- Email: [email protected]
Clinical neurophysiology of neurologic rehabilitation
Handbook of Clinical Neurology, Vol. 161 (3rd series) Clinical Neurophysiology: Diseases and Disorders K.H. Levin and P. Chauvel, Editors https://doi.org/10.1016/B978-0-444-64142-7.00048-5 Copyright © 2019 Elsevier B.V. All rights reserved
Chapter 11
Clinical neurophysiology of neurologic rehabilitation JENS D. ROLLNIK* Institute for Neurorehabilitation Research, BDH Clinic Hessisch Oldendorf, Hannover Medical School (MHH), Hessisch Oldendorf, Germany
Abstract Clinical neurophysiologic testing provides valuable support in predicting outcome in the setting of disorders of consciousness (DOC), including coma and traumatic brain injury (TBI). Electroencephalography (EEG) and evoked potentials (EP) are simple to apply, inexpensive, safe, and available in most rehabilitation facilities. This chapter reviews the use of EEG and EP in postanoxic coma and TBI. Bilateral absence of cortical somatosensory evoked potentials (SSEP) may be regarded as a predictor of poor outcome in hypoxic brain damage. Flash VEP may be useful to differentiate between good and poor outcome. In addition, low EEG frequencies, burst suppression, and isoelectric EEG patterns prognosticate poor outcomes in hypoxic brain damage. While a loss of cortical SSEP is generally regarded as a negative prognostic sign in the acute phase of hypoxic brain damage, absence of cortical SSEP responses is not necessarily associated with poor outcome in TBI. Event-related potentials (ERPs) can provide support in outcome prediction. In particular, the N100, mismatch negativity, P300, and N400 may improve accuracy of outcome prediction DOC of different etiologies. Some evidence suggests that ERPs may be superior to SSEP in predicting functional and DOC outcomes (Lew et al., 2003). ERPs are measured brain responses resulting from specific cognitive tasks, sensory stimulation, or planned motor activity.
INTRODUCTION Outcome prediction in severe neurologic disease remains challenging. While the immediate goal of acute-care hospital treatment is survival, the ultimate goal is recovery from coma and improved functional outcome during rehabilitation. Outcome is limited not only by the neurologic condition itself but also by comorbidities (Rollnik and Janosch, 2010) such as infection or colonization with multidrug resistant bacteria (Rollnik, 2014, 2015a). Along with clinical and neuroimaging findings, clinical neurophysiologic testing may help make more reliable outcome prediction (Rollnik, 2015a). Neurophysiologic techniques such as evoked potentials (EPs) and electroencephalography (EEG) are simple to apply, inexpensive, safe, and available in most rehabilitation facilities. Traumatic brain injury (TBI) shows a wider range of severities and anatomic localizations compared with
hypoxic brain damage; however, some clinical neurophysiologic studies have addressed outcome during long-term rehabilitation in TBI patients. There is some evidence suggesting that event-related potentials (ERPs) may be superior to somatosensory evoked potentials (SSEPs) in predicting functional and disorders of consciousness (DOC) outcomes (Lew et al., 2003). ERPs are measured brain responses resulting from specific cognitive tasks, sensory stimulation, or planned motor activity.
COMA OUTCOME Coma is a DOC characterized by unconsciousness and reflex behavior without eye opening, even to strong painful stimuli (Bodart et al., 2013; Rollnik and Altenm€uller, 2014). In the unresponsive wakefulness syndrome (UWS), previously known as vegetative state, eyes are
*Correspondence to: Prof. Dr. Jens D. Rollnik, BDH-Clinic Hessisch Oldendorf, Institute for Neurorehabilitation Research (InFo), Hannover Medical School, Greitstr. 18-28, 31840 Hessisch Oldendorf, Germany. Tel: +49-5152-781-231, Fax: +49-5152-781-198, E-mail: [email protected]
188
J.D. ROLLNIK
open and reflex behavior occurs, but patients are completely unresponsive (Bodart et al., 2013). Minimally conscious state (MCS) patients show signs of consciousness, such as command following (even if inconsistent), visual pursuit, localization to noxious stimulation, and appropriate responses to emotional stimuli without being able to functionally communicate (Bodart et al., 2013).
Electroencephalography Deep coma generally coincides with a burst-suppression pattern characterized by alternating episodes of isoelectric (“flat”) EEG and bursting slow waves (Amzica, 2015). Further deepening of the coma is associated with continuous isoelectric EEG (Amzica, 2015). There have been efforts to distinguish between UWS and MCS by means of quantitative EEG analysis (power spectrum) (Lehembre et al., 2012). In one study, the relative power in the delta band was higher in the UWS than in the MCS group (Lehembre et al., 2012). A drop in alpha power was observed in UWS compared to MCS patients (Lehembre et al., 2012). Hypoxic brain damage is the main neurologic condition associated with prolonged DOC in the setting of neurologic rehabilitation, in particular after cardiopulmonary resuscitation (Heinz and Rollnik, 2015). Thus this section focuses on hypoxic encephalopathy. Several meta-analyses have evaluated the prognostic value of EEG and EP in the acute phase of brain hypoxia (Wijdicks et al., 2006; Lee et al., 2010; Sandroni et al., 2013). In EEG recordings, suppression (flat EEG), burst suppression, alpha and theta pattern coma, and generalized periodic complexes have been considered the patterns with the worst prognosis (Wijdicks et al., 2006). Within an interval of less than 3 days from cardiopulmonary resuscitation to EEG recording, burst suppression or generalized epileptiform discharges may predict poor outcomes, but with insufficient prognostic accuracy (Wijdicks et al., 2006). In addition, it has been suggested that generalized suppression (amplitudes 20 mV) or generalized periodic complexes on a flat background may be associated with an outcome not better than UWS (Wijdicks et al., 2006). Another review confirmed that burst suppression or isoelectric EEG on the second or third day after coma onset had moderate diagnostic accuracy for poor outcome (Lee et al., 2010). Another meta-analysis concluded that generalized EEG suppression during the first 72 h may be a predictor of poor neurologic outcome (Sandroni et al., 2013). Fig. 11.1 shows an example of a 54-year-old female patient with burst suppression EEG and poor outcome (marginal improvement from UWS to MCS).
EEG proved to be of prognostic value in a sample of 106 UWS and MCS patients with different etiologies, whereas EP did not (Bagnato et al., 2015). Reduced EEG amplitudes and delta frequencies correlated with worse clinical outcomes after 3 months, while alpha frequencies and preserved EEG reactivity predicted better outcomes (Bagnato et al., 2015). One study enrolled 93 patients with hypoxic brain damage, admitted to neurologic rehabilitation 4 weeks after disease onset (Heinz and Rollnik, 2015); 30% of the patients were comatose upon admission, of which only 20% regained consciousness. After 3–4 months of inpatient rehabilitation, 75% had a poor outcome, using the Barthel functional independence index measure of <50. Patients in the good outcome group more frequently had alpha and less frequently theta or delta rhythms, compared with subjects with a poor treatment outcome.
Evoked potentials With respect to EPs, short-latency SSEPs, in particular of the median nerve, have been most frequently studied in the acute phase of hypoxic brain damage (Wijdicks et al., 2006; Lee et al., 2010; Sandroni et al., 2013) while brainstem auditory (AEPs) and visual evoked potentials (VEPs) have not been thoroughly tested for their prognostic value (Wijdicks et al., 2006). SSEPs are much less influenced by sedatives or metabolic influences as compared with EEG and may therefore be more accurate in prognostication (Wijdicks et al., 2006). Several studies have shown that bilateral absence of the SSEP N20 component with median nerve stimulation performed within 3 days after CPR (cardiopulmonary resuscitation) may predict poor outcomes (Wijdicks et al., 2006; Sandroni et al., 2013). One study concluded that bilateral loss of cortical SSEP responses within 24 h of coma onset predicts poor outcome (Lee et al., 2010). Upon admission to rehabilitation, about 4 weeks have passed since coma onset and only one-fourth of patients are still on mechanical ventilation (Pohl et al., 2016). This is a different group of surviving patients, but only a few studies have examined the value of EEG and EP for outcome prediction in this setting. An early study with UWS patients during rehabilitation could not find a correlation between SSEP and AEP and clinical outcomes (Zeitlhofer et al., 1991). A study of MCS patients with diffuse brain damage due to either brain trauma or hypoxia in the subacute phase suggested that SSEP may predict outcome better than AEP (Goldberg and Karazim, 1998). A retrospective cohort study followed 113 consecutive patients admitted to a German inpatient rehabilitation center with severe DOC due to hypoxic brain
CLINICAL NEUROPHYSIOLOGY OF NEUROLOGIC REHABILITATION
189
Fig. 11.1. Burst-suppression EEG of a female 54-year-old hypoxic brain-damaged patient, 10 weeks after cardiopulmonary resuscitation, 1 week after admission into a neurologic early rehabilitation center. Cortical SSEP responses (median nerve stimulation) were present, with latencies prolonged on the right side. The patient showed only moderate improvement of DOC, from UWS to MCS at discharge, 2 months later. An inconsistent visual pursuit and localization to stimulation could be observed.
damage. There was no correlation between loss of SSEP cortical responses and clinical outcomes (Howell et al., 2013). Another study concluded that VEP and SSEP results differed between the good and poor outcome groups (Heinz and Rollnik, 2015). With flash VEP, latency III was significantly prolonged in the poor outcome group (Fig. 11.2). With SSEP of the median nerve, the N20/ P25 amplitude was suppressed in the poor outcome group (on the left stimulation side only). AEP results were not different between the outcome groups.
Summary EEG in the acute phase of hypoxic brain damage seems to be a simple but reliable outcome predictor (Wijdicks et al., 2006; Lee et al., 2010; Sandroni et al., 2013). Suppression, burst suppression, delta frequencies, and generalized periodic complexes are associated with poor outcomes (Wijdicks et al., 2006; Bagnato et al., 2015). In neurologic early rehabilitation, 4 weeks after hypoxic brain damage, patients in the good outcome
group more frequently had alpha and less frequently theta or delta rhythms compared to subjects with poor outcomes (Heinz and Rollnik, 2015). Bilateral absence of the N20 component with median nerve SSEP recorded within 3 days after CPR may predict poor outcomes (Wijdicks et al., 2006; Lee et al., 2010; Sandroni et al., 2013). In neurologic rehabilitation, EP study results are inconsistent. While Howell et al. (2013) found no correlation between loss of SSEP cortical responses and clinical outcomes, other results indicate that flash VEP and SSEP may be useful to differentiate between good and poor outcome groups (Heinz and Rollnik, 2015).
PROGNOSTICATION IN POSTTRAUMATIC BRAIN INJURY While brain hypoxia exhibits a relatively homogeneous and symmetrical pattern of damage, TBI shows a wider range of severities and anatomic localizations. Extensive diffuse axonal injury frequently allows no recovery from coma or UWS (Jellinger, 2013) and comes closest to
190
J.D. ROLLNIK III I
III I
Right
II
Right
II
Left
Left
5 mV
5 mV
50 ms
A
50 ms
B
Fig. 11.2. (A) Flash VEP of a 60-year-old hypoxic brain-damaged patient with a good outcome. Latency III is 85 ms on the right and 86 ms on the left side (Heinz and Rollnik, 2015). (B) Flash VEP of a 50-year-old male with a poor outcome. While latencies I and II are comparable to the example shown in (A), latency III is considerably longer (108 ms right, 107 ms left) (Heinz and Rollnik, 2015).
patients with poor outcomes from brain hypoxia. It is quite difficult to compare different TBI and hypoxic brain damage studies and to draw generally applicable conclusions. However, some clinical neurophysiologic studies have addressed outcome during long-term rehabilitation in TBI patients.
Electroencephalography EEG may be helpful for outcome prediction in TBI. In one study with traumatic UWS patients, the EEG examination was performed repeatedly in 35 subjects from the first days up to several years after injury (Strnad and Strnadová, 1987). The authors found that a gradual improvement from delta to theta and alpha frequencies indicated a good outcome (Strnad and Strnadová, 1987). One report classified EEG patterns of patients with diffuse anoxic and traumatic encephalopathies into five major grades. Grade 3 was characterized by EEG with small amplitudes and diffuse, irregular delta activity. Grade 4 was characterized by burst suppression, and grade 5 by an isoelectric EEG pattern (Synek, 1988). All three grades were associated with poor clinical
outcomes. Another report found support for the hypothesis that delta EEG in comatose patients is associated with poor outcome 1 month after TBI (Beridze et al., 2010). A study with a mixed sample of neurologic early rehabilitation patients (including stroke, TBI, and hypoxia) demonstrated that EEG could predict outcome (Rollnik, 2015b). While about half of the patients with alpha activity belonged to the good outcome group (80/159, 50.3%), only 39/125 (31.2%) with theta activity and 5/41 (12.2%) with delta rhythm had a favorable outcome (w2 ¼ 24.2, P < 0.001). The Barthel index was significantly lower upon admission and at discharge when patients had theta or delta rhythms compared to alpha activity (P < 0.001). In addition, rehabilitation potential (Barthel index at discharge minus Barthel index upon admission) was smaller when patients had delta or theta activity (Fig. 11.3).
Evoked potentials Studies have found that brainstem AEP but not SSEP predicted long-term outcomes after 2 years (Mackey-Hargadine and Hall, 1985; Giaquinto, 2004).
CLINICAL NEUROPHYSIOLOGY OF NEUROLOGIC REHABILITATION 90 80 70 60 50 40 30 20 10 0
191
alpha theta delta
BI admission
BI discharge
Delta BI
Fig. 11.3. When patients had alpha EEG activity, Barthel index (BI) on admission and at discharge and changes of Barthel index (discharge minus admission) were significantly higher than in patients with theta or delta activity (Rollnik, 2015b). Table 11.1 Absence of median SEP on one or both sides and outcome categories (Rollnik, 2015b) Absence of cortical responses (median nerve SEP)
Outcome category Poor (Barthel index < 50) Good (Barthel index 50) Sum
None 200 153 353
Absence right 24 9 33
Absence left 35 12 47
Sum Bilateral absence 14 2 16
273 176 449
Mixed sample of neurologic early rehabilitation patients (e.g., stroke, TBI, anoxic encephalopathy)
This finding is not surprising because AEP may monitor brainstem function and injuries of the brainstem are generally unfavorable. It has been shown that lesions of the thalamus and brainstem are closely associated with poor outcome, in particular among MCS patients (Xu et al., 2016). In TBI, SSEP may help to predict outcome at 8 and 12 months (Shin et al., 1989; Mazzini et al., 1999; € Ozbudak-Demir et al., 1999). Comparable with hypoxic brain damage, bilateral absence of median nerve SSEP cortical responses appears to be strongly predictive of the worst functional outcome (Lew et al., 2003). The specificity and positive predictive value of absent SSEP for death or UWS at 6 months after TBI were as high as 100% in one study with a small sample size (Lew et al., 2003). One report suggested a median nerve SSEP grading system, ranging from grade 1 (bilaterally absent cortical SSEP responses) to grade 6 (bilaterally normal cortical SSEP responses). The study found that grade 3 results in comatose patients had the strongest relationship with the functional outcome (information-processing speed, working memory, and the ability to attend to tasks) 1 year after TBI (Houlden et al., 2010). In another study, outcome prediction by means of SSEP was compared between TBI and hypoxic patients (Schorl et al., 2014). While bilateral loss of
median SSEP cortical responses in hypoxic–ischemic encephalopathy proved to have a poor prognosis, with none of the patients achieving an outcome better than UWS, the situation was different in TBI (Schorl et al., 2014). With TBI, 14/28 patients (50%) with initial bilateral absent SSEP showed reoccurrence of cortical potentials, either uni- or bilaterally, but half of those subjects could be transferred to subsequent neurologic rehabilitation and achieved good functional long-term outcomes (Schorl et al., 2014). In other words, follow-up SSEP testing is required to enable a valid prediction, in particular during the acute phase of TBI. In a mixed sample of neurologic early rehabilitation patients, bilateral loss of SSEP was not uniformly associated with poor outcome, defined by a Barthel index below 50 points (Table 11.1; Rollnik, 2015b). However, while 43.3% (153/353) had a good outcome when SSEP responses were present, only 26.3% (21/80) belonged to the good outcome group when SSEP responses were absent on one side, and 12.5% (2/16) when SSEP responses were missing on both sides (Table 11.1). In regard to brainstem AEP, the loss of responses on one or both sides was associated with poor outcome (Rollnik, 2015b). Subjects with a bilateral loss of cortical flash VEP exhibited a poor outcome (Rollnik, 2015b).
J.D. ROLLNIK
192
Summary Similar to hypoxic brain damage, delta activity, burst suppression and isoelectric EEG patterns indicate poor outcome in TBI (Synek, 1988). During neurologic rehabilitation, improvement in functional independence (Barthel index) may be smaller when patients have delta or theta EEG frequencies (Rollnik, 2015b). While loss of cortical SSEP is generally regarded as a negative prognostic sign in the acute phase of postanoxic encephalopathy, absence of cortical SSEP responses is not uniformly associated with poor outcome in TBI (Schorl et al., 2014; Rollnik, 2015b). It is not uncommon for cortical responses to return during rehabilitation. Up to 50% of TBI patients with initial bilateral absence of SSEP showed reoccurrence of cortical potentials, either uni- or bilaterally (Schorl et al., 2014).
EVENT-RELATED POTENTIALS There is evidence suggesting that ERPs may be superior to SSEP in predicting functional and DOC outcomes (Lew et al., 2003). ERPs are measured brain responses resulting from specific cognitive tasks, sensory stimulation, or planned motor activity. Tasks with higher cognitive requirements, such as those used to measure the error related negativity (ERN) or the Bereitschaftspotential, may be used in conscious, cooperative subjects (Rollnik et al., 2004), but they are not usable in severe DOC. Therefore, this section focuses on (“passive”) ERP elicited in the absence of tasks. Relevant ERP components in this field are N100, mismatch negativity (MMN), P300, and N400 (Giaquinto, 2004; Daltrozzo et al., 2007). A metaanalysis revealed that MMN and P300 appear to be the most reliable predictors of awakening in poorly responsive patients with ischemic, hemorrhagic, traumatic, anoxic, postoperative, and metabolic etiologies of DOC (Daltrozzo et al., 2007). The presence of each of the N100, MMN, and P300 ERP components may be regarded as a highly significant predictor of awakening (Daltrozzo et al., 2007).
N100 The N100 or N1 ERP is a large, negative potential (peak between 80 and 120 ms after the onset of a stimulus), elicited by any unpredictable—primarily auditory— stimulus in the absence of task demands (Joos et al., 2014). In severe TBI, patients with normal N100 latency had a good outcome (Mazzini et al., 2001). When the stimulus was calling the patient’s name, N100 latency correlated significantly with functional outcome measures (Mazzini et al., 2001). One year after TBI, patients with poor outcome (death, UWS) showed a significantly
longer latency and lower amplitude of N100 compared to those with favorable outcomes (Mazzini et al., 2001).
Mismatch negativity MMN, also referred to as mismatch field, is a component of the ERP to an odd stimulus in a sequence of stimuli occurring in any sensory system. It has most frequently been studied with auditory and visual stimuli (N€a€at€anen et al., 1993). With auditory stimulation, the MMN occurs after an infrequent change in a repetitive sequence of sounds and can be elicited regardless of patient attention to the sequence (N€a€at€anen et al., 1993). This ERP may be regarded as an integrity index of cerebral processes that respond automatically to deviations from regularity in the acoustic environment (Giaquinto, 2004). Together with the P300, the MMN reflects discrimination of sounds (Daltrozzo et al., 2007). The presence of MMN predicted recovery from coma with a sensitivity of 89.7% and a specificity of 100% in one study (Kane et al., 1996). Its latency appeared to be the best indicator of 90-day outcome (r ¼ 0.641). With recovery from UWS to consciousness the ability to process auditory stimulus deviance may increase, indicating the consolidation of neural networks underlying overt communication (Wijnen et al., 2007). In a small study with 10 UWS patients, a sudden enhancement in MMN amplitude preceded overt communication with the environment and thus might be used to predict recovery from the vegetative state in the postacute phase after severe TBI (Wijnen et al., 2007).
P300 P300 is an endogenous potential linked to a subject’s reaction, not to the physical attributes of a stimulus. It probably reflects processes involved in cognitive information processing (e.g., memory, attention, executive function) (van Dinteren et al., 2014). The P300 has been used to quantify the extent and severity of TBI (Rappaport et al., 1990). The passive P300 may be recorded in a UWS patient, linked to an emotional response when a patient’s name is spoken (Marosi et al., 1993). In nontraumatic coma, the presence of a P300 appeared to be associated with awakening, but its absence did not exclude recovery from coma (Gott et al., 1991): 30% of the comatose patients studied had a preserved P300, which was associated with a higher Glasgow coma score, and 80% of patients exhibiting a P300 response regained conscious awareness (Gott et al., 1991).
CLINICAL NEUROPHYSIOLOGY OF NEUROLOGIC REHABILITATION
N400 The N400 is part of the response to words and other meaningful (or potentially meaningful) stimuli (Kutas and Federmeier, 2009). It peaks around 400 ms poststimulus (with a range between 250 and 550 ms) and is regarded as a response to a wide array of meaningful and potentially meaningful stimuli, including visual and auditory words (and word-like strings of letters), acronyms, sign language signs, pictures, environmental sounds, and gestures (Kutas and Federmeier, 2009). One study focused on the N400 and the late positive component (LPC) of the ERP in 30 MCS and UWS patients (Rohaut et al., 2015). While N400-like ERP components were recorded in the UWS, MCS, and control groups, only MCS patients and healthy controls showed an LPC response, suggesting that this late effect could be a potential specific marker of conscious semantic processing (Rohaut et al., 2015). The N400 might be useful to predict DOC outcomes (Steppacher et al., 2013). Cognitive ERP elicited by sound (P300) and speech (N400) were used to assess information processing in 92 behaviorally unresponsive patients diagnosed with either UWS (n ¼ 53) or MCS (n ¼ 39) (Steppacher et al., 2013). Within the first year of the condition, many patients showed an intact P300 and several also an N400, indicating residual higher-level semantic information processing. At clinical follow-up, about 25% of the patients recovered and regained communicative capabilities (Steppacher et al., 2013). In addition, there was correlation between the presence of the N400 and the recovery rate, but no association with the preservation of the P300 response (Steppacher et al., 2013).
Summary Among the ERPs, preservation of the N100, the MMN, and the P300 are highly significant predictors of awakening from DOC of different etiologies (Daltrozzo et al., 2007). A later component, called N400, indicates residual higher-level semantic information processing abilities and appears to be associated with recovery from UWS and MCS (Steppacher et al., 2013).
REFERENCES Amzica F (2015). What does burst suppression really mean? Epilepsy Behav 49: 234–237. Bagnato S, Boccagni C, Sant’Angelo A et al. (2015). EEG predictors of outcome in patients with disorders of consciousness admitted for intensive rehabilitation. Clin Neurophysiol 126: 959–966. Beridze M, Khaburzania M, Shakarishvili R et al. (2010). Dominated EEG patterns and their prognostic value in coma caused by traumatic brain injury. Georgian Med News 186: 28–33.
193
Bodart O, Laureys S, Gosseries O (2013). Coma and disorders of consciousness: scientific advances and practical considerations for clinicians. Semin Neurol 33: 83–90. Daltrozzo J, Wioland N, Mutschler V et al. (2007). Predicting coma and other low responsive patients outcome using event-related brain potentials: a meta-analysis. Clin Neurophysiol 118: 606–614. Giaquinto S (2004). Evoked potentials in rehabilitation. A review. Funct Neurol 19: 219–225. Goldberg G, Karazim E (1998). Application of evoked potentials to the prediction of discharge status in minimally responsive patients: a pilot study. J Head Trauma Rehabil 13: 51–68. Gott PS, Rabinowicz AL, DeGiorgio CM (1991). P300 auditory event-related potentials in nontraumatic coma. Association with Glasgow coma score and awakening. Arch Neurol 48: 1267–1270. Heinz U, Rollnik JD (2015). Outcome of hypoxic brain damage patients undergoing neurological early rehabilitation. BMC Res Notes 8: 243. Houlden DA, Taylor AB, Feinstein A (2010). Early somatosensory evoked potential grades in comatose traumatic brain injury patients predict cognitive and functional outcome. Crit Care Med 38: 167–174. Howell K, Grill E, Klein AM et al. (2013). Rehabilitation outcome of anoxic-ischaemic encephalopathy survivors with prolonged disorders of consciousness. Resuscitation 84: 1409–1415. Jellinger KA (2013). Neuropathology of prolonged unresponsive wakefulness syndrome after blunt head injury: review of 100 post-mortem cases. Brain Inj 27: 917–923. Joos K, Gilles A, Van de Heyning P et al. (2014). From sensation to percept: the neural signature of auditory eventrelated potentials. Neurosci Biobehav Rev 42: 148–156. Kane NM, Curry SH, Rowlands CA et al. (1996). Event-related potentials–neurophysiological tools for predicting emergence and early outcome from traumatic coma. Intensive Care Med 22: 39–46. Kutas M, Federmeier KD (2009). N400. Scholarpedia 4: 7790. Lee YC, Phan TG, Jolley DJ et al. (2010). Accuracy of clinical signs, SEP, and EEG in predicting outcome of hypoxic coma: a meta-analysis. Neurology 74: 572–580. Lehembre R, Bruno MA, Vanhaudenhuyse A et al. (2012). Resting-state EEG study of comatose patients: a connectivity and frequency analysis to find differences between vegetative and minimally conscious states. Funct Neurol 27: 41–47. Lew HL, Dikmen S, Slimp J et al. (2003). Use of somatosensoryevoked potentials and cognitive event-related potentials in predicting outcomes of patients with severe traumatic brain injury. Am J Phys Med Rehabil 82: 53–61. Mackey-Hargadine JR, Hall 3rd JW (1985). Sensory evoked responses in head injury. Cent Nerv Syst Trauma 2: 187–206. Marosi M, Prevec T, Masala C et al. (1993). Event-related potentials in vegetative state. Lancet 341: 1473. Mazzini L, Pisano F, Zaccala M et al. (1999). Somatosensory and motor evoked potentials at different stages of recovery from severe traumatic brain injury. Arch Phys Med Rehabil 80: 33–39.
194
J.D. ROLLNIK
Mazzini L, Zaccala M, Gareri F et al. (2001). Long-latency auditory-evoked potentials in severe traumatic brain injury. Arch Phys Med Rehabil 82: 57–65. N€a€at€anen R, Paavilainen P, Titinen H et al. (1993). Attention and mismatch negativity. Psychophysiology 30: 436–450. € Ozbudak-Demir S, Aky€uz M, G€uler-Uysal F et al. (1999). Postacute predictors of functional and cognitive progress in traumatic brain injury: somatosensory evoked potentials. Arch Phys Med Rehabil 80: 252–257. Pohl M, Bertram M, Bucka C et al. (2016). Course of rehabilitation in early neurological/neurosurgical rehabilitation. Results of a 2014 multi-center evaluation in Germany. Nervenarzt 87: 634–644. Rappaport M, Clifford Jr JO, Winterfield KM (1990). P300 response under active and passive attentional states and uni- and bimodality stimulus presentation conditions. J Neuropsychiatry Clin Neurosci 2: 399–407. Rohaut B, Faugeras F, Chausson N et al. (2015). Probing ERP correlates of verbal semantic processing in patients with impaired consciousness. Neuropsychologia 66: 279–292. Rollnik JD (2014). Outcome of MRSA carriers in neurological early rehabilitation. BMC Neurol 14: 34. Rollnik JD (2015a). Outcome of neurological early rehabilitation patients colonized with extended-spectrum beta-lactamase (ESBL) producing bacteria. Open J Ther Rehabil 3: 1–8. Rollnik JD (2015b). May clinical neurophysiology help to predict the recovery of neurological early rehabilitation patients? BMC Neurol 15: 239. Rollnik JD, Altenm€uller E (2014). Music in disorders of consciousness. Front Neurosci 8: 190. Rollnik JD, Janosch U (2010). Current trends in the length of stay in neurological early rehabilitation. Dtsch Arztebl Int 107: 286–292. Rollnik JD, Kohlmetz C, Rodrı´guez-Fornells A et al. (2004). Functional lesions and human action monitoring: combining repetitive transcranial magnetic stimulation and eventrelated brain potentials. Clin Neurophysiol 115: 145–153. Sandroni C, Cavallaro F, Callaway CW et al. (2013). Predictors of poor neurological outcome in adult comatose
survivors of cardiac arrest: a systematic review and metaanalysis. Part 1: patients not treated with therapeutic hypothermia. Resuscitation 84: 1310–1323. Schorl M, Valerius-Kukula SJ, Kemmer TP (2014). Medianevoked somatosensory potentials in severe brain injury: does initial loss of cortical potentials exclude recovery? Clin Neurol Neurosurg 123: 25–33. Shin DY, Ehrenberg B, Whyte J et al. (1989). Evoked potential assessment: utility in prognosis of chronic head injury. Arch Phys Med Rehabil 70: 189–193. Steppacher I, Eickhoff S, Jordanov T et al. (2013). N400 predicts recovery from disorders of consciousness. Ann Neurol 73: 594–602. Strnad P, Strnadova´ V (1987). Long-term follow-up EEG studies in patients with traumatic apallic syndrome. Eur Neurol 26: 84–89. Synek VM (1988). Prognostically important EEG coma patterns in diffuse anoxic and traumatic encephalopathies in adults. J Clin Neurophysiol 5: 161–174. van Dinteren R, Arns M, Jongsma MLA et al. (2014). P300 development across the lifespan: a systematic review and meta-analysis. PLoS One 9:e87347. Wijdicks EF, Hijdra A, Young GB et al. (2006). Practice parameter: prediction of outcome in comatose survivors after cardiopulmonary resuscitation (an evidence-based review): report of the quality standards Subcommittee of the American Academy of neurology. Neurology 67: 203–210. Wijnen VJ, van Boxtel GJ, Eilander HJ et al. (2007). Mismatch negativity predicts recovery from the vegetative state. Clin Neurophysiol 118: 597–605. Xu W, Kaur H, Wang X et al. (2016). The role of magnetic resonance imaging in the prediction of minimally conscious state after traumatic brain injury. World Neurosurg 94: 167–173. Zeitlhofer J, Steiner M, Oder W et al. (1991). The prognostic value of evoked potentials in early neurological rehabilitation of patients with the apallic syndrome. EEG EMG Z Elektroenzephalogr Elektromyogr Verwandte Geb 22: 10–14.
Chapter 11
Clinical neurophysiology of neurologic rehabilitation JENS D. ROLLNIK* Institute for Neurorehabilitation Research, BDH Clinic Hessisch Oldendorf, Hannover Medical School (MHH), Hessisch Oldendorf, Germany
Abstract Clinical neurophysiologic testing provides valuable support in predicting outcome in the setting of disorders of consciousness (DOC), including coma and traumatic brain injury (TBI). Electroencephalography (EEG) and evoked potentials (EP) are simple to apply, inexpensive, safe, and available in most rehabilitation facilities. This chapter reviews the use of EEG and EP in postanoxic coma and TBI. Bilateral absence of cortical somatosensory evoked potentials (SSEP) may be regarded as a predictor of poor outcome in hypoxic brain damage. Flash VEP may be useful to differentiate between good and poor outcome. In addition, low EEG frequencies, burst suppression, and isoelectric EEG patterns prognosticate poor outcomes in hypoxic brain damage. While a loss of cortical SSEP is generally regarded as a negative prognostic sign in the acute phase of hypoxic brain damage, absence of cortical SSEP responses is not necessarily associated with poor outcome in TBI. Event-related potentials (ERPs) can provide support in outcome prediction. In particular, the N100, mismatch negativity, P300, and N400 may improve accuracy of outcome prediction DOC of different etiologies. Some evidence suggests that ERPs may be superior to SSEP in predicting functional and DOC outcomes (Lew et al., 2003). ERPs are measured brain responses resulting from specific cognitive tasks, sensory stimulation, or planned motor activity.
INTRODUCTION Outcome prediction in severe neurologic disease remains challenging. While the immediate goal of acute-care hospital treatment is survival, the ultimate goal is recovery from coma and improved functional outcome during rehabilitation. Outcome is limited not only by the neurologic condition itself but also by comorbidities (Rollnik and Janosch, 2010) such as infection or colonization with multidrug resistant bacteria (Rollnik, 2014, 2015a). Along with clinical and neuroimaging findings, clinical neurophysiologic testing may help make more reliable outcome prediction (Rollnik, 2015a). Neurophysiologic techniques such as evoked potentials (EPs) and electroencephalography (EEG) are simple to apply, inexpensive, safe, and available in most rehabilitation facilities. Traumatic brain injury (TBI) shows a wider range of severities and anatomic localizations compared with
hypoxic brain damage; however, some clinical neurophysiologic studies have addressed outcome during long-term rehabilitation in TBI patients. There is some evidence suggesting that event-related potentials (ERPs) may be superior to somatosensory evoked potentials (SSEPs) in predicting functional and disorders of consciousness (DOC) outcomes (Lew et al., 2003). ERPs are measured brain responses resulting from specific cognitive tasks, sensory stimulation, or planned motor activity.
COMA OUTCOME Coma is a DOC characterized by unconsciousness and reflex behavior without eye opening, even to strong painful stimuli (Bodart et al., 2013; Rollnik and Altenm€uller, 2014). In the unresponsive wakefulness syndrome (UWS), previously known as vegetative state, eyes are
*Correspondence to: Prof. Dr. Jens D. Rollnik, BDH-Clinic Hessisch Oldendorf, Institute for Neurorehabilitation Research (InFo), Hannover Medical School, Greitstr. 18-28, 31840 Hessisch Oldendorf, Germany. Tel: +49-5152-781-231, Fax: +49-5152-781-198, E-mail: [email protected]
188
J.D. ROLLNIK
open and reflex behavior occurs, but patients are completely unresponsive (Bodart et al., 2013). Minimally conscious state (MCS) patients show signs of consciousness, such as command following (even if inconsistent), visual pursuit, localization to noxious stimulation, and appropriate responses to emotional stimuli without being able to functionally communicate (Bodart et al., 2013).
Electroencephalography Deep coma generally coincides with a burst-suppression pattern characterized by alternating episodes of isoelectric (“flat”) EEG and bursting slow waves (Amzica, 2015). Further deepening of the coma is associated with continuous isoelectric EEG (Amzica, 2015). There have been efforts to distinguish between UWS and MCS by means of quantitative EEG analysis (power spectrum) (Lehembre et al., 2012). In one study, the relative power in the delta band was higher in the UWS than in the MCS group (Lehembre et al., 2012). A drop in alpha power was observed in UWS compared to MCS patients (Lehembre et al., 2012). Hypoxic brain damage is the main neurologic condition associated with prolonged DOC in the setting of neurologic rehabilitation, in particular after cardiopulmonary resuscitation (Heinz and Rollnik, 2015). Thus this section focuses on hypoxic encephalopathy. Several meta-analyses have evaluated the prognostic value of EEG and EP in the acute phase of brain hypoxia (Wijdicks et al., 2006; Lee et al., 2010; Sandroni et al., 2013). In EEG recordings, suppression (flat EEG), burst suppression, alpha and theta pattern coma, and generalized periodic complexes have been considered the patterns with the worst prognosis (Wijdicks et al., 2006). Within an interval of less than 3 days from cardiopulmonary resuscitation to EEG recording, burst suppression or generalized epileptiform discharges may predict poor outcomes, but with insufficient prognostic accuracy (Wijdicks et al., 2006). In addition, it has been suggested that generalized suppression (amplitudes 20 mV) or generalized periodic complexes on a flat background may be associated with an outcome not better than UWS (Wijdicks et al., 2006). Another review confirmed that burst suppression or isoelectric EEG on the second or third day after coma onset had moderate diagnostic accuracy for poor outcome (Lee et al., 2010). Another meta-analysis concluded that generalized EEG suppression during the first 72 h may be a predictor of poor neurologic outcome (Sandroni et al., 2013). Fig. 11.1 shows an example of a 54-year-old female patient with burst suppression EEG and poor outcome (marginal improvement from UWS to MCS).
EEG proved to be of prognostic value in a sample of 106 UWS and MCS patients with different etiologies, whereas EP did not (Bagnato et al., 2015). Reduced EEG amplitudes and delta frequencies correlated with worse clinical outcomes after 3 months, while alpha frequencies and preserved EEG reactivity predicted better outcomes (Bagnato et al., 2015). One study enrolled 93 patients with hypoxic brain damage, admitted to neurologic rehabilitation 4 weeks after disease onset (Heinz and Rollnik, 2015); 30% of the patients were comatose upon admission, of which only 20% regained consciousness. After 3–4 months of inpatient rehabilitation, 75% had a poor outcome, using the Barthel functional independence index measure of <50. Patients in the good outcome group more frequently had alpha and less frequently theta or delta rhythms, compared with subjects with a poor treatment outcome.
Evoked potentials With respect to EPs, short-latency SSEPs, in particular of the median nerve, have been most frequently studied in the acute phase of hypoxic brain damage (Wijdicks et al., 2006; Lee et al., 2010; Sandroni et al., 2013) while brainstem auditory (AEPs) and visual evoked potentials (VEPs) have not been thoroughly tested for their prognostic value (Wijdicks et al., 2006). SSEPs are much less influenced by sedatives or metabolic influences as compared with EEG and may therefore be more accurate in prognostication (Wijdicks et al., 2006). Several studies have shown that bilateral absence of the SSEP N20 component with median nerve stimulation performed within 3 days after CPR (cardiopulmonary resuscitation) may predict poor outcomes (Wijdicks et al., 2006; Sandroni et al., 2013). One study concluded that bilateral loss of cortical SSEP responses within 24 h of coma onset predicts poor outcome (Lee et al., 2010). Upon admission to rehabilitation, about 4 weeks have passed since coma onset and only one-fourth of patients are still on mechanical ventilation (Pohl et al., 2016). This is a different group of surviving patients, but only a few studies have examined the value of EEG and EP for outcome prediction in this setting. An early study with UWS patients during rehabilitation could not find a correlation between SSEP and AEP and clinical outcomes (Zeitlhofer et al., 1991). A study of MCS patients with diffuse brain damage due to either brain trauma or hypoxia in the subacute phase suggested that SSEP may predict outcome better than AEP (Goldberg and Karazim, 1998). A retrospective cohort study followed 113 consecutive patients admitted to a German inpatient rehabilitation center with severe DOC due to hypoxic brain
CLINICAL NEUROPHYSIOLOGY OF NEUROLOGIC REHABILITATION
189
Fig. 11.1. Burst-suppression EEG of a female 54-year-old hypoxic brain-damaged patient, 10 weeks after cardiopulmonary resuscitation, 1 week after admission into a neurologic early rehabilitation center. Cortical SSEP responses (median nerve stimulation) were present, with latencies prolonged on the right side. The patient showed only moderate improvement of DOC, from UWS to MCS at discharge, 2 months later. An inconsistent visual pursuit and localization to stimulation could be observed.
damage. There was no correlation between loss of SSEP cortical responses and clinical outcomes (Howell et al., 2013). Another study concluded that VEP and SSEP results differed between the good and poor outcome groups (Heinz and Rollnik, 2015). With flash VEP, latency III was significantly prolonged in the poor outcome group (Fig. 11.2). With SSEP of the median nerve, the N20/ P25 amplitude was suppressed in the poor outcome group (on the left stimulation side only). AEP results were not different between the outcome groups.
Summary EEG in the acute phase of hypoxic brain damage seems to be a simple but reliable outcome predictor (Wijdicks et al., 2006; Lee et al., 2010; Sandroni et al., 2013). Suppression, burst suppression, delta frequencies, and generalized periodic complexes are associated with poor outcomes (Wijdicks et al., 2006; Bagnato et al., 2015). In neurologic early rehabilitation, 4 weeks after hypoxic brain damage, patients in the good outcome
group more frequently had alpha and less frequently theta or delta rhythms compared to subjects with poor outcomes (Heinz and Rollnik, 2015). Bilateral absence of the N20 component with median nerve SSEP recorded within 3 days after CPR may predict poor outcomes (Wijdicks et al., 2006; Lee et al., 2010; Sandroni et al., 2013). In neurologic rehabilitation, EP study results are inconsistent. While Howell et al. (2013) found no correlation between loss of SSEP cortical responses and clinical outcomes, other results indicate that flash VEP and SSEP may be useful to differentiate between good and poor outcome groups (Heinz and Rollnik, 2015).
PROGNOSTICATION IN POSTTRAUMATIC BRAIN INJURY While brain hypoxia exhibits a relatively homogeneous and symmetrical pattern of damage, TBI shows a wider range of severities and anatomic localizations. Extensive diffuse axonal injury frequently allows no recovery from coma or UWS (Jellinger, 2013) and comes closest to
190
J.D. ROLLNIK III I
III I
Right
II
Right
II
Left
Left
5 mV
5 mV
50 ms
A
50 ms
B
Fig. 11.2. (A) Flash VEP of a 60-year-old hypoxic brain-damaged patient with a good outcome. Latency III is 85 ms on the right and 86 ms on the left side (Heinz and Rollnik, 2015). (B) Flash VEP of a 50-year-old male with a poor outcome. While latencies I and II are comparable to the example shown in (A), latency III is considerably longer (108 ms right, 107 ms left) (Heinz and Rollnik, 2015).
patients with poor outcomes from brain hypoxia. It is quite difficult to compare different TBI and hypoxic brain damage studies and to draw generally applicable conclusions. However, some clinical neurophysiologic studies have addressed outcome during long-term rehabilitation in TBI patients.
Electroencephalography EEG may be helpful for outcome prediction in TBI. In one study with traumatic UWS patients, the EEG examination was performed repeatedly in 35 subjects from the first days up to several years after injury (Strnad and Strnadová, 1987). The authors found that a gradual improvement from delta to theta and alpha frequencies indicated a good outcome (Strnad and Strnadová, 1987). One report classified EEG patterns of patients with diffuse anoxic and traumatic encephalopathies into five major grades. Grade 3 was characterized by EEG with small amplitudes and diffuse, irregular delta activity. Grade 4 was characterized by burst suppression, and grade 5 by an isoelectric EEG pattern (Synek, 1988). All three grades were associated with poor clinical
outcomes. Another report found support for the hypothesis that delta EEG in comatose patients is associated with poor outcome 1 month after TBI (Beridze et al., 2010). A study with a mixed sample of neurologic early rehabilitation patients (including stroke, TBI, and hypoxia) demonstrated that EEG could predict outcome (Rollnik, 2015b). While about half of the patients with alpha activity belonged to the good outcome group (80/159, 50.3%), only 39/125 (31.2%) with theta activity and 5/41 (12.2%) with delta rhythm had a favorable outcome (w2 ¼ 24.2, P < 0.001). The Barthel index was significantly lower upon admission and at discharge when patients had theta or delta rhythms compared to alpha activity (P < 0.001). In addition, rehabilitation potential (Barthel index at discharge minus Barthel index upon admission) was smaller when patients had delta or theta activity (Fig. 11.3).
Evoked potentials Studies have found that brainstem AEP but not SSEP predicted long-term outcomes after 2 years (Mackey-Hargadine and Hall, 1985; Giaquinto, 2004).
CLINICAL NEUROPHYSIOLOGY OF NEUROLOGIC REHABILITATION 90 80 70 60 50 40 30 20 10 0
191
alpha theta delta
BI admission
BI discharge
Delta BI
Fig. 11.3. When patients had alpha EEG activity, Barthel index (BI) on admission and at discharge and changes of Barthel index (discharge minus admission) were significantly higher than in patients with theta or delta activity (Rollnik, 2015b). Table 11.1 Absence of median SEP on one or both sides and outcome categories (Rollnik, 2015b) Absence of cortical responses (median nerve SEP)
Outcome category Poor (Barthel index < 50) Good (Barthel index 50) Sum
None 200 153 353
Absence right 24 9 33
Absence left 35 12 47
Sum Bilateral absence 14 2 16
273 176 449
Mixed sample of neurologic early rehabilitation patients (e.g., stroke, TBI, anoxic encephalopathy)
This finding is not surprising because AEP may monitor brainstem function and injuries of the brainstem are generally unfavorable. It has been shown that lesions of the thalamus and brainstem are closely associated with poor outcome, in particular among MCS patients (Xu et al., 2016). In TBI, SSEP may help to predict outcome at 8 and 12 months (Shin et al., 1989; Mazzini et al., 1999; € Ozbudak-Demir et al., 1999). Comparable with hypoxic brain damage, bilateral absence of median nerve SSEP cortical responses appears to be strongly predictive of the worst functional outcome (Lew et al., 2003). The specificity and positive predictive value of absent SSEP for death or UWS at 6 months after TBI were as high as 100% in one study with a small sample size (Lew et al., 2003). One report suggested a median nerve SSEP grading system, ranging from grade 1 (bilaterally absent cortical SSEP responses) to grade 6 (bilaterally normal cortical SSEP responses). The study found that grade 3 results in comatose patients had the strongest relationship with the functional outcome (information-processing speed, working memory, and the ability to attend to tasks) 1 year after TBI (Houlden et al., 2010). In another study, outcome prediction by means of SSEP was compared between TBI and hypoxic patients (Schorl et al., 2014). While bilateral loss of
median SSEP cortical responses in hypoxic–ischemic encephalopathy proved to have a poor prognosis, with none of the patients achieving an outcome better than UWS, the situation was different in TBI (Schorl et al., 2014). With TBI, 14/28 patients (50%) with initial bilateral absent SSEP showed reoccurrence of cortical potentials, either uni- or bilaterally, but half of those subjects could be transferred to subsequent neurologic rehabilitation and achieved good functional long-term outcomes (Schorl et al., 2014). In other words, follow-up SSEP testing is required to enable a valid prediction, in particular during the acute phase of TBI. In a mixed sample of neurologic early rehabilitation patients, bilateral loss of SSEP was not uniformly associated with poor outcome, defined by a Barthel index below 50 points (Table 11.1; Rollnik, 2015b). However, while 43.3% (153/353) had a good outcome when SSEP responses were present, only 26.3% (21/80) belonged to the good outcome group when SSEP responses were absent on one side, and 12.5% (2/16) when SSEP responses were missing on both sides (Table 11.1). In regard to brainstem AEP, the loss of responses on one or both sides was associated with poor outcome (Rollnik, 2015b). Subjects with a bilateral loss of cortical flash VEP exhibited a poor outcome (Rollnik, 2015b).
J.D. ROLLNIK
192
Summary Similar to hypoxic brain damage, delta activity, burst suppression and isoelectric EEG patterns indicate poor outcome in TBI (Synek, 1988). During neurologic rehabilitation, improvement in functional independence (Barthel index) may be smaller when patients have delta or theta EEG frequencies (Rollnik, 2015b). While loss of cortical SSEP is generally regarded as a negative prognostic sign in the acute phase of postanoxic encephalopathy, absence of cortical SSEP responses is not uniformly associated with poor outcome in TBI (Schorl et al., 2014; Rollnik, 2015b). It is not uncommon for cortical responses to return during rehabilitation. Up to 50% of TBI patients with initial bilateral absence of SSEP showed reoccurrence of cortical potentials, either uni- or bilaterally (Schorl et al., 2014).
EVENT-RELATED POTENTIALS There is evidence suggesting that ERPs may be superior to SSEP in predicting functional and DOC outcomes (Lew et al., 2003). ERPs are measured brain responses resulting from specific cognitive tasks, sensory stimulation, or planned motor activity. Tasks with higher cognitive requirements, such as those used to measure the error related negativity (ERN) or the Bereitschaftspotential, may be used in conscious, cooperative subjects (Rollnik et al., 2004), but they are not usable in severe DOC. Therefore, this section focuses on (“passive”) ERP elicited in the absence of tasks. Relevant ERP components in this field are N100, mismatch negativity (MMN), P300, and N400 (Giaquinto, 2004; Daltrozzo et al., 2007). A metaanalysis revealed that MMN and P300 appear to be the most reliable predictors of awakening in poorly responsive patients with ischemic, hemorrhagic, traumatic, anoxic, postoperative, and metabolic etiologies of DOC (Daltrozzo et al., 2007). The presence of each of the N100, MMN, and P300 ERP components may be regarded as a highly significant predictor of awakening (Daltrozzo et al., 2007).
N100 The N100 or N1 ERP is a large, negative potential (peak between 80 and 120 ms after the onset of a stimulus), elicited by any unpredictable—primarily auditory— stimulus in the absence of task demands (Joos et al., 2014). In severe TBI, patients with normal N100 latency had a good outcome (Mazzini et al., 2001). When the stimulus was calling the patient’s name, N100 latency correlated significantly with functional outcome measures (Mazzini et al., 2001). One year after TBI, patients with poor outcome (death, UWS) showed a significantly
longer latency and lower amplitude of N100 compared to those with favorable outcomes (Mazzini et al., 2001).
Mismatch negativity MMN, also referred to as mismatch field, is a component of the ERP to an odd stimulus in a sequence of stimuli occurring in any sensory system. It has most frequently been studied with auditory and visual stimuli (N€a€at€anen et al., 1993). With auditory stimulation, the MMN occurs after an infrequent change in a repetitive sequence of sounds and can be elicited regardless of patient attention to the sequence (N€a€at€anen et al., 1993). This ERP may be regarded as an integrity index of cerebral processes that respond automatically to deviations from regularity in the acoustic environment (Giaquinto, 2004). Together with the P300, the MMN reflects discrimination of sounds (Daltrozzo et al., 2007). The presence of MMN predicted recovery from coma with a sensitivity of 89.7% and a specificity of 100% in one study (Kane et al., 1996). Its latency appeared to be the best indicator of 90-day outcome (r ¼ 0.641). With recovery from UWS to consciousness the ability to process auditory stimulus deviance may increase, indicating the consolidation of neural networks underlying overt communication (Wijnen et al., 2007). In a small study with 10 UWS patients, a sudden enhancement in MMN amplitude preceded overt communication with the environment and thus might be used to predict recovery from the vegetative state in the postacute phase after severe TBI (Wijnen et al., 2007).
P300 P300 is an endogenous potential linked to a subject’s reaction, not to the physical attributes of a stimulus. It probably reflects processes involved in cognitive information processing (e.g., memory, attention, executive function) (van Dinteren et al., 2014). The P300 has been used to quantify the extent and severity of TBI (Rappaport et al., 1990). The passive P300 may be recorded in a UWS patient, linked to an emotional response when a patient’s name is spoken (Marosi et al., 1993). In nontraumatic coma, the presence of a P300 appeared to be associated with awakening, but its absence did not exclude recovery from coma (Gott et al., 1991): 30% of the comatose patients studied had a preserved P300, which was associated with a higher Glasgow coma score, and 80% of patients exhibiting a P300 response regained conscious awareness (Gott et al., 1991).
CLINICAL NEUROPHYSIOLOGY OF NEUROLOGIC REHABILITATION
N400 The N400 is part of the response to words and other meaningful (or potentially meaningful) stimuli (Kutas and Federmeier, 2009). It peaks around 400 ms poststimulus (with a range between 250 and 550 ms) and is regarded as a response to a wide array of meaningful and potentially meaningful stimuli, including visual and auditory words (and word-like strings of letters), acronyms, sign language signs, pictures, environmental sounds, and gestures (Kutas and Federmeier, 2009). One study focused on the N400 and the late positive component (LPC) of the ERP in 30 MCS and UWS patients (Rohaut et al., 2015). While N400-like ERP components were recorded in the UWS, MCS, and control groups, only MCS patients and healthy controls showed an LPC response, suggesting that this late effect could be a potential specific marker of conscious semantic processing (Rohaut et al., 2015). The N400 might be useful to predict DOC outcomes (Steppacher et al., 2013). Cognitive ERP elicited by sound (P300) and speech (N400) were used to assess information processing in 92 behaviorally unresponsive patients diagnosed with either UWS (n ¼ 53) or MCS (n ¼ 39) (Steppacher et al., 2013). Within the first year of the condition, many patients showed an intact P300 and several also an N400, indicating residual higher-level semantic information processing. At clinical follow-up, about 25% of the patients recovered and regained communicative capabilities (Steppacher et al., 2013). In addition, there was correlation between the presence of the N400 and the recovery rate, but no association with the preservation of the P300 response (Steppacher et al., 2013).
Summary Among the ERPs, preservation of the N100, the MMN, and the P300 are highly significant predictors of awakening from DOC of different etiologies (Daltrozzo et al., 2007). A later component, called N400, indicates residual higher-level semantic information processing abilities and appears to be associated with recovery from UWS and MCS (Steppacher et al., 2013).
REFERENCES Amzica F (2015). What does burst suppression really mean? Epilepsy Behav 49: 234–237. Bagnato S, Boccagni C, Sant’Angelo A et al. (2015). EEG predictors of outcome in patients with disorders of consciousness admitted for intensive rehabilitation. Clin Neurophysiol 126: 959–966. Beridze M, Khaburzania M, Shakarishvili R et al. (2010). Dominated EEG patterns and their prognostic value in coma caused by traumatic brain injury. Georgian Med News 186: 28–33.
193
Bodart O, Laureys S, Gosseries O (2013). Coma and disorders of consciousness: scientific advances and practical considerations for clinicians. Semin Neurol 33: 83–90. Daltrozzo J, Wioland N, Mutschler V et al. (2007). Predicting coma and other low responsive patients outcome using event-related brain potentials: a meta-analysis. Clin Neurophysiol 118: 606–614. Giaquinto S (2004). Evoked potentials in rehabilitation. A review. Funct Neurol 19: 219–225. Goldberg G, Karazim E (1998). Application of evoked potentials to the prediction of discharge status in minimally responsive patients: a pilot study. J Head Trauma Rehabil 13: 51–68. Gott PS, Rabinowicz AL, DeGiorgio CM (1991). P300 auditory event-related potentials in nontraumatic coma. Association with Glasgow coma score and awakening. Arch Neurol 48: 1267–1270. Heinz U, Rollnik JD (2015). Outcome of hypoxic brain damage patients undergoing neurological early rehabilitation. BMC Res Notes 8: 243. Houlden DA, Taylor AB, Feinstein A (2010). Early somatosensory evoked potential grades in comatose traumatic brain injury patients predict cognitive and functional outcome. Crit Care Med 38: 167–174. Howell K, Grill E, Klein AM et al. (2013). Rehabilitation outcome of anoxic-ischaemic encephalopathy survivors with prolonged disorders of consciousness. Resuscitation 84: 1409–1415. Jellinger KA (2013). Neuropathology of prolonged unresponsive wakefulness syndrome after blunt head injury: review of 100 post-mortem cases. Brain Inj 27: 917–923. Joos K, Gilles A, Van de Heyning P et al. (2014). From sensation to percept: the neural signature of auditory eventrelated potentials. Neurosci Biobehav Rev 42: 148–156. Kane NM, Curry SH, Rowlands CA et al. (1996). Event-related potentials–neurophysiological tools for predicting emergence and early outcome from traumatic coma. Intensive Care Med 22: 39–46. Kutas M, Federmeier KD (2009). N400. Scholarpedia 4: 7790. Lee YC, Phan TG, Jolley DJ et al. (2010). Accuracy of clinical signs, SEP, and EEG in predicting outcome of hypoxic coma: a meta-analysis. Neurology 74: 572–580. Lehembre R, Bruno MA, Vanhaudenhuyse A et al. (2012). Resting-state EEG study of comatose patients: a connectivity and frequency analysis to find differences between vegetative and minimally conscious states. Funct Neurol 27: 41–47. Lew HL, Dikmen S, Slimp J et al. (2003). Use of somatosensoryevoked potentials and cognitive event-related potentials in predicting outcomes of patients with severe traumatic brain injury. Am J Phys Med Rehabil 82: 53–61. Mackey-Hargadine JR, Hall 3rd JW (1985). Sensory evoked responses in head injury. Cent Nerv Syst Trauma 2: 187–206. Marosi M, Prevec T, Masala C et al. (1993). Event-related potentials in vegetative state. Lancet 341: 1473. Mazzini L, Pisano F, Zaccala M et al. (1999). Somatosensory and motor evoked potentials at different stages of recovery from severe traumatic brain injury. Arch Phys Med Rehabil 80: 33–39.
194
J.D. ROLLNIK
Mazzini L, Zaccala M, Gareri F et al. (2001). Long-latency auditory-evoked potentials in severe traumatic brain injury. Arch Phys Med Rehabil 82: 57–65. N€a€at€anen R, Paavilainen P, Titinen H et al. (1993). Attention and mismatch negativity. Psychophysiology 30: 436–450. € Ozbudak-Demir S, Aky€uz M, G€uler-Uysal F et al. (1999). Postacute predictors of functional and cognitive progress in traumatic brain injury: somatosensory evoked potentials. Arch Phys Med Rehabil 80: 252–257. Pohl M, Bertram M, Bucka C et al. (2016). Course of rehabilitation in early neurological/neurosurgical rehabilitation. Results of a 2014 multi-center evaluation in Germany. Nervenarzt 87: 634–644. Rappaport M, Clifford Jr JO, Winterfield KM (1990). P300 response under active and passive attentional states and uni- and bimodality stimulus presentation conditions. J Neuropsychiatry Clin Neurosci 2: 399–407. Rohaut B, Faugeras F, Chausson N et al. (2015). Probing ERP correlates of verbal semantic processing in patients with impaired consciousness. Neuropsychologia 66: 279–292. Rollnik JD (2014). Outcome of MRSA carriers in neurological early rehabilitation. BMC Neurol 14: 34. Rollnik JD (2015a). Outcome of neurological early rehabilitation patients colonized with extended-spectrum beta-lactamase (ESBL) producing bacteria. Open J Ther Rehabil 3: 1–8. Rollnik JD (2015b). May clinical neurophysiology help to predict the recovery of neurological early rehabilitation patients? BMC Neurol 15: 239. Rollnik JD, Altenm€uller E (2014). Music in disorders of consciousness. Front Neurosci 8: 190. Rollnik JD, Janosch U (2010). Current trends in the length of stay in neurological early rehabilitation. Dtsch Arztebl Int 107: 286–292. Rollnik JD, Kohlmetz C, Rodrı´guez-Fornells A et al. (2004). Functional lesions and human action monitoring: combining repetitive transcranial magnetic stimulation and eventrelated brain potentials. Clin Neurophysiol 115: 145–153. Sandroni C, Cavallaro F, Callaway CW et al. (2013). Predictors of poor neurological outcome in adult comatose
survivors of cardiac arrest: a systematic review and metaanalysis. Part 1: patients not treated with therapeutic hypothermia. Resuscitation 84: 1310–1323. Schorl M, Valerius-Kukula SJ, Kemmer TP (2014). Medianevoked somatosensory potentials in severe brain injury: does initial loss of cortical potentials exclude recovery? Clin Neurol Neurosurg 123: 25–33. Shin DY, Ehrenberg B, Whyte J et al. (1989). Evoked potential assessment: utility in prognosis of chronic head injury. Arch Phys Med Rehabil 70: 189–193. Steppacher I, Eickhoff S, Jordanov T et al. (2013). N400 predicts recovery from disorders of consciousness. Ann Neurol 73: 594–602. Strnad P, Strnadova´ V (1987). Long-term follow-up EEG studies in patients with traumatic apallic syndrome. Eur Neurol 26: 84–89. Synek VM (1988). Prognostically important EEG coma patterns in diffuse anoxic and traumatic encephalopathies in adults. J Clin Neurophysiol 5: 161–174. van Dinteren R, Arns M, Jongsma MLA et al. (2014). P300 development across the lifespan: a systematic review and meta-analysis. PLoS One 9:e87347. Wijdicks EF, Hijdra A, Young GB et al. (2006). Practice parameter: prediction of outcome in comatose survivors after cardiopulmonary resuscitation (an evidence-based review): report of the quality standards Subcommittee of the American Academy of neurology. Neurology 67: 203–210. Wijnen VJ, van Boxtel GJ, Eilander HJ et al. (2007). Mismatch negativity predicts recovery from the vegetative state. Clin Neurophysiol 118: 597–605. Xu W, Kaur H, Wang X et al. (2016). The role of magnetic resonance imaging in the prediction of minimally conscious state after traumatic brain injury. World Neurosurg 94: 167–173. Zeitlhofer J, Steiner M, Oder W et al. (1991). The prognostic value of evoked potentials in early neurological rehabilitation of patients with the apallic syndrome. EEG EMG Z Elektroenzephalogr Elektromyogr Verwandte Geb 22: 10–14.