Elsevier

Clinical Neurophysiology

Volume 110, Issue 9, 1 September 1999, Pages 1601-1610
Clinical Neurophysiology

Mismatch negativity and late auditory evoked potentials in comatose patients

https://doi.org/10.1016/S1388-2457(99)00131-5Get rights and content

Abstract

Objectives: The purpose of this study was to assess the patterns of mismatch negativity (MMN) and N100 component in comatose patients and to evaluate their prognostic value vis-à-vis return of consciousness.

Methods: MMN and auditory (early, middle-latency and late) evoked potentials were recorded in 52 normals and in 128 comatose patients (comas due to neurosurgical and neurological problems). At the time of recording, all patients scored lower than 8 on the Glasgow scale.

Results: Visually detected N100 and MMN were confirmed by cross-correlation of sub-averages. The MMN was present in 33/128 patients and the N100 component in 84/128. The amplitudes of MMN and N100 waves detected in comatose patients were statistically different from those of normal subjects. By 3 months after the onset of coma, 95 patients had returned to consciousness, most of them with moderate to severe disability. A ratio of 30/33 patients with MMN and 70/84 with N100 had regained consciousness. The presence of a MMN together with a N100 component was more specific (90.9%) than the presence of a N100 component irrespective of MMN (57.6%) in terms of predicting return to consciousness, but its sensitivity was lower (respectively 31.6% for MMN and 73.7% for N100). The mean period that elapsed between the recording of evoked potentials and a return to consciousness was 6.3±4 days. MLAEPs were also highly specific, but BAEPs were not.

Conclusion: MMN and auditory evoked potentials provide a reliable assessment of the functional status of comatose patients. When present, MMN and the N100 differ from those found in normal subjects in terms of latencies and amplitudes. As a predictor of return of consciousness MMN had high specificity and low sensitivity, whereas the N100 had high sensitivity and low specificity. This study demonstrates that the recording of MMN and the auditory N100 can be a very useful aid in the assessment of coma and in predicting whether or not a patient will regain consciousness.

Introduction

Over the last 20 years, interest has been growing in the assessment of coma and in predictors of awakening and recovery. Certain clinical (e.g. the Glasgow Coma scale) and physiological (e.g. the Saps II scale) scales are widely used but measurements of evoked potentials less so. The recording of evoked potentials is a non-invasive technique which can be performed as a complement to a physical examination and which has been shown to be of great value in the evaluation of coma. Visual (Greenberg et al., 1981), somatosensory (Greenberg et al., 1981; Cant et al., 1986; Rothstein et al., 1991; Chen et al., 1996), and motor (Ying et al., 1992; Zentner and Rohde, 1992) evoked potentials have all been used either on their own or in combination. With respect to auditory evoked potentials, early responses (BAEPs) which only explore the brainstem have been used more often (Greenberg et al., 1981; Cant et al., 1986) than middle-latency responses (MLAEPs) (Litscher, 1995). Only one study deals with early, middle-latency and late auditory potentials together: in this relatively early study of 25 patients, it was shown that survival correlated with the simultaneous presence of long, middle-latency and brain stem components and that sustained middle-latency and/or brain stem components alone were not predictive of survival (Rosenberg et al., 1984).

It is only recently that cognitive components have been studied in comatose subjects. The most studied is the P300 component (Gott et al., 1991; De Giorgio et al., 1993; Signorino et al., 1995) but mismatch negativity (MMN) has only been recorded by one investigator (Kane et al., 1993; Kane et al., 1996). MMN was first described by Näätänen in 1978 and is a component of event related potentials generated by any perceptible change in an auditory stimulus; it can be used as an indicator of the status of pre-attentive sensory memory (Näätänen et al., 1978; Näätänen and Michie, 1979; Näätänen, 1995). As shown both in patients with electrodes implanted in the brain (Kropotov et al., 1995) and by bipolar modeling (Giard et al., 1995; Frodl-Bauch et al., 1997), this component seems to be exclusively or preferentially generated in the auditory area of the temporal cortex. As it is an index of the functional state of the cortex, and strongly attentive-independent (Näätänen, 1995), MMN might be recorded in the prognosis of comatose patients.

The aim of this study was to determine MMN and N100 component patterns in comatose patients and to evaluate their predictive value vis-à-vis the patient regaining consciousness.

Section snippets

Patients

The sample consisted of 128 patients (including 54 females and 74 males) aged between 15 and 93 years (mean=51.1±16.9). All were in Intensive Care in hospital and were comatose with a Glasgow score less than 8 at the time of the recording of evoked potentials. All patients had undergone at least one brain CAT scan or MRI. The causes of coma were head injury (34 cases), cardiac and respiratory failure (19), ischemic or hemorrhagic stroke syndrome (41), complications of neurosurgery (31) and

MMN and N100 in normal subjects

The N100 component, N100 deviant and MMN were all detected in 52 normal subjects. In 45% of them, the MMN had more than one peak. Mean peak latencies and mean peak amplitudes are presented in Table 1 together with the measurements made in comatose patients.

MMN and N100 in comatose patients

The MMN was observed in 33/128 patients and was absent in the other 95. The N100 component was observed in 84/128 patients and was absent in the other 44. MMN was never observed without either the N100 component or N100 deviant and the N100

Discussion

The mismatch negativity can be recorded in comatose subjects and, in the population in this study consisting of comatose patients with Glasgow scores of less than 8, it was detected in 33 out of 128 cases (it was absent in the other 95 cases). It is unlikely that failure to detect MMN could reflect inadequate stimulation or recording because the same techniques were used in the normal population in all 52 of whom MMN, the N100 component and the N100 deviant were detected. Our technique of

Acknowledgements

This work has been supported by EU grant BMH4-CT96-0819 (Cobrain) and Fondation de l'Avenir.

References (24)

  • C.M. De Giorgio et al.

    Predictive value of P300 event-related potentials compared with EEG and somatosensory evoked potentials in non-traumatic coma

    Acta Neurol Scand

    (1993)
  • C. Fischer et al.

    Auditory early and middle-latency evoked potentials in patients with quadrigeminal plate tumors

    Neurosurgery

    (1994)
  • Cited by (214)

    • Mismatch Negativity and Theta Oscillations Evoked by Auditory Deviance in Early Schizophrenia

      2023, Biological Psychiatry: Cognitive Neuroscience and Neuroimaging
    • Neural responses to sensory novelty with and without conscious access

      2022, NeuroImage
      Citation Excerpt :

      Auditory stimuli that violate the predictions of the central auditory system elicit a MMN response at around 200 ms after odd stimulus onset. This event is independent of attentional states, and presumably reflects early sensory memory and automatic error-detection mechanisms involved in the identification of mismatching sensory representations (Bekinschtein et al., 2009; Fischer et al., 1999; Garrido et al., 2009; Näätänen et al., 2007, 2011; Tiitinen et al., 1994). Novel auditory events that are attended to and consciously accessed additionally elicit a P3 response, an event that has been proposed to reflect attentional reorientation, revision of stimulus-driven mental representations and memory-dependent cognitive operations (Comerchero and Polich, 1999; Polich, 1987, 2007; Sutton et al., 1965).

    View all citing articles on Scopus
    View full text