Long-latency auditory-evoked potentials in severe traumatic brain injury☆☆☆★

Article Outline

• Abstract
• Methods
  • Brainstem auditory-evoked potentials
  • LLAEPs
  • Data analysis
  • Statistical analysis
• Results
  • N100 and P150
  • P250 and P300
• Discussion
• Conclusions
• Acknowledgements
• Suppliers
• References
• Copyright

Abstract 

Mazzini L, Zaccala M, Gareri F, Giordano A, Angelino E. Long-latency auditory-evoked potentials in severe traumatic brain injury. Arch Phys Med Rehabil 2001;82:57-65. Objective: To detect the effects of different deviant stimuli on long-latency auditory-evoked potentials (LLAEPs) in patients with severe impairment of consciousness from traumatic brain injury (TBI) and to define their prognostic value for late functional outcome. Design: Correlational study on a prospective cohort. Setting: Brain injury rehabilitation center. Patients: Eleven volunteers and 21 consecutively sampled patients with severe TBI referred to the inpatient intensive rehabilitation unit of primary care in a university-based system. Main Outcome Measures: The LLAEPs recorded with different paradigms; and the Glasgow Outcome Scale (GOS), Disability Rating Scale (DRS), FIM™ instrument, and Neurobehavioural Rating Scale (NBHRS). Results: N100-P150 complex showed high reliability. Patients with good outcomes showed N100 and P150 mean latencies similar to those of unimpaired patients and shorter than patients with unfavorable outcomes. When the deviant stimulus was the patient's name, N100 latency showed high correlations with DRS (p < .007), FIM (p < .01), and NBHRS (p < .009). P250 and P300 showed a low percentage of occurrence with passive paradigms in both patients and controls. Their scores were inversely correlated to the Glasgow Coma Scale (p < .03) and the Innsbruck Coma Scale (p < .003), but no significant correlations were found with functional and behavioral outcomes. Patients with GOS score 1-2 1 year posttrauma had significantly longer latency and lower amplitude of N100 and P150 than those with GOS score 4-5. Conclusions: LLAEPs can be recorded in patients with severe impairment of consciousness by means of passive paradigms. The use of a stimulus that is relevant for the patient can enhance the accuracy of the test and its relationship with functional outcome. © 2001 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation

Keywords:  Auditory, Brain injuries, Evoked potentials, Rehabilitation

AUDITORY-EVOKED potentials have been widely studied in patients with severe traumatic brain injury (TBI),¹, ², ³ particularly in intensive care units, because of the ease of the method and its high reproducibility.⁴ Most studies have focused on short latency potentials reflecting brainstem activity but a few have investigated long-latency auditory-evoked potentials (LLAEPs), either in the acute phase⁵, ⁶ or in the long term.⁷, ⁸, ⁹ Cortical activity is first observed in the 15 to 50msec of the midlatency potentials, and these are followed up over the next several hundred milliseconds by several components that are influenced by the type and configuration of the stimulus. These waves are related to aspects of both sensory and cognitive processing¹⁰ and include a sequence of negative and positive components: a negative wave (N100), a positive wave (P150), a positive wave (P250), and a positive component at approximately 300ms (P300).¹¹

Most studies of the long latency potentials focus on the P300 wave.¹², ¹³, ¹⁴, ¹⁵, ¹⁶, ¹⁷ Usually this wave is obtained by an “oddball” paradigm that requires the subject to detect an improbable target or “oddball” stimulus that occurs unpredictably in a sequence of standard stimuli.⁹, ¹¹ A vertex-positive peak is obtained only after the rare stimuli at a latency of approximately 300ms. This paradigm is usually active because the subject must act on detecting the target stimulus.¹⁴ Thus, its clinical use is limited to collaborative patients, and all subjects with vigilance impairment are excluded. Efforts have been made to elicit a typical parietally distributed P300 with procedures that do not use an active task¹¹, ¹⁸ in control subjects. Farwell and Donchin¹⁹ showed that the P300 is enhanced when relevant stimuli are presented, even when the stimuli are not explicitly task relevant. Berlad and Pratt²⁰ reported that P300 amplitude was larger when the subject's name was used as a relevant stimulus in a passive oddball paradigm. The effects of relevant stimuli on the N100-P150 complex are unknown.

In this study, we recorded LLAEPs by using 3 different passive paradigms and an active paradigm both in unimpaired subjects and in patients with severe TBI. The aims were: (1) to test the effects on LLAEPs of deviant stimuli with different relevance for the subject; (2) to detect changes in patients and to correlate them with the most important clinical and neuroradiologic features; and (3) to define the prognostic value of LLAEPs for late functional outcome and to verify whether different stimulus presentations may influence their capability to predict outcome.

Back to Article Outline

Methods 

We investigated 11 healthy volunteers (6 men, 5 women) between the ages of 22 and 58 years (mean age, 32 ± 11yr). They had no history of neurologic disease and none had ever undergone electrophysiologic studies. Only 2 were university graduates.

Twenty-one patients (19 men, 2 women) in a vegetative state or coma from severe TBI were also studied. The patients, who were referred from June 1996 to June 1997 to our inpatient intensive rehabilitation unit of primary care in a university-based system, were sequentially recruited for the study. Their mean age was 27.6 ± 12 years (age range, 16-57yr). Patients were excluded if they had experienced neurologic deficits before the acute event. The mean Glasgow Coma Scale (GCS) was 8.2 ± 1.2 (range, 6-10), whereas the mean period from the time of trauma was 54 ± 26 days (range, 16-100d). No patient was receiving sedatives at the time of recording and only 3 had been treated with carbamazepine for posttraumatic seizures. A subgroup of 10 patients who emerged from coma and became collaborative was examined a second time 1 year after the trauma.

The GCS,²¹ Innsbruck Coma Scale (ICS),²² Glasgow Outcome Scale (GOS),²³ Disability Rating Scale (DRS),²⁴ and FIM instrument²⁵ were recorded at entry and at 6 and 12 months posttrauma. The patients who emerged from coma were also examined with the Neurobehavioural Rating Scale (NBHRS)²⁶ at 12 months posttrauma and the 4 factor scores of the scale were separately studied.

Nuclear magnetic resonance (NMR) and single-photon emission tomography (SPET) of the brain were performed in all patients 3 months posttrauma. The level of hydrocephalus and the severity of the lesions and of hypoperfusion in each lobe and brainstem were classified with a score from 0 (normal) to 4 (severe).

The study was approved by the hospital ethics committee.

Brainstem auditory-evoked potentials 

All control subjects and patients underwent a basal recording of brainstem auditory evoked potentials to verify the integrity of the peripheral and brainstem auditory pathways. Recordings were performed with a Nicolet Viking II.^a The stimulus was applied to 1 ear at a time with rarefaction clicks of 60dB sound pressure level (SPL) in controls and 90dB in patients at 10.3Hz frequency. Two trials of 2000 stimulus repetitions were performed and then superimposed. Recordings were performed from the mastoids (M1, M2) and referred to Cz with a filter bandpass of 100 to 2000Hz. Analysis time was 10ms.

LLAEPs 

LLAEPs were recorded from electrodes attached to the scalp at Fz, Cz, and Pz, according to the international 10/20 system with the reference point at M1 and M2. Electrode impedance was kept at less than 3kΩ. The electroencephalogram signal was amplified and filtered with an analog recording bandpass of 0.2 to 30Hz. The analysis epoch was 1000ms and included 300ms of prestimulus delay.

All subjects underwent 3 experimental conditions in a passive oddball paradigm. For each condition, 98 stimuli were delivered, 28 of which were deviant stimuli. In the first experiment, the deviant stimulus consisted of a 2000-Hz tone; in the second, a word that had no specific relevance to the subject (an object); and in the third, the subject's first name. Both the word without relevance and the subject's name were recorded from the voice of a relative with a custom-made system based on a multimedia personal computer.^b The voice was directly digitized into the computer by microphone and a specifically designed application program automatically produced the entire stimuli sequence, interspersing deviant stimuli between neutral ones (1000-Hz tones) in a pseudorandom way and with a 20% occurrence probability. The sequence was recorded by using a high fidelity tape recorder and was reproduced through earphones. In all control subjects and in the 10 patients who recovered from the vegetative state and became collaborative, a fourth experiment was performed based on an active oddball paradigm. The subjects were asked to consider the deviant tones as “targets” and to keep a mental count of them.

All control subjects performed all experiments 3 times at 1-week intervals to verify the reproducibility of the tests.

Data analysis 

The prestimulus delay was used as a baseline for amplitude measurements of the evoked responses. The latency was calculated from the stimulus onset to maximal peak of the wave. In agreement with Garcìa-Larrea et al,¹¹ we measured the latency and amplitude of N100 (maximal negativity in the 70-120ms interval after the stimulus), P150 (first positivity after N100 between 120-200ms), P250 (positive peak between 220-280ms), and P300 (positive peak of greatest amplitude between 280-400ms).

The waves were then graded as follows: (1) normal latency; (2) latency longer than 3 standard deviations (SDs) above the mean values of the controls; and (3) absence of the response.

Statistical analysis 

Statistical analysis was performed by using StatView SE+.^c Repeated-measures analysis of variance (ANOVA) was used to test the effect of stimulus type and repetition and the effect of the electrode sites on latencies and amplitudes. The correlations between electrophysiologic measures and clinical and radiologic variables were computed with the Spearman rank nonparametric test. To verify the reproducibility of the tests, we calculated the correlation coefficient between the first and second trials. The difference between percentages was calculated with the Fisher exact test 2 × 2 contingency table. Mean values were compared by using Student's unpaired t test. By means of cross-tables,²⁷ we computed sensitivity, specificity, and positive and negative predictive values for the relevant neurophysiologic parameters. Findings were considered significant if p values were equal to or lower than .05.

Back to Article Outline

Results 

Representative traces obtained from 1 control subject are shown in figure 1.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  Example of LLAEPs recorded in 1 control subject with all stimulus types at the 3 recording sites.

The sequence of negative and positive components N100, P150, P250, and P300 is evident. The latencies were unchanged at the different scalp sites. The amplitudes were largest at the midline; N100, P150, and P250 were maximal at Cz and P300 parietally. The differences between the amplitudes of the waves at the different scalp sites were statistically significant for each wave (ANOVA, p < .0001) both in controls and in patients, except for P300 recorded with passive paradigms, which showed a relatively constant amplitude at each scalp site.

N100 and P150 

In all control subjects for each experimental condition, we recorded a first negative wave with a mean latency of 98.7 ± 9.9ms (range, 85-100ms), called N100, and a following positive wave with a mean latency of 173.8 ± 7.6ms (range, 154-192ms), called P150. The latency and the amplitude of N100 and P150 were not influenced by the experimental paradigm, by the repetition of the test, or by the age and gender of the subject. N100 and P150 were recorded in 100% of the subjects with the active paradigm. With passive paradigms, N100 was recorded in 100% and P150 in 82% when the deviant stimulus was the subject's name, and in a slightly lower percentage of patients with the other 2 stimuli (fig 2, left panel).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 2. 

  Percentage of occurrence and mean values of latencies and amplitudes of N100 and P150 with all stimulus types in controls and in patients.

However, the difference between the percentages of response to the different stimulus types, calculated with the Fisher exact test 2 × 2 contingency table, was not statistically significant.

The mean latencies of N100 and P150 (fig 2, middle panels) were longer in patients than in control subjects in all experimental conditions, but the differences—compared by using Student's unpaired t test—were statistically significant only with passive paradigms (table 1). N100 mean amplitude was significantly lower in each group of patients than in the controls (fig 2, upper right panel) in all experimental conditions (table 1). A trend was found with passive paradigms toward a higher amplitude of the wave when the subject's name was the deviant stimulus (ANOVA, p < .03). With the passive paradigm the mean amplitude of P150 was significantly lower in each group of patients than in the controls (fig 2, lower right panel), whereas no difference was found with the active paradigm (table 1).

Table 1: Mean Latencies and Amplitudes in Controls and in Patients

NOTE. Values presented as mean latencies and amplitudes ± SD.

Abbreviation: NS, not significant.

Table 2 shows the correlations, computed with Spearman's rank nonparametric test, between N100 and P150 latencies and the most important clinical and neuroradiologic variables. Clinical and functional status of the patient, measured by GCS, ICS, and the functional rating scales at the time of entry into the study, showed no correlation with the degree of abnormality of either N100 or P150 with all stimulus types. However, N100 latency was significantly correlated with the duration of coma—a most important clinical parameter of the severity of the trauma—and with the degree of cortical atrophy and hydrocephalus and with the severity of the lesions in the thalamus as evaluated by NMR. It is evident that the N100 latency showed significantly higher correlations with the patients' outcomes, measured by functional rating scales and NBHRS, when the deviant stimulus was the subject's first name. The P150 latency did not show any significant correlation (table 1).

Table 2: Correlations Between LLAEP Latency and Clinical and Neuroradiologic Variables

The probability of occurrence and both the latency and amplitude of N100 and P150 showed a relation to the outcome of the patients 1 year posttrauma. Figure 3 shows the percentages of occurrence (left panel), as well as the mean values of latency (middle panel) and amplitude (right panel) of N100 and P150 in patients classified in 3 groups based on GOS scores measured 1 year posttrauma.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 3. 

  Percentage of occurrence and mean values of latencies and amplitudes of N100 and P150 with passive paradigms in controls and in patients. Patients are classified in 3 groups based on the GOS score measured 1 year posttrauma.

It is evident that the waves were recorded in all patients with good outcome (GOS score, 4-5), with a mean latency similar to that of the control subjects and significantly shorter than patients with unfavorable outcomes (table 3).

Table 3: Mean Latencies and Amplitudes in Controls and in Patients Classified on the GOS Score

NOTE. Values presented as mean latencies and amplitudes ± SD.

All groups of patients showed a significantly lower amplitude of N100 and P150 than did controls; however, the patients with favorable outcomes showed higher values than patients with unfavorable outcomes. It is also evident from figure 3 (right panel) that the mean amplitude of both N100 and P150 in patients with unfavorable outcomes was significantly higher when the deviant stimulus was the subject's first name (ANOVA, p < .02). The relation between N100 and P150 and outcome was also studied in terms of the tests' sensitivity and specificity (table 4).

Table 4: Specificity, Sensitivity, and Positive and Negative Predictive Value of LLAEPs

We performed cross-table analysis between N100 and P150 grades 1 and 2,3 versus DRS grades 0 to 11 and 12 to 30. This analysis showed high sensitivity and low specificity of both N100 and P150 for all stimulus types; hence, these waves had a high negative predictive value and a low positive predictive value.

P250 and P300 

In all control subjects with the active paradigm, we recorded a large positive wave with a mean latency of 339 ± 31ms (range, 300-384ms), called P300. In 63% of these subjects, this wave was preceded by another positive wave with a mean latency of 48 ± 22ms (range, 215-277ms), called P250. As shown in figure 4, the percentage of occurrence of P300 with the passive paradigms was small.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 4. 

  Percentage of occurrence of P250 and P300 with all stimulus types in controls and in patients.

In the other subjects, we recorded a positivity corresponding to P250. Also, with the patients the P300 percentage of occurrence (fig 4) was high with the active paradigm whereas it was small with all passive paradigms. In another small percentage of patients we recorded a positive wave corresponding to P250. The percentage of occurrence of both P300 and P250 was not statistically different between patients and controls with passive and active paradigms.

Latency and amplitude of P300 and P250 of both controls and patients greatly changed between subjects (ANOVA, p < .001) with all passive and active paradigms, but they were not influenced by age or gender. In controls with passive paradigms, the P300 amplitude significantly decreased when the test was repeated and when the deviant stimulus was the name of the object (ANOVA, p < .002) or the deviant tones (ANOVA, p < .03). It was unchanged, however, when the deviant stimulus was the subject's first name. The P300 mean latency was longer in patients than in controls, but the difference was statistically significant only when the deviant stimulus was the subject's first name (table 1). However, P300 mean latencies showed no statistically significant differences between the stimulus types. The P300 mean amplitude was significantly lower in patients than in controls when the deviant stimulus was the name of the object or the deviant tones, but it was not statistically different when the deviant stimulus was the subject's first name, or with the active paradigm (table 1).

All patients in whom P250 or P300 were recorded with a passive paradigm awoke from their comas, but no significant correlations were found with the functional rating scales measured 1 year posttrauma. Cross-table analysis (P250 and P300 grade 1 and 2,3 vs DRS grades 0-11 and 12-30), however, showed a high specificity and low sensitivity of P300 (table 2); hence, it showed a considerable negative predictive value but no positive predictive value. Moreover, both P250 and P300 scores correlated inversely to the level of impairment of consciousness at the time of recording as measured by GCS (p < .03) and ICS (p < .003), but only when the deviant stimulus was the subject's first name. Because of the great variability of these waves and the low occurrence, we did not consider them for further correlations.

Back to Article Outline

Discussion 

The first finding of this study is that LLAEPs can be recorded in patients with severe impairment of consciousness with a passive paradigm, and that a stimulus that is relevant for the subject, such as his/her own name, can enhance the accuracy of the test and its relation to functional outcome. In control subjects the N100-P150 complex is not influenced by an active or passive paradigm or by the repetition of the test, hence, it seems the most reliable LLAEP of those we studied. In our patients, N100 showed a delayed latency and a decreased amplitude with respect to awake control subjects. This result is similar to that previously reported in sleeping subjects.²⁸, ²⁹ It suggests that external sensory stimuli is computed also when the level of consciousness is severely impaired, but that the process is delayed. Previous studies³⁰, ³¹, ³² have shown that N100 is greatly influenced by the alerting state.³⁰, ³¹

Our study found that a stimulus relevant for the subject can enhance the N100-P150 complex. These data have not been previously reported. In our patients, in fact, N100 was recorded in a higher percentage of cases and with an increased amplitude when the deviant stimulus was the subject's first name, indicating that this type of stimulus can also enhance the level of alertness when consciousness is severely impaired. The lack of correlation between the N100-P150 complex and the level of impairment of consciousness measured by GCS and ICS at the time of entry suggests that this complex may be the expression of a nonspecific state of alertness evoked by the stimulus mediated by reticular and thalamic inputs, which influence the primary auditory cortex. This hypothesis is supported by the significant correlation found between the N100-P150 complex and the severity of damage to the thalamus as evaluated by NMR. We also found that a relevant deviant stimulus enhances the predictability of the N100 wave on functional outcome, as measured by the most important functional rating scales. Significant relationships were also found between N100 and cognitive and behavioral outcome as evaluated by the NBHRS, particularly with scale items that explore memory and frontal functions. Modifications of LLAEPs have been reported in patients with psychiatric disorders³³ and an effect of serotoninergic neurotransmission has been postulated. It has been proposed, in fact, that the N100-P150 complex may be modulated by cortical serotoninergic innervation³⁴ and that this system could be well qualified for adjusting individual levels of sensory processing, especially in the primary auditory cortex in which the N100-P150 component is mainly generated. In patients who recovered from coma, the N100-P150 complex with active paradigms showed a latency similar to control subjects whereas the amplitude of N100—but not that of P150—remained significantly lower than in controls. This data could be explained by the impairment of attention that is often evident in the sequelae of TBI.³⁵

The waves after P150 have shown a great variability within both controls and patients and have been recorded in only a few subjects with the passive paradigm. As reported by others,¹⁶, ³⁶, ³⁷, ³⁸ we found that P300 in controls is greatly influenced by the paradigm used; it is, in fact, more easily evoked with larger amplitude with an active paradigm and decreases in amplitude with repetition. However, we found that the repetition effect is less evident when the deviant stimulus is relevant for the subject. Moreover, we found that the use of a stimulus relevant for the subject in a passive paradigm enhances the probability of recording P300 both in controls and in patients. In a small number of controls and patients, P300 with an active paradigm was preceded by a positivity called P250; this positivity was also recorded in some controls and in some patients without the following P300. According to Garcìa-Larrea et al¹¹ P250 could represent a preliminary step before a P300 can be elicited; it could reflect the identification of the stimulus as a target whereas P300 could be related to a cognitive process of recognition. In our unconscious patients, we found a straight correlation between the scores of these P300 and P250 waves and the level of impairment of consciousness as measured by GCS and ICS at the time of the examination. These data agree with those data from previous studies.⁵, ⁶, ³⁹ We hypothesize that these waves could be associated with an automatic orienting of attention toward the stimulus and not just with a nonspecific level of alertness, as is true for the N100-P150 complex. P300 has been correlated in previous studies with recovery from traumatic coma⁵, ⁶, ⁷; also in our study, the presence of P300 related to a good prognosis but its absence did not preclude such a prognosis. The limitations of P300, however, must be considered. In agreement with Barnes,⁴⁰ we found that it is more easily elicited in attentive subjects and may be absent in control individuals. Hence, we think that it must be considered cautiously as a prognostic index in unconscious patients.

In our opinion, LLAEPs may be more useful than short latency potentials in the evaluation of patients in the postacute phase after severe TBI when consciousness is severely impaired. Short latency-evoked potentials are predominantly of sensory origin, are generated close to the primary projection area, and are related to the physical nature of the stimulus. Long-latency components are generated by subcortical/cortical and cortico/cortical circuits, modulated by the ascending reticular activating system.¹⁰ Therefore, their presence depends on the integrity of a more extensive network of connections than do short-latency responses.⁴¹ Moreover, early components are often normalized in the postacute phase of TBI,⁴² probably because as the repair of secondary damage related to brain edema or conduction blocks.

Back to Article Outline

Conclusions 

The presence of LLAEPs and the influence exerted by a relevant stimulus seems to suggest that at least rudimentary processing or discrimination of auditory stimuli occurs in unconscious patients. The use of a stimulus that is relevant to the subject may activate some mechanism of automatic sensory analysis and memory trace of auditory features, even in patients with severe impairment of consciousness. The method that we applied is simple, noninvasive, relatively inexpensive, and appropriate for the monitoring of comatose or minimally responsive patients. However, further studies are warranted to verify our data so that other passive paradigms can be devised to enhance the sensibility of the test in unconscious patients.

In conclusion, LLAEPs may extend the battery of neurophysiologic tests currently available for determining the functional integrity of the central nervous system and the capacity of cognition in patients with severe impairment of consciousness. Use of stimuli relevant for the patient can enhance the probability of recording these waves and their correlations to outcome. We suggest that LLAEPs, particularly the N100-P150 complex, be included in the neurophysiologic monitoring of minimally responsive patients affected by severe TBI.

Back to Article Outline

Acknowledgements 

The authors thank Professor Marco Schieppati for his helpful suggestions and critical reading and Rosemary Allpress for revising the English text.

Back to Article Outline

Suppliers 

a. Nicolet Biomedical Inc, 5225 Verona Rd, Bldg 2, Madison, WI 53711-4495.

b. Macintosh LC, MasOS 7.0.1; Apple Computer, Inc, One Infinite Loop, Cupertino, CA 95014.

c. SAS Institute, Inc, SAS Campus Dr, Cary, NC 27513.

Back to Article Outline

References 

1. Anderson DC, Bundlie S, Rockswold GL. Multimodality evoked potentials in closed head trauma. Arch Neurol. 1984;41:369–374
   • View In Article
   • MEDLINE
2. Cant BR, Hume AL, Shaw NA. The assessment of severe head injury by short-latency somatosensory and brain-stem auditory evoked potentials. Electroencephalogr Clin Neurophysiol. 1986;65:188–195
   • View In Article
   • MEDLINE
3. Rumpl E, Prugger M, Gerstenbrand F, Brunhaber W, Badry F, Hacke JM. Central somatosensory conduction time and acoustic brainstem transmission time in post-traumatic coma. J Clin Neurophysiol. 1988;5:237–260
   • View In Article
   • MEDLINE
   • CrossRef
4. Chatrian GE, Bergamasco B, Bricolo A, Frost JD, Prior PF. IFCN recommended standards for electrophysiologic monitoring in comatose and other unresponsive states. Report of an IFCN committee. Electroencephalogr Clin Neurophysiol. 1996;99:103–122
   • View In Article
   • Full-Text PDF (2188 KB)
   • CrossRef
5. Kane NM, Curry SH, Rowlands CA, Manara AR, Lewis T, Moss T, et al.  Event-related potentials neurophysiological tools for predicting emergence and early outcome from traumatic coma. Intensive Care Med. 1996;22:39–46
   • View In Article
   • MEDLINE
   • CrossRef
6. Yingling CD, Hosobuchi Y, Harrington M. P300 as a predictor of recovery from coma. Lancet. 1990;336:873
   • View In Article
   • CrossRef
7. Doebrich HM, Nau HE, Lodeman E, Zerbin D, Schmit-Neuerburg KP. Evoked potentials assessment of mental function during recovery from severe head injury. Surg Neurol. 1986;26:112–118
   • View In Article
   • MEDLINE
   • CrossRef
8. Unsal A, Segalowitz SJ. Sources of P300 attenuation after head injury: single-trial amplitude, latency jitter and EEG power. Psychophysiology. 1995;32:249–256
   • View In Article
   • MEDLINE
   • CrossRef
9. Rugg MD, Pickles CD, Potter DD, Doyle MC, Pentland B, Roberts RC. Cognitive brain potentials in a three-stimulus auditory “oddball” task after closed head injury. Neuropsychologia. 1993;31:373–393
   • View In Article
   • MEDLINE
   • CrossRef
10. Greenberg RP, Ducker TB. Evoked potentials in the clinical neurosciences. J Neurosurg. 1982;56:1–18
    • View In Article
    • MEDLINE
    • CrossRef
11. Garcìa-Larrea L, Lukaszewicz AC, Mauguière F. Revisiting the oddball paradigm, non-target vs neutral stimuli and the evaluation of ERP attentional effects. Neuropsychologia. 1992;30:723–741
    • View In Article
    • MEDLINE
    • CrossRef
12. Gooding DS, Squires KC, Starr A. Long latency event related components of the auditory evoked potentials in dementia. Brain. 1978;101:635–648
    • View In Article
    • MEDLINE
13. Pfefferbaum A, Wenegrat BG, Ford JM, Roth WT, Kopell BS. Clinical application of the P3 component of event-related potentials. II. Dementia, depression, schizophrenia. Electroencephalogr Clin Neurophysiol. 1984;59:104–124
    • View In Article
    • MEDLINE
14. Pfefferbaum A, Ford JM. Clinical utility of long-latency cognitive event-related potentials (P3): the cons. Electroencephalogr Clin Neurophysiol. 1990;76:6–12
    • View In Article
    • Full-Text PDF (540 KB)
    • CrossRef
15. Starr A, Don M. Brain potentials evoked by acoustic stimuli. In:  Picton TW editors. Human event-related potentials EEG handbook rev series. Vol. 3:Amsterdam: Elsevier; 1988;p. 97–157
    • View In Article
16. Roth WT. Auditory evoked responses to unpredictable stimuli. Psychophysiology. 1973;10:125–138
    • View In Article
    • MEDLINE
    • CrossRef
17. Picton TW, Hillyard SA, Krausz HI, Galambos R. Human auditory evoked potentials. I. Evaluation of components. Electroencephalogr Clin Neurophysiol. 1974;36:179–190
    • View In Article
    • Abstract
    • Full-Text PDF (1236 KB)
    • CrossRef
18. Polich J. P300 from a passive auditory paradigm. Electroencephalogr Clin Neurophysiol. 1989;74:312–320
    • View In Article
    • MEDLINE
19. Farwell LA, Donchin E. The truth will out: interrogative polygraphy (“lie detection”) with event-related potentials. Psychophysiology. 1991;28:531–546
    • View In Article
    • MEDLINE
    • CrossRef
20. Berlad I, Pratt H. P300 in response to the subject's own name. Electroencephalogr Clin Neurophysiol. 1995;96:472–474
    • View In Article
    • MEDLINE
21. Teasdale G, Jennett B. Assessment of coma and impaired consciousness: a practical scale. Lancet. 1974;2:814
    • View In Article
    • MEDLINE
22. Benzer A, Mitterschiffthaler G, Marosi M. Prediction of non-survival after trauma: Innsbruck coma scale. Lancet. 1991;338:977–978
    • View In Article
    • Abstract
    • CrossRef
23. Jennett WB, Bond MR. Assessment of outcome after severe brain damage. Lancet. 1975;i:480–484
    • View In Article
24. Rappaport M, Hall KM, Hopkins HK, Belleza T, Cope DN. Disability rating scale for severe head trauma: coma to community. Arch Phys Med Rehabil. 1982;63:118–123
    • View In Article
    • MEDLINE
25. Granger CV, Hamilton BB, Keith RA, Zielezny M, Sherwin FS. Advances in functional assessment for medical rehabilitation. Top Geriatr Rehabil. 1986;1:59–74
    • View In Article
26. Levin HS, High WM, Goethe KE, Sisson RA, Overall JE, Rhoades HM, et al.  The neurobehavioural rating scale: assessment of the behavioural sequelae of head injury by the clinician. J Neurol Neurosurg Psychiatry. 1987;50:183–193
    • View In Article
    • MEDLINE
    • CrossRef
27. Altman DG. Practical statistics for medical research. London: Chapman and Hall; 1990;
    • View In Article
28. Nordby H, Hugdahl K, Stickgold R, Bronnick KS, Hobson JA. Event-related potentials (ERPs) to deviant auditory stimuli during sleep and waking. Neuroreport. 1996;7:1082–1086
    • View In Article
    • MEDLINE
    • CrossRef
29. Bastuji H, Garcìa-Larrea L, Franc C, Mauguière F. Brain processing of stimulus deviance during slow-wave and paradoxical sleep: a study of human auditory evoked responses using the oddball paradigm. J Clin Neurophysiol. 1995;12:155–167
    • View In Article
    • MEDLINE
    • CrossRef
30. Näätänen R, Picton TW. The N1 of the human electric and magnetic response to sound. A review and an analysis of the component structure. Psychophysiology. 1987;24:375–425
    • View In Article
    • MEDLINE
    • CrossRef
31. Fruhstorfer H, Bergström RM. Human vigilance and auditory evoked responses. Electroencephalogr Clin Neurophysiol. 1969;27:346–355
    • View In Article
    • Abstract
    • Full-Text PDF (747 KB)
    • CrossRef
32. Kevanishvili Z, von Spetch H. Human slow auditory evoked potentials during natural and drug-induced sleep. Electroencephalogr Clin Neurophysiol. 1979;47:280–288
    • View In Article
    • Abstract
    • Full-Text PDF (596 KB)
    • CrossRef
33. Boutros N, Nasrall H, Leighty R, Torello M, Tueting P, Olson S. Auditory evoked potentials, clinical vs. research applications. Psychiatry Res. 1997;69:183–195
    • View In Article
    • Abstract
    • Full-Text PDF (1322 KB)
    • CrossRef
34. Hegerl U, Juckel G. Intensity dependence of auditory evoked potentials as an indicator of central serotoninergic neurotransmission: a new hypothesis. Biol Psychiatry. 1993;33:173–187
    • View In Article
    • Abstract
    • Full-Text PDF (1719 KB)
    • CrossRef
35. Levin HS, Goldstein FC. Neurobehavioral aspects of traumatic brain injury. In: Bach-y-Rita, editor. Traumatic brain injury. Vol 2. New York: Demos: Comprehensive Neurologic Rehabilitation; 53-72.
    • View In Article
36. Knight RT. Decreased response to novel stimuli after prefrontal lesions in man. Electroencephalogr Clin Neurophysiol. 1984;59:9–20
    • View In Article
    • MEDLINE
37. Woods DL, Hillyard SA, Courchesne E, Galambos R. Electrophysiological signs of split-second decision-making. Science. 1980;207:655–657
    • View In Article
    • MEDLINE
38. Polich J. Task difficulty, probability and inter-stimulus interval as determinants of P300 from auditory stimuli. Electroencephalogr Clin Neurophysiol. 1987;63:311–320
    • View In Article
39. Guérit JM, Verougstraete D, de Tourtchaninoff M, Debatisse D, Witdoechkt C. ERPs obtained with auditory oddball paradigm in coma and altered states of consciousness: clinical relationship, prognostic value, and origin of components. Clin Neurophysiol. 1999;110:1260–1269
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (385 KB)
    • CrossRef
40. Barnes MP. Outcome of head injury. Curr Opin Neurol. 1991;4:12–16
    • View In Article
41. Pfurtscheller G, Schwarz G, Gravenstein N. Clinical relevance of long-latency SEPs and VEPs during coma and emergence from coma. Electroencephalogr Clin Neurophysiol. 1985;62:88–98
    • View In Article
    • MEDLINE
42. Mazzini L, Pisano F, Zaccala M, Miscio G, Gareri F, Galante M. Somatosensory and motor evoked potentials at different stages of recovery from severe traumatic brain injury. Arch Phys Med Rehabil. 1999;80:33–39
    • View In Article
    • Abstract
    • Full-Text PDF (879 KB)
    • CrossRef