The temporal relationship between the brainstem and primary cortical auditory evoked potentials

Progress in Neurobiology Vol. 47, pp. 95 to 103, 1995 Copyright 0 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0301-0082/95/$29.00

Pergamon

THE TEMPORAL RELATIONSHIP BETWEEN THE BRAINSTEM AND PRIMARY CORTICAL AUDITORY EVOKED POTENTIALS N. A. SHAW* Departmnt

of Physiology, School of Medicine. University of Auckland, Private Bag Auckland I, New Zealand (Received 9 May 1995)

Abstract-h!lany methods are employed in order to define more precisely the generators of an evoked potential (El’) waveform. One technique is to compare the timing of an EP whose origin is well established with that of one whose origin is less certain. In the present article, the latency of the primary cortical auditory evoked potential (PCAEP) was compared to each of the seven subcomponents which compose the brainstem auditory evoked potential (BAEP). The data for this comparison was derived from a retrospective: analysis of previous recordings of the PCAEP and BAEP. Central auditory conduction time (CACT) was calculated by subtracting the latency of the cochlear nucleus BAEP component (wave III) from that of the PCAEP. It was found that CACT in humans is 12 msec which is more than double that of central somatosensory conduction time. The interpeak latencies between BAEP waves V, VI, and VII and the PCAEP were also calculated. It was deduced that all three waves must have an origin rather more caudally within the central auditory system than is commonly supposed. In addition, it is demonstrated that the early components of the middle latency AEP (No and Na) largely reside within the time domain between the termination of the BAEP components and the PCAEP which would be consistent with their being far field reflections of midbrain and subcortical auditory activity. It is concluded that as the afferent volley ascends the central auditory pathways, it generates not a sequence of high frequency BAEP responses but rather a succession of slower post-synaptic waves. The only means of reconciling the timing of the BAEP waves with that of the PCAEP is to assume that the generation of all the BAEP components must be largely restricted to a quite confined region within the auditory nerve and the lower half of the pons.

CONTENTS 95 97 98 99 100 101 101 102

1. Introduction 2. Animal studies 3. Central auditory conduction time in man 4. The origins of the late BAEP waves 5. The origins of the early MAEP waves 6. Conclusions Acknowledgements References

ABBREVIATIONS AEP BAEP CACT EP IPL LAEP

MAEP MGB Pl PCAEP SNlO SP6

Auditory evoked potential Brainstem auditory evoked potential Central auditory conduction time Evoked potentials Interpeak latency Long latency auditory evoked potential

1. INTRODUCTION

Middle latency auditory evoked potential Medial geniculate body Primary cortical positivity Primary cortical auditory evoked potential Slow negative 10 Slow positive 6

which can be obtained from the human scalp (Picton er al., 1974). AEPs were divided into three categories depending upon the time domain they occupied. Initially, there were the early or short latency AEPs which arose within the first 10 msec following stimulus onset. Second, there were the middle latency AEPs (MAEPs) which were generated mostly between 10 and 50 msec. MAEPs were followed by the long

Twenty years ago, Picton and his colleagues published their now classic description of the principal types of auditory evoked potentials (AEPs)

*Correspondence: Fax No [64-91373-7499. 95

96

N. A. Shaw

latency AEPs (LAEPs). As a rule, the earlier the AEP, the lower its amplitude, the higher its frequency and the more robust and stable it was. Conversely, the later the AEP, the higher its amplitude, the lower its frequency and the more la bile and sensitive it was to the subject’s state of arousal. The early AEPs are now more commonly called the brainstem AEP (BAEP) or some similar moniker and arise principally within the eighth nerve and the tracts and nuclei of the auditory brainstem. The MAEP was thought to reflect activity mostly in the subcortex and the temporal lobe, while the LAEP was considered to be generated by intracortical processing especially within the frontal association areas. Despite much research, especially during the last decade, the origins of both the MAEP and the LAEP still remain uncertain although there is increasing evidence that activity arising in the superior temporal plane plays a major role in the generation of both waveforms (e.g., Vaughan and Ritter, 1970; Knight et al., 1980; Cohen, 1982; Ozdamar et al., 1982; Wood and Wolpaw, 1982; Kileny et al., 1983; Scherg and Von Cramon, 1985, 1986; Deiber et al., 1988; Knight et al., 1988; Richer et al., 1989). In addition to the BAEP, the MAEP and the LAEP, there exists a fourth type of AEP which cannot be detected with a surface (i.e., scalp) recording. This is the primary cortical AEP (PCAEP) which is the response generated by the arrival of the afferent volley at the primary cortical receiving area. In man, the primary auditory cortex is located in the tranverse temporal gyri (Heschl’s gyri) which lies buried within the sylvian fissure. As a consequence, the PCAEP can only be obtained with an intracranial electrode (Celesia, 1976; Lee et al., 1984; LiegeoisChauvel et al., 1991). As Picton et aI. (1974) confirmed, there appears to be no correlate of the PCAEP in any of the AEPs recorded from the scalp. It is important not to confuse the MAEP with the PCAEP, as is sometimes done (e.g., Hall, 1992). Although the two waveforms do overlap in time and may share some generators, they should be treated as largely separate responses. Of the three scalp-recorded AEPs, most attention has focussed on refining the origins of the BAEP, presumably because of its value in otoneurological investigation (Chiappa, 1990). Although the BAEP was first described a quarter of a century ago, the nature and location of the generators of its waveform are still not definitively understood (Zaaroor and Starr, 1991a). The human BAEP consists of a series of up to seven (five early major plus two late minor) high frequency vertex-positive far field waves which are conventionally labelled with Roman numerals. According to the classical theory, each wave is generated by the sequential activation of successively higher auditory structures. This concept is exemplified by Stockard’s well known illustration, variations of which are widely reproduced (e.g., Stockard et al., 1980; Dorfman, 1983). In accordance with this model, wave I of the BAEP arises in the eighth nerve, wave II in the cochlear nucleus, wave III in the environs of the superior olivary complex, wave IV in the pathway or nucleus of the lateral lemniscus, wave V in the inferior colliculus, wave VI in the

medial geniculate body (MGB), and wave VII in the auditory radiations, presumably just prior to the onset of cortical activity. There have, however, been a number of modifications to this model. For example, it is now widely accepted that because of its length, there is sufficient time for two waves to be generated within the human auditory nerve. Therefore wave I is believed to be generated within the more distal part of the eighth nerve, while wave II arises in the more proximal intracranial portion (Hashimoto et al., 1981; Moller et al., 1981, 1988; Hall, 1992). As a consequence, wave III is now assumed to largely reflect cochlear nucleus activity (Moller and Jannetta, 1983; Zaaroor and Starr, 1991b), while wave IV is thought to arise in or near the superior olivary complex (Wada and Starr, 1983~; Zaaroor and Starr, 1991b; Hall, 1992). There have also been persistent doubts about whether the last of the principal BAEP waves (wave V) actually does arise in the midbrain (Achor and Starr, 1980a, b; Moller and Jannetta, 1982; Wada and Starr, 1983~; Moller and Burgess, 1986; Hall, 1992). An origin no more rostra1 than the terminal part of the lateral lemniscus has been proposed (Moller and Jannetta, 1982). If there is, in fact, a collicular contribution to the BAEP waveform, it may be the trough of negativity which follows the generation of the principal BAEP waves (Hashimoto et al., 1981; Hashimoto, 1982; Moller and Jannetta, 1982). This slow waveform appears to be analogous to the slow negative 10 (SNlO) component of the slow brainstem response described by Davis and Hirsh (1979) and the early No component of the MAEP waveform (Picton et al., 1974). A wide range of techniques have been employed in an attempt to elucidate the origins of the BAEP components. These have included clinico-pathologic studies of patients with focal brainstem lesions (e.g., Stockard and Rossiter, 1977), intracranial recordings of patients undergoing neurosurgical procedures (e.g., Hashimoto et al., 1981), topographic mapping studies (e.g., Starr and Squires, 1982), correlation between depth and surface recordings in animals (e.g., Achor and Starr, 1980a), micro-electrode recordings from animals (e.g., Huang and Buchwald, 1977), experimental lesion and ablation studies in animals (e.g., Achor and Starr, 1980b), and the use of neural inactivating agents (e.g., Zaaroor and Starr, 1991a). In addition, the timing of other AEPs with well-established origins has been compared with that of the BAEP waves. For instance, a comparison of the electrocochleogram with the BAEP helped prove that wave I, at least, reflects the compound action potential in the eighth nerve (Jewett, 1970). Similarly, such a technique might also be employed to help confirm or reject the putative origins of some of the later BAEP waves. This could be achieved by examining the temporal correspondence between them and the PCAEP. Simultaneous recordings of the BAEP and PCAEP have been made in animals such as the rat (Shaw, 1990a) and cat (Chen and Buchwald, 1986). However, in man, such a comparison is hampered by the difficulty in recording the PCAEP referred to previously. Even when the PCAEP has been successfully recorded from humans, the BAEP has not been concurrently obtained, or else

Temporal Relationship Between Brainstem and Cortical Potentials no data actually reported (e.g., Liegeois-Chauvel et al., 1991). Be that as it may, sufficient information presently exists on both the BAEP and PCAEP recorded from man to allow a reliable comparison to be made between the two waveforms, even if the individual recordings were not obtained from the same subject. To investigate the temporal relationship between the PCAEP and the later BAEP waves in man was therefore the main purpose of the present review.

2. ANIMAL STUDIES Unlike man, the PCAEP in the rat can be quite readily obtained with an epidural electrode (Shaw, 1990b). The relative ease with which the PCAEP can be recorded therefo:re facilitates a direct comparison between the PCAE.P and BAEP waveforms in this animal. Examining the temporal correspondence between the PCAEP and BAEP in the rat might reveal insights into the timing of the later BAEP waves equally relevant to man. Figure 1 shows an example of the PCAEP and BAEP waveforms recorded simultaneously from the rat. The interpeak latency (IPL) between each of the BAEP waves and the primary cortical response (Pl) is also indicated. In this regard, it should be noted that animals such as rat, cat and monkey possess one less major BAEP wave than man, i.e. just four rather than five. The reason for this discrepancy is thought to be because the auditory nerve in these animals is only long enough to allow the generation of a single wave (Moller and Burgess, 1986). Therefore, when comparing human BAEPs to animal BAEPs, waves I and II (human) are equal to wave I (animal), wave III (human) is equal to wave II (animal), wave IV

(human) is equal to wave III (animal) and so on (Zaaroor and Starr, 1991a; Hall, 1992). Central auditory conduction time (CACT) is defined as the time it takes for an afferent volley to traverse the central auditory pathways. CACT can be calculated by subtracting the latency of a potential generated at or near the entry point of the auditory nerve into the brainstem from that of the primary cortical response (Pl) of the PCAEP. Clearly the most appropriate BAEP wave to use in the calculation of CACT in the rat is the cochlear nucleus response (wave II). The wave II-PI IPL yields a CACT value of 6.5 msec. This is more than double the central somatosensory conduction time previously calculated for the rat (Shaw and Cant, 1981). The discrepancy in central conduction time between the two sensory systems would seem to be largely accounted for by the additional three or four synaptic delays interposed in the central auditory system. Figure 1 also shows the IPLs between the later BAEP waves and the primary cortical response (Pl). For example, the wave IV-P1 IPL is 4.9 msec. If wave IV genuinely did originate in the inferior colliculus as the standard theory supposes, then it would be difficult to account for such a long conduction allowing that there are only two further synapses to traverse. The comparison of the timing of wave IV and Pl therefore casts further doubt on the likelihood that wave IV (or wave V in man) could have a mesencephalic origin. Similarly, the wave V-P1 IPL was 4.0 msec. If wave V did arise within the MGB, then a thalamo-cortical conduction time of 4 ms would be quite anomalous. Thalamo-cortical conduction time in the somatosensory system of the rat has been previously calculated at less than 1.5 msec (Shaw and Cant, 1981) and there seems no reason to believe that the comparable conduction time in the

BAEP’

;c

7.2

91

I

II

Ill

IV

v

1.3

2.0

2.7

3.6

4.5

31 :/

Pl

(8.5)

PCAIEP 2 M8

Fig. 1. Bralnstem auditory evoked potentials (BAEPs) and primary cortical AEPs (PCAEPs) recorded simultaneously from the rat. The high frequency BAEP components are identified with Roman numerals. Pl represents the primary cortical response. Actual latencies (msec) and interpeak (IPLs) between the BAEP waves and Pl are also indicated. A description of how these waveforms were recorded can be found elsewhere (Shaw, 199Oa).The amplitude and latency scale applies to both waveforms.

N. A. Shaw

98

A

RECORDED

Pl

(8.7)

OVER

THALAMUS

RECORDED INFERIOR

OVER COLLICULUS

1

II

III

IV

v

1.3

2.1

2.0

3.7

4.8

Fig. 2. Auditory waveforms recorded consecutively from different skull locations in the rat. The actual latencies of the PI component of the PCAEP waveform and the high frequency BAEP waves are indicated. Also shown is the latency of the principal positive component of the potentials recorded from over the inferior colliculus and the thalamus. A description of how these waveforms were recorded can be found elsewhere (Shaw, 1993).

auditory system could be much longer. While there has never been any consensus on the origin of wave V (or wave VI in humans), the present analysis suggests that it must be generated somewhat more caudally than the diencephalon. In fact, the only way of satisfactorily reconciling the timing of the BAEP waves as a whole with the PCAEP is to assume that the generation of the entire waveform must be restricted to the eighth nerve and the auditory brainstem. Irrespective of the exact origins of the BAEP waves, it is clear from Fig. 1 that there is a gap of several milliseconds between the termination of the BAEP waveform and the Pl component of the PCAEP. It seemed likely therefore that this time zone could harbour additional AEPs which might be detected with appropriate electrode positioning. When epidural electrodes were inserted over likely generator sites such as the inferior colliculus and the thalamus, this prediction appeared to be confirmed (Shaw, 1993). Figure 2 shows examples of the characteristic auditory waveforms recorded from over the inferior colliculus and thalamus plus their temporal relationship to both the BAEP and PCAEP. Two basic types of slow positivities could be distinguished. The earlier wave could be recorded only over the inferior colliculus and had a peak latency which occurred approximately 1 msec after BAEP wave V. In contrast, the second slow positivity could be optimally recorded only over the thalamus and occurred just under 2 msec later than the putative midbrain potential, and just over 1 msec earlier than the primary cortical response (Pl). While the origins of these two waves have not yet been established with

certainty, it is notable that their latencies are about what would be predicted for post-synaptic activity arising in the inferior colliculus and MGB, respectively. In addition, their waveforms are similar in both timing and morphology to those recorded from their presumed generators. At the least, their timing is more consistent with a collicular or thalamic origin than that of waves IV or V of the BAEP. The temporal relationship between the BAEP and PCAEP described in this section is not unique to the rat. When the relevant data is collated for both cat (e.g., Teas and Kiang, 1964; Achor and Starr, 1980a; Chen and Buchwald, 1986) and monkey (e.g. Arezzo et al., 1975; Legatt et al., 1986a), then the same relative timing between the two waveforms is revealed. As will be demonstrated in the next section, the same temporal relationship between the BAEP and the PCAEP also seems to exist in man.

3. CENTRAL AUDITORY CONDUCTION TIME IN MAN The schematic and idealized AEPs shown in Fig. 3 illustrate how CACT can be estimated in man. The artificial BAEP waveform shown in the figure was constructed using the latency data published by Allison et al. (1983). The PCAEP waveform was constructed by amalgamating and then averaging the latencies reported by Celesia (1976), Lee et al. (1984) and Liegeois-Chauvel et al. (1991). The PCAEP waveform recorded using an electrode inserted in or near Heschl’s gyri has a configuration virtually identical to that recorded from the rat (Figs 1 and 2),

Temporal Relationship Between Brainstem and Cortical Potentials although with longer latencies. The mean value of Pl derived from the three human studies was 16 msec. Figure 3 also shows, the IPLs between Pl and the seven BAEP waves. CACT in man is calculated by subtracting the latency of cochlear nucleus wave III (the homologue of wave II in the rat)1 from that of Pl of the PCAEP. According to Fig. 3, this would indicate that CACT in man is of the order of 12 msec. This is more than double the time taken for a somatosensory volley to traverse the central pathways (Cant and Shaw, 1986). Presumably the same explanation given for the comparably longer conduction time in the central auditory pathways of the rat is equally applicable to man. It is also notable that the wave V-PI IPL is 10 msec, a conduction time at least twice as long as would be predicted if wave V had a genuine midbrain origin. The timing of wave V relative to Pl therefore provides additional, if indirect, evidence in support of the notion that wave V must arise in a location below that of the inferior Icolliculus.

4. THE ORIGINS OF THE LATE BAEP WAVES Evidence that the late BAEP waves VI and VII may arise from the MGB and auditory radiations has always remained tenuous and insubstantial (Hall,

99

1992). Stockard and Rossiter (1977) described a patient in whom an apparently distorted wave VI was associated with a thalamic lesion. However, the intrinsic variability and instability of these later components make such clinico-pathologic correlations suspect and unreliable (Rowe, 1978; Chiappa et al., 1979; Hughes et al., 1988). There is also a dearth of relevant animal studies on this matter. Although some investigators using both cats (Buchwald et al., 1981; Knight and Brailowsky, 1990) and monkeys (Legatt et al., 1986b) have concluded that a diencephalic origin for the later waves is possible, these claims are fraught with the same difficulties discussed earlier over trying to account for unrealistically long conduction times in the auditory radiations. The same problem is also encountered in the human auditory system when attempting to reconcile the timing of BAEP wave VI with a thalamic origin. For example, it is apparent from Fig. 3 that the time intervals between wave VI and Pl, and wave VII and Pl are 9 and 7 msec, respectively. Theoretically, it would be difficult to account for a thalamo-cortical conduction time much beyond 2 msec. This assumes a pathway length of no more than 70 mm, a conduction velocity of 40 m/set and allows an intracortical synaptic delay of less than 0.5 msec. Desmedt and Cheron (1980) have previously estimated transmission time in the thalamo-cortical pathways of the somatosensory system to also be

BAEP

PCAEP

b

1

2

3

4

6

0

7

B

0

1

10 11 12 13 14 16 10 17 18 19 20 21 22 23 24 LATENCY

(MS)

Fig. 3. A sch’ematicillustration showing the temporal relationship between the BAEP and the PCAEP recorded from man. The seven high frequency components which compose the BAEP and the Pl component of the PCAEP are identified. The IPLs between each of the BAEP waves and Pl are also indicated. The size of the artificial waveforms is arbitrary and does not represent relative amplitudes.

N.

100

A. Shaw

close to 2 msec. It seems, therefore, that like their more robust early counterparts, the timing of waves VI and VII relative to Pl is compatible only with an origin in the more caudal parts of the central auditory system. In fact, given their timing and lability, it seems plausible to conclude that they may represent nothing more complex than the repetitive firing of brainstem nuclei. Among the most widely quoted evidence (e.g. Robinson and Rudge, 1982) that the thalamus is the generator of wave VI is a set of depth recordings from a patient reported by Hashimoto and his colleagues (Hashimoto et al., 1981). When the electrode was inserted into the MGB, a quite well-defined wave VI was recorded but this potential rapidly diminished in amplitude following small displacements of the electrode. As Hughes et al. (1988) point out, the behaviour of wave VI is not incompatible with its origin being elsewhere and recorded via volume conduction at the MGB. There is, however, perhaps a more serious flaw in this data. Assuming the primary cortical response has a latency of 16 msec and a thalamo-cortical transit time of 2 msec, then it can be presumed that the thalamic response will not occur much earlier than 14 msec. An examination of the recordings made by Hashimoto and his associates reveals that the time-base employed was merely 10 msec. A sweep length of at least twice this duration would seem to be necessary to adequately capture any thalamic auditory activity. Irrespective of the observation that wave VI seemed to be optimally recorded by an electrode in the MGB, the timing of this potential seems far too early to represent a genuine thalamic response.

5. THE ORIGINS

OF THE EARLY MAEP WAVES

If the artificial waveforms in Fig. 3 are re-examined, it is apparent that there is a hiatus between the termination of the BAEP waveform (around 9-10 msec) and the primary cortical response (at 16 msec). It is within this time-frame that collicular and thalamic AEPs would be predicted to occur. The question therefore arises whether any additional scalp AEPs can be recorded during this period which might represent far field reflections of collicular and thalamic activity. Such responses would presumably be homologous to the putative collicular and thalamic AEPs detected during the comparable time period in the rat (Fig. 2). The most likely candidates for such hypothetical midbrain and subcortical AEPs would be the two early, low amplitude, negative components of the MAEP. These are usually labelled No and Na. The latencies of No and Na do fall more or less within the expected time domain, although Na is often reported to have a latency slightly beyond the predicted range. Mean values for MAEP components including No and Na have frequently been reported in the literature (e.g. Picton et al., 1974; McFarland et al., 1975; Streletz et al., 1977; Cohen, 1982; Hashimoto, 1982; Ozdamar et al., 1982; Kileny et al., 1983, 1987; Suzuki et al., 1984; Woods and Clayworth, 1986; Yokoyama et al., 1987; Deiber et al., 1988). In these twelve studies, the mean latency

for Na was 17 msec (range 15-20 msec). Only seven of these papers also reported values for No, where the mean latency was 10 msec (range 9-13 msec). It has long been suspected that the early MAEP waves probably did arise in or near the inferior colliculus and the MGB (Hall, 1992) although there was always the implicit difficulty of reconciling such a notion with the belief that the later BAEP waves may somehow share the same generators. Of the two responses, a collicular origin for No is rather better established than is a thalamic origin for Na. As discussed earlier, the No component of the MAEP appears to correspond to the SNlO response described by Davis and Hirsh (1979) (Hashimoto, 1982). Considering its latency, Davis and Hirsh suggested a collicular source was most likely for SNlO. More direct evidence for a midbrain generator was obtained by depth recordings conducted by Hashimoto et al. (1981), Hashimoto (1982), and Moller and Jannetta (1982). These investigators showed that a large slow negative potential of probable post-synaptic origin could be recorded from or near the inferior colliculus with a peak latency which coincided with the SNlO waveform recorded from the scalp. When this potential was systematically traced from the midbrain to the scalp, it diminished markedly in amplitude but was still clearly visible in surface recordings (Hashimoto, 1982). In contrast, the evidence that Na arises in or near the thalamus is more circumstantial. In fact, Hashimoto (1982) decided that Na was generated in the same midbrain location as No, although the relative timing of Na suggests a diencephalic origin is more realistic. This is consistent with the scalp mapping study conducted by Deiber et al. (1988) who concluded that Na was a far field response of probable subcortical generation. Similarly, Kileny et al. (1987) reported the preservation of Na in patients with lesions of the temporal lobe. which is also compatible with a subcortical source for this component. Determining the origin of Na in a more definitive manner has been partly stymied by the range in its latency reported by different studies. If Na is a thalamic potential, then its latency should precede that of the primary cortical response by no more than 2 msec. Unfortunately, the variable conditions under which MAEPs have been recorded make comparisons with the PCAEP unreliable and presumably account for the occasions when the latency appears to slightly exceed that of the Pl component of the PCAEP. It is most likely that these aberrant latencies are a product of inappropriately narrow settings of the bandpass of the recording system (e.g., McFarland et al., 1975; Suzuki et al., 1984; Kraus et al., 1987; Hall, 1992). This problem may be compounded by the occurrence of post-auricular muscle activity which tends to coincide with the latency of Na and therefore may contaminate and obscure the component (Jacobson et al., 1990; Hall, 1992). There is also a dearth of intracranial recordings of thalamic auditory activity with which to correlate Na. For instance, Hashimoto (1982) failed to detect a distinct thalamic component when a depth electrode was positioned in the vicinity of the MGB. Even at

Temporal Relationship Between Brainstem and Cortical Potentials NOMENCLATURE

0

SNlO OR

SPS

101

Pl

No

POLARITY

SAEP

0

POSITIVE

NEQATIVE

NEQATIVE

POSITIVE

‘WAVES

SUPERIMPOSED ON

RISINQ

Ill -

SLOPE

,

o

1’

1

‘1

2

3

‘1’

4

6

S

7

*

‘.

8

0 10111213141316171019202122

““a.

LATENCY

‘1

.‘*

4

(MS)

Fig. 4. A schematic diagram of the proposed revision of the physiological organisation of the auditory pathway. The four principal slow waves which are serially generated from brainstem to cortex are shown along with their usual nomenclature, putative origin and polarity of their dominant component. The point of occurrence of each of the five principal BAEP waves on the rising slope of the first slow component (SP6) is also indicated.

this level, the waveform recorded still continued to be dominated by the large ostensibly collicular negativity. However, only one of the five patients studied by Hashimoto was actually illustrated in the paper. On balance, it therefore seems likely that Na does reflect a volume-conducted thalamic response, given its far field origin and close temporal proximity to the primary cortical response. This implies the existence of a slow negative potential arising in the thalamus or auditory radiations which, under optimal stimulating and recording conditions, should have a latency of 14-16 msec. Unfortunately, no such potential has yet been discovered. There is one additional characteristic of the No and Na components which super6cially seems at odds with their proposed midbrain and thalamic origins. This is their negative polarity. Conventionally, neuro-electric events recorded from the skull or scalp at a distance from I heir generators are supposed to be of positive polarity (Desmedt and Cheron, 1981). This is exemplified by the three far field non-cortical waveforms illustrated in Fig. 2. This incongruity has been remarked upon by both Deiber et al. (1988) and Jacobson et al. (1990) who pointed out that a precedent seems to exist in the human somatosensory pathways. Following median nerve stimulation, a far field negative waveform (N18) can be recorded which is thought to arise in the thalamus or thalamo-cortical tracts (Desmedt and Cheron, 1981). In summary, the polarity of No and Na would not seem to seriously compromise the present conclusions regarding their likely origins. Even if their timing does not always fit perfectky with that of the PCAEP, their latencies are still much more congruent with a midbrain and subcortical origin than are those of waves V, VI and VII of the BAEP.

6. CONCLUSIONS The present comparison between the timing of the BAEP and PCAEP waveforms suggests that some revision is required to the current understanding of the physiological organisation of the central auditory pathway. According to the standard theory, as the afferent volley ascends to successively higher relay stations within the auditory system, there is a concomitant generation of a high frequency wave. This focal generation theory is now recognised as an over-simplification and it has been acknowledged for some time that most of the fast BAEP components represent an amalgam of activity from more than one source (Achor and Starr, 1980a,b; Legatt ef al., 1986a,b; Hall, 1992). Nonetheless, the basic notion of the serial generation of the BAEP waves still persists, both as a framework with which to analyse patient data (e.g. Chiappa, 1990) and as a paradigm to interpret the findings of experimental animal data (e.g. Legatt er al., 1986b). The conclusions of the present review suggest that this model is, at best, only partially correct. As the sensory signal ascends the tracts and nuclei of the central auditory system, it appears to generate not so much a series of fast components but rather a succession of slow overlapping waves of varying positive and negative polarity with a frequency content which denotes a post-synaptic origin. The first of these slow waves is a positivity whose peak latency almost coincides with wave V of the BAEP. In the present review, it is labelled slow positive 6 (SP6) and represents the initial component of the slow brainstem response described by Davis and Hirsh (1979). SP6 may be an amorphous mixture of brainstem activity, although Moller and Jannetta (1982) have suggested a more

102

N. A. Shaw

discrete generation in the tracts of the lateral lemniscus. The homologue of SP6 in the rat can be seen in Fig. 2 as the low amplitude positivity with a peak latency of about 3 msec which precedes the thalamic and cortical waveforms. SP6 is followed by SNlO which represents the negative half of the slow brainstem response and is identical to the earliest component of the MAEP waveform (No). The generator of the SNlO/No seems most likely to be in the midbrain while the most reasonable origin for the subsequent negativity (Na) remains the MGB or auditory radiations. The final component in this sequence of auditory waveforms arises in the primary auditory cortex (PI). This modified concept of the behaviour of the primary auditory system is illustrated in Fig. 4. It can be seen from the diagram that the principal high frequency BAEP waves are almost entirely superimposed on the rising slope of the earliest of the slow components (SP6) (Hashimoto et al., 1981). The slow positivity which underlies the BAEP is not usually apparent in routine recordings because it is normally filtered out by setting the high pass filter at about 100 Hz, as was done in the BAEP recordings shown in Figs 1 and 2. If the present analysis is correct, then the BAEP must be a rather more parochial measure of activity within the central auditory system than is commonly assumed. In order to reconcile the timing of the BAEP with the PCAEP it seems unlikely that the former potential could reflect any significant activity being generated more rostrally than about the level of the superior olivary complex. If the sources of the central BAEP waves actually are confined to the more caudal part of the auditory brainstem and are not serially generated throughout much of the auditory pathway, then the question must be raised as to what is their functional significance. It seems probable therefore that the high frequency BAEP waves reflect the non-linear intra-pontine processing, integration and modulation that the auditory signal undergoes soon after entering the brainstem. Such a conclusion is very much in accord with the series of animal studies conducted by Starr and his colleagues since 1980 (Achor and Starr, 1980a,b; Wada and Starr, 1983a,b, 1983~; Zaaroor and Starr, 1991a,b). In these, it has been consistently found that the principal BAEP waves do not seem to arise from locations much beyond the lower half of the pons. It could also be inferred from the present analysis that infarction of the upper pons should nonetheless leave the BAEP waveform basically intact and this is exactly what has been found (e.g. Ferbert et al., 1988). Acknowledgemenrs-Preparation of this review was funded by a grant from the Deafness Research Foundation of New Zealand. The author thanks Dr Suzanne Purdy and Dr Peter Thome for their help and encouragement. REFERENCES Achor, L. J. and Starr, A. (198Oa) Auditory brainstem responses in the cat. 1. Intracranial and extracranial recordings. Electroenceph. Clin. Neurophysiol. 48, 15&l 73. Achor, L. J. and Starr, A. (1980b) Auditory brainstem responses in the cat. II. Effects of lesions. Electroenceph. Clin. Neurophysiol. 4% 174-190.

Allison, T., Wood, C. C. and Goff, W. R (1983) Brainstem auditory,

pattern-reversal visual, and short-latency somatosensory evoked potentials: latencies in relation to age, sex, and brain and body size. Electroenceph. Clin. Neurophysiol. 55, 619-636. Arerzo, J., Pickoff, A. and Vaughan, H. G. (1975) The sources and intracerebral distribution of auditory evoked potentials in the alert rhesus monkey. Brain Res. 90, 57-73. Buchwald, J. S., Hinman, C., Norman, R. J., Huang, C-M. and Brown, K. A. (1981) Middle- and long-latency auditory evoked responses recorded from the vertex of normal and chronically lesioned cats. Brain Res. 205, 91-109. Cant, B. R. and Shaw, N. A. (1986) Central somatosensory conduction time: method and clinical applications. In: Evoked Potentials, pp. 58-67, Eds. R. Q. Cracco and I. Bodis Wollner. Alan R. Liss: New York. Celesia, G. G. (1976) Organization of auditory cortical areas in man. Brain 99, 403414. Chen, B. M. and Buchwald, J. S. (1986) Midlatency auditory evoked responses: differential effects of sleep in the cat. E/ectroenceph. Clin. Neurophysiol. 65, 373-382.

Chiappa, K. H. (1990) Brainstem auditory evoked potentials: interpretation. In: Evoked Potentials in Clinical Medicine, 2nd edn, pp. 223-284, Ed. K. H. Chiappa. Raven Press: New York. Chiappa, K. H., Gladstone, K. J. and Young, R. R. (1979) Brainstem auditory evoked responses: studies of waveform variations in 50 normal human subjects, Arch. Neural. 36, 81-87. Cohen, M. M. (1982) Corona1 topography of the middle latency auditory evoked potentials (MLAEPs) in man. Electroenceph. Clin. Neurophysiol. 53, 231-236.

Davis,H. and Hirsh, S. K. (1979) A slow brain stem response for low-frequency audiometry. Audiology 18, 44%l61. Deiber, M. P., Ibanez, V., Fischer, C., Perrin, F. and Mauguiere, F. (1988) Sequential mapping favours the hypothesis of distinct generators for Na and Pa middle latency auditory evoked potentials. Electroenceph. C/in. Neurophysiol. 71, 1X7-197. Desmedt, J. E. and Cheron, G. (1980) Central somatosensory conduction in man: Neural generators and interpeak latencies of the far-field components recorded from neck and right or left scalp and earlobes. Electroenceph. C/in. Neurophysiol. 50, 382403.

Desmedt, J. E. and Cheron, G. (1981) Non-cephalic reference recording of early somatosensory potentials to finger stimulation in adult or aging normal man: Differentiation of widespread N18 and contralateral N20 from the prerolandic P22 and N30 components. Electroenceph. C/in. Neurophysiol. 52, 553-570.

Dorfman, L. J. (1983) Sensory evoked potentials: clinical applications in medicine. A. Rev. Med. 34, 473-489. Ferbert, A., Buchner, H., Bruckmann, H., Zeumer, H. and Hacke, W. (1988) Evoked potentials in basilar artery thrombosis: Correlation with clinical and angiographic findings. Electroenceph. Clin. Neurophysiol. 69, 136-147. Hall, J. W. (1992) Handbook of Auditory Evoked Responses. Allyn and Bacon: Boston. Hashimoto, I. (1982) Auditory evoked potentials from the human midbrain: slow brain stem responses. Electroenceph. C/in. Neurophysiol. 53, 652-657.

Hashimoto, I., Ishiyama, Y., Yoshimoto, T. and Nemoto, S. (1981) Brain-stem auditory-evoked potentials recorded directly from human brain-stem and thalamus. Brain 104, 841-859. Huang, C.-M. and Buchwald, J. S. (1977) Interpretation of the vertex short-latency acoustic response: A study of single neurcms in the brainstem. Brain Res. 137, 291-303. Hughes, J. R., Helgason, C. M. and Wilbur, A. (1988) Neuroanatomic correlations with the late waves of the brain-stem auditory evoked potential. Electroenceph. Clin. Newophysiol. 71, 367-374.

Jacobson, G. P., Privitera, M., Neils, J. R., Grayson, A. S. and Yeh, H.-S. (1990) Effects of anterior temporal lobectomy (ATL) on the middle-latency auditory evoked potential (MLAEP) Electroenceph. C/in. Neurophysiol. 75, 230-241.

Jewett, D. L. (1970) Volume-conducted potentials in response to auditory stimuli as detected by averaging in thecat. Electroenceph. Clin. Neurophysiol. 28, 609618.

Kileny, P., Dobson, D. and Gelfand, E. T. (1983) Middle-latency auditory evoked responses during open-heart surgery with hypothermia. Electroenceph. Clin. Neurophysiol. 55, 268-276. Kileny. P.. Paccioretti. D. and Wi!sw \ F. (lop- rfyx <>fcortical

Temporal Relationship Between Brainstem and Cortical Potentials lesions on middle-latency

auditory

evoked responses (MLR).

Electroenceph. Clin. Neurophysiol. 66, 108-l 20.

Knight, R. T. and Brailowsky, S. (1990) Auditory evoked potentials from the primary auditory cortex of the cat: Topographic and pharmacological studies. Electroenceph. C/in. Neurophysiol. 77, 225-232.

Knight, R. T., Hillyard, S. A., Woods, D. L. and Neville, H. J. (1980) The effects of frontal and temporal-parietal lesions on the auditory evoked potential in man. Electroenceph. Clin. Neurophysiol. 50, 112-124. Knight, R. T., Scabini, D., Woods, D. L. and Clayworth, C. (1988) The effects of lesions ol’the superior temporal gyrus and inferior parietal lobe on temporal and vertex components of the human AEP. Electroenceph. C/in. Neurophysiol. 70, 49%509. Kraus, N., Reed, N., Smth, D. I., Stein, L. and Cartee, C. (1987) High-pass filter settings affect the detectability of MLRs in humans. Electroenceph. Clin. Neurophysiol. 68, 236236. Lee, Y. S., Lueders, H., Dinner, D. S., Lesser, R. P., Hahn, J. and Klem, G. (1984) Recording of auditory evoked potentials in man using chronic subdural electrodes. Brain 107, 115-13 1. Legatt, A. D., Arezzo, J. C. and Vaughan, H. G. (1986a) Short-latency auditory evoked potentials in the monkey. I. Wave shape and surface topography. Electroenceph. C/in. Neurophysiol. 64, 41-52.

Legatt, A. D., Arezzo, J. C. and Vaughan, H. G. (1986b) Short-latency auditory evoked potentials in the monkey. II. Intracranial generators. Electroenceph. Clin. Neurophysiol. 64, 53-73.

Liegeois-Chauvel, C., hdusolino, A. and Chauvel, P. (1991) Localization of the primary auditory area in man. Brain 114, 139-I 53. McFarland, W. H., Vivicln, M. C., Wolf, K. E. and Goldstein, R. (1975) Re-examination of effects of stimulus rate and number on the middle components of the averaged electroencephalic response. Audiology 14,, 456-465. Moller, A. R. and Burgess, J. (1986) Neural generators of the brain-stem auditory evoked potentials (BAEPs) in the rhesus monkey. Electroenceph. Clin. Neurophysiol. 65, 361-372. Moller, A. R. and Jannetta, P. J. (1982) Evoked potentials from the inferior colliculus in man. Electroenceph. Clin. Neurophysiol. 53, 612-620. Moller, A. R. and Jannetra, P. J. (1983) Auditory evoked potentials recorded from the cochlear nucleus and its vicinity in man. J. Neurosurg 59, 1013-1018. Moller, A. R., Jannetta, P., Bennett, M. and Moller, M. B. (1981) Intracranially recorded responses from the human auditory nerve: New insights into the origin of brain stem evoked potentials (BSEPs). Electroenceph. Clin. Neurophysiol. 52, 18-27. Moller, A. R., Jannetta, P. J. and Sekhar, L. N. (1988) Contributions from the auditory nerve to the brain-stem auditory evoked potentials (BAEPs): Results of intracranial recording in man. Electroenceph. Clin. Neurophysiol. 71, 198-211.

Ozdamar, O., Kraus, N. and Curry, F. (1982) Auditory brain stem and middle latency responses in a patient with cortical deafness. Electroenceph. Clin. Neurophysiol. 53, 224-230.

Picton, T. W., Hillyard, S. A., Krausz, H. I. and Galambos, R. (1974) Human auditory evoked potentials. I: Evaluation of components. Electroenceph. Clin. Neurophysiol. 36, 179190. Richer, F., Alain, C., Achim, A., Bouvier, G. and Saint-Hilaire, J.-M. (1989) Intracerebral amplitude distributions of the auditory evoked potential. Electroenceph. Clin. Neurophysiol. 74, 202-208.

Robinson, K. and Rudge, P. (1982) Centrally generated auditory potentials. In: Evoked Potentials in Clinical Testing, pp. 345-372, Ed. A. M. Halliday. Churchill Livingstone: Edinburgh. Rowe, M. J. (1978) Normal variability of the brain-stem auditory evoked response in young and old adult subjects. Electroenceph. Clin. Neurophysiol. 44, 459-470.

103

Scherg, M. and Von Cramon, D. (1985) Two bilateral sources of the late AEP as identified by a spatio-temporal dipole model. Electroenceph. Clin. Neurophysiol. 62, 32-44.

Scherg, M. and Von Cramon, D. (1986) Evoked dipole source potentials of the human auditory cortex. Electroenceph. Clin. Neurophysiol. 65, 344-360.

Shaw, N. A. (1990a) Central auditory conduction time in the rat. Expl. Brain Res. 79, 217-220.

Shaw, N. A. (1990b) A simple technique for localising the primary auditory cortex in the rat. physiol. Behau. 48, 207-209. Shaw, N. A. (1993) Auditory evoked potentials recorded from different skull locations in the rat. ht. J. Neuroscience. 70, 277-283.

Shaw, N. A. and Cant, B. R. (1981) The effect of pentobarbital on central somatosensory conduction time in the rat. Electroenceph. C/in. Neurophysiol. 51, 674-677. Starr, A. and Squires, K. (1982) Distribution of auditory brainstem potentials over the scalp and nasopharynx in humans. Ann N. Y. Acad. Sci. 388, 427442. Stockard, J. J. and Rossiter, V. S. (1977) Clinical and pathologic correlates of brainstem auditory response abnormalities. Neurology (Minneap.) 27, 316325. Stockard, J. J., Stockard, J. E. and Sharbrough, F. W. (1980) Brainstem auditory evoked potentials in neurology: Methodology, interpretation, clinical application. In: Electrodiagnosis in Clinical Neurology, pp. 370413, Ed. M. J. Aminoff. Churchill Livingstone: New York. Streletz, L. J., Katz, L., Hohenberger, M. and Cracco, R. Q. (1977) Scalp recorded auditory evoked potentials and sonomotor responses: An evaluation of components and recording techniques. Electroenceph. Clin. Neurophysiol. 43, 192-206. Suzuki, T., Hirabayashi, M. and Kobayashi, K. (1984) Effects of analog and digital filtering on auditory middle latency responses in adults and young children. Ann. Otol. Rhinol. Lar. 93,267-270. Teas, D. C. and Kiang, N. Y.-S. (1964) Evoked responses from the auditory cortex. Exp. Neural. 10, 91-119. Vaughan, H. G. and Ritter, W. (1970) The sources of auditory evoked responses recorded from the human head. Electroenceph. Clin. Neurophysiol. 28, 360-367.

Wada, S-I. and Starr, A. (1983a) Generation of auditory brainstem responses (ABRs). I. Effects of injection of a local anesthetic (procaine HCl) into the trapezoid body of guinea pigs and cat. Electroenceph. Chin. Neurophysiol. 56, 326339.

Wada, S.-I. and Starr, A. (1983b) Generation of auditory brain stem responses (ABRs). II. Effects of surgical section of the trapezoid body on the ABR in guinea pigs and cat. Electroenceph. Clin. Neurophysiol. 56, 340-351.

Wada, S.-I. and Starr, A. (1983~) Generation of auditory brainstem responses (ABRs). III. Effects of lesions of the superior olive, lateral lemniscus and inferior colliculus on the ABR in guinea pig. Electroenceph. Clin. Neurophysiol. 56, 352-366.

Wood, C. C. and Wolpaw, J. R. (1982) Scalp distribution of human auditory evoked potentials. II. Evidence for overlapping sources and involvement of auditory cortex. Electroenceph. Clin. Neurophysiol. 54, 2S38.

Woods, D. L. and Clayworth, C. C. (1986) Age-related changes in human middle latency auditory evoked potentials. Electroenceph. Clin. Neurophysiol. 65, 297-303.

Yokoyama, T., Ryu, H., Uemura, K., Miyamaoto, T. and Imamura, Y. (1987) Study of the constant wave from of ML-AEP in humans. Electroenceph. Clin. Neurophysiol. 67, 372-378. Zaaroor, M. and Starr, A. (1991a) Auditory brain-stem evoked potentials in cat after kainic acid induced neuronal loss. I. Superior olivary complex. Electroenceph. Clin. Neurophysiol. 80, 422435.

Zaaroor, M. and Starr, A. (1991b) Auditory brain-stem evoked potentials in cat after kainic acid induced neuronal loss. II. Cochlear nucleus. Electroenceph. Clin. Neurophysiol. 80,43&445.