45-year-old left-handed woman with medically refractory localization-related epilepsy since 15 years of age. Seizures were thought to relate to a possible history of encephalitis. The patient underwent a previous left temporal lobectomy and vagal nerve stimulator placement, but with continued seizures. The postoperative seizures were described primarily as sudden loss of consciousness with falls, occurring 10 times per day.
Long-term video scalp EEG monitoring captured 15 typical events, corresponding to rhythmic alpha and beta activity lateralized to the left hemisphere but difficult to further localize. The interictal EEG showed frequent left temporal epileptiform discharges, focal left temporal slowing, and a left temporal breach. Imaging included MRI demonstrating a left frontotemporoparietal enhancing lesion (Fig. 1A–C). Repeated neuropsychological testing was notable for variable deficits, including language impairment, verbal memory loss, and executive dysfunction. Given the need for better localization, a decision was made to proceed with intracranial monitoring. A total of 250 seizures were captured over five days. Eighty-seven seizures had associated video available. Semiology consisted of an atonic slump of the head to the left, left arm flexion, and loss of consciousness. Eleven seizures had no clear clinical correlate; the vast majority of these electrographic seizures occurred during sleep or stuporous states, with her head already positioned to the left such that a behavioral change would be difficult to detect. Seizures had multifocal onsets from the left temporal and frontal regions, although multiple seizures were nonlocalizable. Ictal activity invariably spread to the left MCC, often involving the right MCC as well. At baseline, the left DM nucleus demonstrated rhythmic bursts of gamma activity, evident more often and with greater amplitude during wakefulness. This activity ceased as ictal discharges spread to the MCC and consciousness was lost, and it recurred with the termination of each seizure as the patient regained awareness (Fig. 2). Analysis of the spectrograms demonstrated statistical significance of this DM pattern (Fig. 3). When seizures occurred during wakefulness (n = 43), there was lower DM ictal power (p b 0.0001) and higher DM postictal power (p b 0.0001) relative to the preictal epoch. As the seizures spread to the MCC, there was higher MCC ictal power (NS,
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
217
Fig. 2. Cessation of DM gamma (30–40 Hz) bursts during seizure propagation to the MCC. An example of a peri-ictal EEG recording from left thalamic/MCC depth electrodes (LTD 1–8) is shown. Contact #2 is in the DM nucleus, and Contacts #7–8 are in the MCC. Distal EEG onset refers to onset in cortical electrodes. The images reflect different filter settings applied to the same seizure. The conventional view (right) was obtained with a low frequency filter at 1 Hz, high frequency filter at 70 Hz, and notch filter at 60 Hz. The tailored filter settings (left) included a low frequency filter at 30 Hz, the high frequency filter set to “off”, and the notch filter at 60 Hz.
30–40 Hz; p b 0.05, 40–50 Hz) with lower MCC postictal power relative to the preictal epoch (p b 0.0001). This pattern of activity is also evident in the power spectral density plots, averaged across all seizures during wakefulness (Fig. 4). In contrast, this pattern was not clearly seen on the right, where the thalamic contact was placed more laterally, and the spread of ictal activity to the MCC was less robust. Nor was this spectral pattern evident when
seizures occurred during sleep (n = 44). While there were occasional sinusoidal bursts of gamma activity in the pre- and postictal periods during sleep (as seen in Fig. 5), their presence was variable with clearly less gamma activity than what was seen during wakefulness. Furthermore, no rebound of gamma activity was evident in the DM nucleus during the postictal period as had been seen after seizures occurring during wakefulness. When the subject was asleep, there was a higher ictal DM
Fig. 3. Spectrogram demonstrating loss of gamma power in the DM nucleus during ictal activity within the MCC. The spectrograms correspond to the seizure in Fig. 2. (A) Gamma power within the DM nucleus during pre- and postictal periods is not seen during MCC ictal activity. (B) An increase in gamma power during propagation of the seizure is evident in the MCC.
218
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
Fig. 4. Power spectral density (PSD) graphs of seizures during wakefulness. The PSD plots demonstrate the power for frequencies from 0 to 50 Hz, during the preictal (dark blue), ictal (red), and postictal (light blue) periods. The graph represents the mean power-frequency distribution for seizures occurring during the awake state (n = 43). (A) The PSD plot demonstrates lower ictal and higher postictal 30- to 40-Hz power relative to the preictal epoch within the DM nucleus (p b 0.0001 for both comparisons). (B) The PSD plot demonstrates lower postictal 30- to 40-Hz power relative to the preictal epoch within the MCC (p b 0.0001). The increase in ictal 30- to 40-Hz power compared to the preictal period was not statistically significant, while the increase in ictal 40- to 50-Hz power compared to the preictal period was significant at p b 0.05.
gamma power relative to the preictal epoch (p b 0.0001), but no significant difference between pre- and postictal 30- to 40-Hz activity. In the MCC, there was higher ictal and lower postictal power relative to the preictal epoch (p b 0.0001, both pairings). 4. Discussion Rhythmic bursts of 30- to 40-Hz gamma activity were evident within the DM nucleus at baseline, particularly during the awake state. The DM bursts terminated when ictal activity spread to the MCC and the patient lost consciousness, and recurred at the end of each seizure as consciousness was regained. This bursting pattern appeared to be characteristic of the DM nucleus, not seen in the other brain regions recorded. We posit that the seizures originated from the frontal and temporal lesions in this patient, and that frontal and limbic inputs to the MCC served as pathways for propagation of seizure activity. While our planned analysis focused on the 30- to 40-Hz range, an increase in 40to 50-Hz power within the MCC during seizures was evident in the power spectral density plot (Fig. 4). This 40- to 50-Hz activity likely represented ictal activity and can be seen in the raw EEG (Fig. 2). The MCC is believed to have reciprocal connectivity with the DM nucleus based upon nonhuman primate data [3,4]; hence, ictal activity within the MCC may have disrupted DM function. Alteration of DM activity may, in part, account for the impairment in cognition and consciousness seen with complex partial seizures, consistent with functional imaging, lesion, and connectivity studies suggesting that the DM nucleus participates in aspects of attention, executive function, memory, orientation, and arousal [5–7]. This finding is also consistent with previous research suggesting that consciousness is impaired in temporal lobe seizures due to diminished subcortical arousal [8]. By extension, altered interictal DM function (e.g., via interictal epileptiform activity or enduring effects of seizures) may contribute to the cognitive dysfunction evident on baseline neuropsychological testing. The functional significance of greater postictal DM gamma activity is unknown but may represent a rebound phenomenon following seizure termination. A role of DM gamma bursts during wakefulness in the maintenance of cognition and consciousness may also explain the relative absence of this gamma bursting pattern during sleep. Rodent data suggest that the DM nucleus participates in primary limbic seizure networks, and that the DM nucleus may act as a “control point” for seizure duration, propagation, and behavioral manifestations [9,10]. In humans with temporal lobe epilepsy, studies demonstrated preferential neuronal loss within the DM nucleus on postmortem immunohistochemical analysis [11], reduced glucose metabolism and benzodiazepine receptor binding in the ipsilateral DM nucleus on PET imaging
[12], and ipsilateral medial thalamic atrophy on MRI correlating with frontal and temporal cortical thinning and duration of epilepsy [13]. If the DM nucleus lies within the epileptogenic network, perhaps it could serve as an effective target for seizure therapy. While DM stimulation in a rat model of limbic epilepsy was ineffective [14], lesions of the DM nucleus prevented the pharmacologic induction of spike–wave discharges [15]. In humans receiving stimulation of the anterior nucleus of the thalamus for localization-related epilepsy, the deepest 1–2 electrode contacts may be located within the DM nucleus [16]. Although not stimulated directly, the current may spread to the DM nucleus. Whether this spread relates to the efficacy of stimulation is unknown. Stimulation of the thalamus has also been explored with respect to restoration of consciousness in vegetative or minimally conscious states [17]. Stimulation of the central lateral nucleus, for example, has been shown to decrease EEG slowing, increase cortical multiunit activity, and restore exploratory behavior when administered in the postictal period following secondarily generalized seizures in a rat model [18]. Effects of DM stimulation on arousal, however, are largely unknown. Westmoreland et al. [19] reported depth electrode placement within the left dorsal and right medial thalamus of a patient with right temporal seizures. Stimulation across these electrodes during sleep elicited behavioral and electrographic wakefulness. Electrode placement could not be verified by modern imaging techniques; however, the results suggest that the region of the DM nucleus plays a role in consciousness and that arousal can be produced by electrical stimulation of this area. To our knowledge, effects of selective DM stimulation in patients with epilepsy have not been studied and serve as a direction for future research. Alternative interpretations of the data should be considered. It is possible that the DM rhythmic bursting resulted from either the epileptogenic process or instrumentation and does not reflect normal activity. Seizure activity spreading into the MCC may impair cognitive processing and arousal, and cessation of the DM bursting could represent an epiphenomenon due to DM connectivity with the MCC or other regions. Disruption of consciousness may be due to propagation of the seizure to other brain regions that underlie arousal, or the seizures may have spread to other regions that led indirectly to depressed activity within the DM nucleus and/or impaired consciousness (“network inhibition hypothesis”), which may or may not have been captured with the available electrode contacts [8]. Finally, thalamic slowing late in the seizure, as seen in Fig. 2 (conventional view, right panel), may have contributed to an ongoing disruption of consciousness. This slow activity was unrelated to the onset of unresponsiveness, however, as it occurred well after the initial head slump and loss of awareness. Nevertheless, better understanding of the neural mechanisms underlying consciousness may help us develop new treatments that, if they do not abort seizures,
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
219
Fig. 5. Minimal pre- and postictal gamma (30–40 Hz) activity in the DM nucleus during sleep. An example of a peri-ictal EEG recording from left thalamic/MCC depth electrodes (LTD 1–8) during sleep is shown. Contact #2 is in the DM nucleus, and Contacts #7–8 are in the MCC. The images reflect different filter settings applied to the same seizure. (A) The tailored filter settings included a low frequency filter at 30 Hz, the high frequency filter set to “off”, and the notch filter at 60 Hz. (B) The conventional view was obtained with a low frequency filter at 1 Hz, high frequency filter at 100 Hz, and notch filter at 60 Hz. While there were occasional left DM sinusoidal bursts of gamma activity in the pre- and postictal periods during sleep, as evident in this example, their presence was variable with clearly less gamma activity than seen during wakefulness. In this seizure, high frequency activity was evident in the left DM nucleus as well as in the bilateral MCC (LTD #7–8 [left], RTD #7–8 [right]) during the ictal periods, likely representing the spread of seizure activity. No rebound of gamma activity was evident in the left DM nucleus during the postictal period as had been seen in seizures occurring during wakefulness.
may at least allow patients to retain awareness and lessen the impact of intractable epilepsy. Acknowledgment The authors wish to thank Klaus Mewes, PhD, for his assistance with electrode localization.
Disclosures Dr. Leeman-Markowski received a Clinical Research Training Fellowship from the American Brain Foundation. Dr. Smart reports no disclosures. Dr. Faught reports no disclosures. Dr. Gross reports no disclosures.
220
B.A. Leeman-Markowski et al. / Epilepsy & Behavior 51 (2015) 215–220
Dr. Meador is currently funded by NIH grants 2 U01 NS038455-13 and 3 U01 NS038455-13S1, and receives additional research support from NIH grant 1 R01 NS076665-01, PCORI 527, and the Human Epilepsy Project. Prior support was provided by UCB, Pfizer, and the Epilepsy Therapy Project. References [1] Lee KH, Meador KJ, Park YD, et al. Pathophysiology of altered consciousness during seizures: subtraction SPECT study. Neurology 2002;59:841–6. [2] Blumenfeld H. Impaired consciousness in epilepsy. Lancet Neurol 2012;11:814–26. [3] Giguere M, Goldman-Rakic PS. Mediodorsal nucleus: areal, laminar and tangential distribution of afferents and efferents in the frontal lobe of rhesus monkeys. J Comp Neurol 1988;277:195–213. [4] Russchen FT, Amaral DG, Price JL. The afferent input to the magnocellular division of the mediodorsal thalamic nucleus in the monkey, Macaca fascicularis. J Comp Neurol 1987;256:175–210. [5] Fernández-Espejo D, Junque C, Bernabeu M, Roig-Rovira T, Vendrell P, Mercader JM. Reductions of thalamic volume and regional shape changes in the vegetative and the minimally conscious states. J Neurotrauma 2010;27:1187–93. [6] Kumral E, Gulluoglu H, Dramali B. Thalamic chronotaraxis: isolated time disorientation. J Neurol Neurosurg Psychiatry 2007;78:880–2. [7] Schmahmann JD. Vascular syndromes of the thalamus. Stroke 2003;34:2264–78. [8] Motelow JE, Wei Li, Zhan Q, et al. Decreased subcortical cholinergic arousal in focal seizures. Neuron 2015;85:561–72. [9] Cassidy RM, Gale K. Mediodorsal thalamus plays a critical role in the development of limbic motor seizures. J Neurosci 1998;18(21):9002–9.
[10] Bertram EH, Zhang D, Williamson JM. Multiple roles of midline dorsal thalamic nuclei in induction and spread of limbic seizures. Epilepsia 2008;49:256–68. [11] Sinjab B, Martinian L, Sisodiya SM, Thom M. Regional thalamic neuropathology in patients with hippocampal sclerosis and epilepsy: a postmortem study. Epilepsia 2013;54:2125–33. [12] Juhász C, Nagy F, Watson C, et al. Glucose and [11C]flumazenil positron emission tomography abnormalities of thalamic nuclei in temporal lobe epilepsy. Neurology 1999;53:2037–45. [13] Bernhardt BC, Bernasconi N, Kim H, Bernasconi A. Mapping thalamocortical network pathology in temporal lobe epilepsy. Neurology 2012;78:129–36. [14] Wang S, Wu DC, Fan XN, et al. Mediodorsal thalamic stimulation is not protective against seizures induced by amygdaloid kindling in rats. Neurosci Lett 2010;481: 97–101. [15] Banerjee PK, Snead III OC. Thalamic mediodorsal and intralaminar nuclear lesions disrupt the generation of experimentally induced generalized absence-like seizures in rats. Epilepsy Res 1994;17:193–205. [16] Zumsteg D, Lozano AM, Wennberg RA. Rhythmic cortical EEG synchronization with low frequency stimulation of the anterior and medial thalamus for epilepsy. Clin Neurophysiol 2006;117:2272–8. [17] Gummadavelli A, Kundishora AJ, Willie JT, et al. Neurostimulation to improve level of consciousness in patients with epilepsy. Neurosurg Focus 2015;38(6):E10. [18] Gummadavelli A, Motelow JE, Smith N, Zhan Q, Schiff ND, Blumenfeld H. Thalamic stimulation to improve level of consciousness after seizures: evaluation of electrophysiology and behavior. Epilepsia 2015;56(1):114–24. [19] Westmoreland BF, Groover RV, Klass DW. Spontaneous sleep and induced arousal. A depth-electroencephalographic study. J Neurol Sci 1976;28:353–60.