Electroencephalography and clinical Neurophysiology , 84 (1992) 196-200 © 1992 Elsevier Scientific Publishers Ireland, Ltd. 0168-5597/92/$05.00
196
EVOPOT 91145
Short communication
Cellular generators of the cortical auditory evoked potential initial c o m p o n e n t Mitchell Steinschneider a,b, Craig E. Tenke b, Charles E. Schroeder b, Daniel C. Javitt b,c, Gregory V. Simpson a,b, Joseph C. Arezzo a,b and Herbert G. Vaughan, Jr. a,b Departments of" Neurology, b Neuroscience and c Psychiatry, Rose F. Kennedy Center, Albert Einstein College of Medicine, Bronx, N Y 10461 (U.S.A.) (Accepted for publication: 26 November 1991)
Summary Cellular generators of the initial cortical auditory evoked potential (AEP) component were determined by analyzing laminar profiles of click-evoked AEPs, current source density, and multiple unit activity (MUA) in primary auditory cortex of awake monkeys. The initial AEP component is a surface-negative wave, N8, that peaks at 8-9 msec and inverts in polarity below lamina 4. N8 is generated by a lamina 4 current sink and a deeper current source. Simultaneous MUA is present from lower lamina 3 to the subjacent white matter. Findings indicate that thalamocortical afferents are a generator of N8 and support a role for lamina 4 stellate cells. Relationships to the human AEP are discussed. Key words: Auditory cortex; Evoked potential; Current source density; Multiple unit activity; Cellular generators
The advent of sophisticated source analysis of middle latency auditory evoked potentials (AEPs) (e.g., Scherg and Von Cramon 1986) and the increased availability of intracortically recorded responses performed during epilepsy surgery evaluation (e.g., Liegeois-Chauvel et al. 1991) have facilitated examination of the earliest activity evoked by sound in human auditory cortex. The utility of these short latency components for understanding auditory cortical physiology would be enhanced if the underlying cellular generators were known. We determined the generators of the initial cortical AEP component by analyzing laminar profiles of concurrently recorded AEPs, the derived current source density (CSD), and multiple unit activity (MUA) in primary auditory cortex (A1) of awake monkeys.
Methods and materials Six adult male Macaca fascicularis were studied. All participated in other experiments. Methodological procedures have been previously described (Steinschneider et al. 1982; Schroeder et al. 1991). Briefly, under general anesthesia (sodium pentobarbital) and using aseptic techniques, matrices of adjacently placed 18-gauge tubes were stereotaxically positioned to target A1 and to serve as guides for the recording electrodes. The matrices were positioned vertically with respect to stereotaxic coordinates in 4 monkeys, and at an angle to approximate the tilt of the superior temporal plane in 2 others. Animals were housed in our AAALAC-accredited animal facility and were monitored daily for assessment of their well-being. Recordings were obtained from multicontact electrodes containing 8-15 recording contacts evenly spaced at 75-400/zm intervals. The reference was an occipital bone electrode. Brain potentials were recorded using unity gain headstage amplification, followed by amplification at a gain of 5000 and a frequency response down 3 dB at 3 and 3000 Hz. AEPs were averaged by computer with a sampling rate greater
Correspondence to: Dr. M. Steinschneider, Department of Neurology, Rose F. Kennedy Center, Room 322, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461 (U.S.A.). Tel.: 212-430-4115; Fax: 212-824-3058.
than 3 kHz. Neuroelectric signals which generated the AEPs were simultaneously high-pass filtered above 500 Hz, digitized, full-wave rectified, and averaged by computer for analysis of multiple unit activity (MUA). MUA represented the summed action potential activity of neuronal ensembles within a radius of 50-100 tzm. Onedimensional current source density (CSD) analyses were performed to identify the pattern of current sources and sinks within auditory cortex. CSD analysis computes the second spatial derivative of voltage across electrode sites and defines the location, direction and amplitude of transmembrane currents that generate evoked potentials. Current sinks represent locations of net inward transmembrane current flow. They are produced by either depolarizing events (excitatory post-synaptic potentials and axonal depolarizations) or by passive, circuit-completing transmembrane currents evoked by hyperpolarizing potentials. Current sources indicate sites of net outward transmembrane currents and are produced by either hyperpolarizing events or current return for depolarizing potentials. Both 3- and 5-point approximation formulas were used to calculate the CSD, using differentiation grids of 75-400/xm. Stimuli were condensation clicks generated by 100 /Lsec positive square-wave pulses delivered every 650 msec. Clicks were generally delivered at 80 dB SPL to the ear contralateral to the recording sites. Subjects were awake and seated in a primate chair. Electrodes were moved within the brain using a microdrive. Penetration into auditory cortex was indicated on-line by the inversion of AEP components in deeper electrode channels and by the presence of MUA. One to two hundred presentations of the clicks were typically used to generate the averages. Recordings were taken in each monkey over a period of 3-12 months. At the end of the recording series, the animals were deeply anesthetized with sodium pentobarbital and perfused with 10% buffered formalin. Tissue was histologically examined to reconstruct the electrode tracks, locate selected recorded sites which had been marked with iron deposition, and identify A1 using previously published criteria (Pandya and Sanides 1973).
Results Results are based upon 80 electrode penetrations into A1, surrounding auditory fields on the superior temporal plane, and the auditory thalamocortical (TC) radiations within the white matter
CELLULAR GENERATORS OF THE INITIAL CORTICAL AEP
197
medial to and below auditory cortex. Of these, 49 penetrations traversed A1. The click-evoked AEP within A1 exhibits a characteristic wave shape that begins with a surface-negative component, N8, that has an onset of 5-6 msec and a peak at 8-9 msec. Occasionally, the following surface-positive wave is diminished in amplitude by continuation of the negative component. N8 can be traced with decreasing amplitude from recordings in auditory cortex to the dorsolateral surface of the brain. Like the other early components of the AEP, N8 is of smaller amplitude in surrounding auditory cortical fields. The laminar profile of the AEP within A1 represents a complex sequence of changes in potential (Figs. 1 and 2). N8 increases in amplitude in supragranular laminae and inverts in polarity across the lamina 4-5 boundary. In contrast to N8, the following positivity inverts in polarity in supragranular layers. AEP wave forms labeled "A" and "B" in Fig. 1A were recorded at the indicated sites in Fig. 3, which also illustrates the electrode track traversing A1. The
laminar profiles from an adjacent electrode penetration (Fig. 1B), labeled "C" in Fig. 3, illustrate the reliability of the data. CSD analysis reveals that N8 arises from the initial current sink in A1, which is centered in lamina 4 with frequent extension into lower portions of lamina 3. Its duration and amplitude are rather variable. Despite this variability, the lamina 4 sink is always the initial manifestation of local cortical activation. The sink is invariably paired with a current source just below it in lamina 5. Occasionally, a small coincident source is also present in lower lamina 3. In contrast to the source-sink distribution associated with N8, the later positivity is coincident with a large lamina 3 sink and more superficial sources. Timing of MUA varies with intracortical depth. The earliest MUA can be traced from the subcortical white matter into lamina 4 and usually into lower lamina 3 (Fig. 2). It has an onset identical to the lamina 4 sink and a peak at 8-10 msec (see also click-evoked MUA, Figs. 2-8, Steinschneider et al. 1982). In subgranular laminae and subjacent white matter, this burst is typically 7-10 msec in
A
AEP
CSD
B
MUA
AEP
CSD
MUA
m~ *o.Te
*o.Io
®= *O,II
I
~
~
....
I.Ii
fll
I.le
IV s.IO
VI , . . I.IO
|'11
J.Ii
~.,~ ise°
o' ' ;, ' " ' '-- '
I
isoll
j,o +,
i.,.
,o~cK
I,.,-...
~,~,~,~,
.,:.
[] s,~,~E
.~ -r~
Ie°(I
u,-,
Fig. 1. Laminar patterns of the click-evoked AEP, CSD and MUA in A1 recorded at 125/.~m intervals from 2 parallel electrode penetrations. Five-point approximation formulas are used to calculate the CSD. The electrode tracks and recording locations of the wave forms labeled "A" and "B" are shown in Fig. 3. The initial cortical AEP component, N8, inverts in polarity within lamina 5. The earliest transmembrane current flow in the CSD occurs simultaneously with N8 and consists of a lamina 4 current sink and deeper current sources. MUA with an onset identical to N8 and the lamina 4 sink can be traced from the underlying white matter into lower lamina 3. The similarity in laminar profiles recorded from the two electrode penetrations illustrates the reliability of the data and supports the validity of the 1-dimensional CSD analysis by demonstrating synchronous activation of an extended region of A1. Laminar boundaries are shown to the left of each set of profiles.
M. STEINSCHNEIDER ET AL.
198 MUA
CSD
AEP
HI
IV
V
inveru~ ....J/
J 0
6
~ 12
18
24
1
2,0
36
0
6
msec
12
18 msec
+
[ 50pV
~ ~
trli lnl~
24
30
!!i][ii]]ii
36
':,!i%i source
] I m V / m m 2 ["'"""" ] sink
0
6
12
. . . . Tr~v~
15
2.4
30
36
msec
12rv . . . . . . . . . . . . . . . . . . . .
Fig. 2. Laminar patterns of the click-evoked AEP, CSD and MUA recorded within A1 from another subject at 200 p,m intervals. A 3-point approximation formula is used to calculate the CSD. Time lines at 8 msec, and approximate laminar boundaries, are also shown. Inversion of N8 occurs in lamina 5 and is deeper than the inversion of the following positive waves, which occurs in supragranular regions. The initial sink is simultaneous with N8 and is located in lamina 4, with deeper current sources. The small, superficial sources and sinks seen during N8 are inconsistent and represent variability in the CSD. Later current sinks are located in lamina 3, have prominent superficial sources, and reflect the transmembrane current flows generating the early positive components of the AEP. This laminar pattern is consistent with the classical current dipole of pyramidal cell activation. The earliest MUA has an onset identical with N8 and the initial cortical sink and is present from the subjacent white matter to mid-lamina 3 (solid arrowheads). It is absent in more superficial MUA (unfilled arrowhead).
duration and is often followed by a small reduction in firing rate (Fig. 1). Slightly later MUA occurs within supragranular layers and has an onset of 6-8 msec and a peak at about 11 msec.
Discussion
The initial cortical AEP component, N8, is a surface-negative wave that inverts in polarity across the lamina 4-5 boundary and is primarily generated by a lamina 4 sink and a deeper lamina 5 source. The timing and laminar distribution of the concurrent MUA indicate that a portion of this sink represents activity of the afferent volley in TC fibers. The initial MUA burst can be traced into the underlying white matter, indicating its axonal origin. It has an upper limit that extends into lower' lamina 3, consistent with the depths of TC fiber terminations in A1 (Pandya and Sanides 1973; Jones and Burton 1976). Timing of this MUA burst is appropriate for activation of
medial geniculate neurons and their projections within the auditory radiations (Legatt et al. 1986). The initial portion of the lamina 4 sink and the earliest MUA have identical onsets, further supporting the axonal origin of this transmembrane current flow because postsynaptic contributions to the initial sink should be delayed by the time of synaptic transmission. The finding that the initial lamina 4 current sink is partially generated by depolarization of axonal terminals disputes earlier investigations, which minimized the role these neural elements play in cortical CSD analyses (e.g., Muller-Preuss and Mitzdorf 1984; Mitzdorf 1985; Kulics and Cauller 1986). TC fibers from the medial geniculate nucleus enter A1 and branch profusely to form a dense terminal plexus (Jones and Burton 1976). We postulate that depolarization of this plexus generates a net extracellular current sink within the confines of the axon terminal ramifications. The current source necessary to balance this depolarization is drawn from more proximal portions of the axons within deeper cortical laminae, establishing
CELLULAR GENERATORS OF THE INITIAL CORTICAL AEP
199
t
f~
Fig. 3. Coronal section of the brain stained with cresyl violet, illustrating the 2 electrode tracks whose laminar profiles are shown in Fig. 1. T h e electrodes traversed A1, located on the lower bank of the sylvian fissure. Wave forms labeled " A " and " B " in Fig. 1A are recorded at the correspondingly labeled iron deposition sites. Iron deposition site " A " is located on the upper bank of the sylvian fissure, 0.5 m m above A1. Iron deposition site " B " is located near the boundary of laminae 3 and 4 in A1. Laminar profiles from Fig. 1B are recorded from the electrode penetration labeled "C." Separation between the two tracks is 0.8 mm.
a surface-negative dipole. A surface-positive c o m p o n e n t should precede N8, during the time when the axon terminals act as a source for the approaching depolarization. This c o m p o n e n t was identified in
the cortical recordings, but at a much smaller amplitude than N8, and corresponds to wave 8 of the simian brain-stem A E P that reflects activity in the TC radiations (Legatt et al. 1986). Lamina 4
200 current sinks generated by terminal portions of TC fiber tracts have also been described in monkey visual cortex (Schroeder et al. 1991). The duration of the lamina 4 sink is longer than the MUA burst in TC fibers, indicating that the initial lamina 4 sink also represents, in part, activation of cortical cells. Elements implicated are lamina 4 stellate cells, which receive monosynaptic input from TC fibers (Smith and Moskowitz 1979). Many of these cells are characterized by an asymmetrical and vertical orientation of their dendritic trees that ascend into lamina 3 (McMillen and Glaser 1982). Because stellate cells in auditory cortex receive TC afferents mainly on dendritic branches (Smith and Moskowitz 1979) concentrated in lower lamina 3 and upper lamina 4 (Jones and Burton 1976), the initial stetlate cell depolarization would lead to an open-field configuration with a sink surrounding the dendrites and a source near the cell bodies, resulting in a surface-negative potential. This cortical activation would temporally overlap the pre-synaptic responses for all but the initial 1-2 msec of the surface N8 component. Similar conclusions have been reached for the initial cortical component of the primate visual evoked response (Schroeder et al. 1991). The monkey N8 component corresponds to the initial polarity-inverting response in human auditory cortex, which is a negative wave with an onset of 8-10 msec and a peak at 13 msec (Liegeois-Chauvel et al. 1991). The short latency of this component is consistent with slightly earlier activity noted within the human medial geniculate (Velasco et al. 1982). Thus, we propose that terminal portions of TC fibers are a generator of this early cortical component in man. We also hypothesize that the initial negative component in the monkey cortical AEP is partially generated by lamina 4 stellate cells. Stellate cells with a vertical dendritic orientation also occur in middle cortical laminae of the human auditory cortex (Seldon 1981). This anatomical configuration is consistent with a vertically oriented, open-field generator situated in the thalamo-recipient zone. These data and the present findings support a role for stellate cells in the generation of the initial auditory cortical component. The intracortical N8 in monkey and N13 in man (Liegeois-Chauvel et al. 1991) appear to represent the initial segment of the earliest scalp-recorded auditory cortical component in man (Seherg and Von Cramon 1986). This scalp-recorded component is a negativity with a peak at 19 msec and an onset at about 10 reset. Inability to completely isolate the initial intracorticai negativity by scalp recordings may be due to several factors. There is temporal overlap of this early cortical activity with negative waves generated within the midbrain (Hashimoto 1982), which may obscure generator identification from scalp recordings. The 300 Hz upper cut-off frequency used for the scalp recordings may be insufficient to fully characterize the wave form of the high frequency intracortical negativity. Further work will be required to resolve this issue. We thank the excellent technical and secretarial assistance of Mona Litwak, Shirley Seto, Chester Freeman and Linda O'Donnell. Supported by Grants DC00657 from NIH, MH06723 from NIMH, and 90-31 from the McDonnell-Pew Program in Cognitive Neuroscience.
M. STEINSCHNEIDER ET AL. References
Hashimoto, I. Auditory evoked potentials from the human midbrain: slow brain stem responses. Electroenceph. clin. Neurophysiol., 1982, 53: 652-657. Jones, E.G. and Burton, H. Areal differences in the laminar distribution of thalamic afferents in cortical fields of the insular, parietal and temporal regions of primates. J. Comp. Neurol., 1976, 168: 197-248. Kulics, A.T. and Cauller, L.J. Cerebral cortical somatosensory evoked responses, multiple unit activity and current source-densities: their interrelationships and significance to somatic sensation as revealed by stimulation of the awake monkey's hand. Exp. Brain Res., 1986, 62: 46-60. Legatt, A.D., Arezzo, J.C. and Vaughan, Jr., H.G. Short-latency auditory evoked potentials in the monkey. II. Intracranial generators. Electroenceph. olin. Neurophysiol., 1986, 64: 53-73. Liegeois-Chauvel, C., Musolino, A. and Chauvel, P. Localization of the primary auditory area in man. Brain, 1991, 114: 139-153. McMillen, N.T. and Glaser, E.M. Morphology and laminar distribution of nonpyramidal neurons in the auditory cortex of the rabbit. J. Comp. Neurol., 1982, 208: 85-106. Mitzdorf, U. Current source-density method and application in cat cerebral cortex: investigations of evoked potentials and EEG phenomena. Physiol. Rev., 1985, 65: 37-100. Muller-Preuss, P. and Mitzdorf, U. Functional anatomy of the inferior colliculus and the auditory cortex: current source density analyses of click-evoked potentials. Hearing Res., 1984, 16: 133142. Pandya, D.N. and Sanides, F. Architectonic parcellation of the temporal operculum in rhesus monkey and its projection pattern. Z. Anat. Entwickl.-Gesch., 1973, 139: 127-161. Scherg, M. and Von Cramon, D. Evoked dipole source potentials of the human auditory cortex. Electroenceph. clin. Neurophysiol., 1986, 65: 344-360. Schroeder, C.E., Tenke, C.E., Givre, S.J., Arezzo, J.C. and Vaughan, Jr., H.G. Striate cortical contribution to the surface-recorded pattern-reversal VEP in the alert monkey. Vision Res., 1991, 31: 1143-1157. Seldon, H.L. Structure of human auditory cortex. I. Cytoarchitectonics and dendritic distributions. Brain Res., 1981, 229: 277-294. Smith, D.E. and Moskowitz, N. Ultrastructure of layer IV of the primary auditory cortex of the squirrel monkey. Neuroscience, 1979, 4: 349-359. Steinschneider, M., Arezzo, J.C. and Vaughan, Jr., H.G. Speech evoked activity in the auditory radiations and cortex of the awake monkey. Brain Res., 1982, 252: 353-365. Velasco, M., Velasco, F., Almanza, X. and Coats, A.C. Subcortical correlates of the auditory brain stem potentials in man: bipolar EEG and multiple unit activity and electrical stimulation. Electroenceph, clin. Neurophysiol., 1982, 53: 133-142.
196
EVOPOT 91145
Short communication
Cellular generators of the cortical auditory evoked potential initial c o m p o n e n t Mitchell Steinschneider a,b, Craig E. Tenke b, Charles E. Schroeder b, Daniel C. Javitt b,c, Gregory V. Simpson a,b, Joseph C. Arezzo a,b and Herbert G. Vaughan, Jr. a,b Departments of" Neurology, b Neuroscience and c Psychiatry, Rose F. Kennedy Center, Albert Einstein College of Medicine, Bronx, N Y 10461 (U.S.A.) (Accepted for publication: 26 November 1991)
Summary Cellular generators of the initial cortical auditory evoked potential (AEP) component were determined by analyzing laminar profiles of click-evoked AEPs, current source density, and multiple unit activity (MUA) in primary auditory cortex of awake monkeys. The initial AEP component is a surface-negative wave, N8, that peaks at 8-9 msec and inverts in polarity below lamina 4. N8 is generated by a lamina 4 current sink and a deeper current source. Simultaneous MUA is present from lower lamina 3 to the subjacent white matter. Findings indicate that thalamocortical afferents are a generator of N8 and support a role for lamina 4 stellate cells. Relationships to the human AEP are discussed. Key words: Auditory cortex; Evoked potential; Current source density; Multiple unit activity; Cellular generators
The advent of sophisticated source analysis of middle latency auditory evoked potentials (AEPs) (e.g., Scherg and Von Cramon 1986) and the increased availability of intracortically recorded responses performed during epilepsy surgery evaluation (e.g., Liegeois-Chauvel et al. 1991) have facilitated examination of the earliest activity evoked by sound in human auditory cortex. The utility of these short latency components for understanding auditory cortical physiology would be enhanced if the underlying cellular generators were known. We determined the generators of the initial cortical AEP component by analyzing laminar profiles of concurrently recorded AEPs, the derived current source density (CSD), and multiple unit activity (MUA) in primary auditory cortex (A1) of awake monkeys.
Methods and materials Six adult male Macaca fascicularis were studied. All participated in other experiments. Methodological procedures have been previously described (Steinschneider et al. 1982; Schroeder et al. 1991). Briefly, under general anesthesia (sodium pentobarbital) and using aseptic techniques, matrices of adjacently placed 18-gauge tubes were stereotaxically positioned to target A1 and to serve as guides for the recording electrodes. The matrices were positioned vertically with respect to stereotaxic coordinates in 4 monkeys, and at an angle to approximate the tilt of the superior temporal plane in 2 others. Animals were housed in our AAALAC-accredited animal facility and were monitored daily for assessment of their well-being. Recordings were obtained from multicontact electrodes containing 8-15 recording contacts evenly spaced at 75-400/zm intervals. The reference was an occipital bone electrode. Brain potentials were recorded using unity gain headstage amplification, followed by amplification at a gain of 5000 and a frequency response down 3 dB at 3 and 3000 Hz. AEPs were averaged by computer with a sampling rate greater
Correspondence to: Dr. M. Steinschneider, Department of Neurology, Rose F. Kennedy Center, Room 322, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461 (U.S.A.). Tel.: 212-430-4115; Fax: 212-824-3058.
than 3 kHz. Neuroelectric signals which generated the AEPs were simultaneously high-pass filtered above 500 Hz, digitized, full-wave rectified, and averaged by computer for analysis of multiple unit activity (MUA). MUA represented the summed action potential activity of neuronal ensembles within a radius of 50-100 tzm. Onedimensional current source density (CSD) analyses were performed to identify the pattern of current sources and sinks within auditory cortex. CSD analysis computes the second spatial derivative of voltage across electrode sites and defines the location, direction and amplitude of transmembrane currents that generate evoked potentials. Current sinks represent locations of net inward transmembrane current flow. They are produced by either depolarizing events (excitatory post-synaptic potentials and axonal depolarizations) or by passive, circuit-completing transmembrane currents evoked by hyperpolarizing potentials. Current sources indicate sites of net outward transmembrane currents and are produced by either hyperpolarizing events or current return for depolarizing potentials. Both 3- and 5-point approximation formulas were used to calculate the CSD, using differentiation grids of 75-400/xm. Stimuli were condensation clicks generated by 100 /Lsec positive square-wave pulses delivered every 650 msec. Clicks were generally delivered at 80 dB SPL to the ear contralateral to the recording sites. Subjects were awake and seated in a primate chair. Electrodes were moved within the brain using a microdrive. Penetration into auditory cortex was indicated on-line by the inversion of AEP components in deeper electrode channels and by the presence of MUA. One to two hundred presentations of the clicks were typically used to generate the averages. Recordings were taken in each monkey over a period of 3-12 months. At the end of the recording series, the animals were deeply anesthetized with sodium pentobarbital and perfused with 10% buffered formalin. Tissue was histologically examined to reconstruct the electrode tracks, locate selected recorded sites which had been marked with iron deposition, and identify A1 using previously published criteria (Pandya and Sanides 1973).
Results Results are based upon 80 electrode penetrations into A1, surrounding auditory fields on the superior temporal plane, and the auditory thalamocortical (TC) radiations within the white matter
CELLULAR GENERATORS OF THE INITIAL CORTICAL AEP
197
medial to and below auditory cortex. Of these, 49 penetrations traversed A1. The click-evoked AEP within A1 exhibits a characteristic wave shape that begins with a surface-negative component, N8, that has an onset of 5-6 msec and a peak at 8-9 msec. Occasionally, the following surface-positive wave is diminished in amplitude by continuation of the negative component. N8 can be traced with decreasing amplitude from recordings in auditory cortex to the dorsolateral surface of the brain. Like the other early components of the AEP, N8 is of smaller amplitude in surrounding auditory cortical fields. The laminar profile of the AEP within A1 represents a complex sequence of changes in potential (Figs. 1 and 2). N8 increases in amplitude in supragranular laminae and inverts in polarity across the lamina 4-5 boundary. In contrast to N8, the following positivity inverts in polarity in supragranular layers. AEP wave forms labeled "A" and "B" in Fig. 1A were recorded at the indicated sites in Fig. 3, which also illustrates the electrode track traversing A1. The
laminar profiles from an adjacent electrode penetration (Fig. 1B), labeled "C" in Fig. 3, illustrate the reliability of the data. CSD analysis reveals that N8 arises from the initial current sink in A1, which is centered in lamina 4 with frequent extension into lower portions of lamina 3. Its duration and amplitude are rather variable. Despite this variability, the lamina 4 sink is always the initial manifestation of local cortical activation. The sink is invariably paired with a current source just below it in lamina 5. Occasionally, a small coincident source is also present in lower lamina 3. In contrast to the source-sink distribution associated with N8, the later positivity is coincident with a large lamina 3 sink and more superficial sources. Timing of MUA varies with intracortical depth. The earliest MUA can be traced from the subcortical white matter into lamina 4 and usually into lower lamina 3 (Fig. 2). It has an onset identical to the lamina 4 sink and a peak at 8-10 msec (see also click-evoked MUA, Figs. 2-8, Steinschneider et al. 1982). In subgranular laminae and subjacent white matter, this burst is typically 7-10 msec in
A
AEP
CSD
B
MUA
AEP
CSD
MUA
m~ *o.Te
*o.Io
®= *O,II
I
~
~
....
I.Ii
fll
I.le
IV s.IO
VI , . . I.IO
|'11
J.Ii
~.,~ ise°
o' ' ;, ' " ' '-- '
I
isoll
j,o +,
i.,.
,o~cK
I,.,-...
~,~,~,~,
.,:.
[] s,~,~E
.~ -r~
Ie°(I
u,-,
Fig. 1. Laminar patterns of the click-evoked AEP, CSD and MUA in A1 recorded at 125/.~m intervals from 2 parallel electrode penetrations. Five-point approximation formulas are used to calculate the CSD. The electrode tracks and recording locations of the wave forms labeled "A" and "B" are shown in Fig. 3. The initial cortical AEP component, N8, inverts in polarity within lamina 5. The earliest transmembrane current flow in the CSD occurs simultaneously with N8 and consists of a lamina 4 current sink and deeper current sources. MUA with an onset identical to N8 and the lamina 4 sink can be traced from the underlying white matter into lower lamina 3. The similarity in laminar profiles recorded from the two electrode penetrations illustrates the reliability of the data and supports the validity of the 1-dimensional CSD analysis by demonstrating synchronous activation of an extended region of A1. Laminar boundaries are shown to the left of each set of profiles.
M. STEINSCHNEIDER ET AL.
198 MUA
CSD
AEP
HI
IV
V
inveru~ ....J/
J 0
6
~ 12
18
24
1
2,0
36
0
6
msec
12
18 msec
+
[ 50pV
~ ~
trli lnl~
24
30
!!i][ii]]ii
36
':,!i%i source
] I m V / m m 2 ["'"""" ] sink
0
6
12
. . . . Tr~v~
15
2.4
30
36
msec
12rv . . . . . . . . . . . . . . . . . . . .
Fig. 2. Laminar patterns of the click-evoked AEP, CSD and MUA recorded within A1 from another subject at 200 p,m intervals. A 3-point approximation formula is used to calculate the CSD. Time lines at 8 msec, and approximate laminar boundaries, are also shown. Inversion of N8 occurs in lamina 5 and is deeper than the inversion of the following positive waves, which occurs in supragranular regions. The initial sink is simultaneous with N8 and is located in lamina 4, with deeper current sources. The small, superficial sources and sinks seen during N8 are inconsistent and represent variability in the CSD. Later current sinks are located in lamina 3, have prominent superficial sources, and reflect the transmembrane current flows generating the early positive components of the AEP. This laminar pattern is consistent with the classical current dipole of pyramidal cell activation. The earliest MUA has an onset identical with N8 and the initial cortical sink and is present from the subjacent white matter to mid-lamina 3 (solid arrowheads). It is absent in more superficial MUA (unfilled arrowhead).
duration and is often followed by a small reduction in firing rate (Fig. 1). Slightly later MUA occurs within supragranular layers and has an onset of 6-8 msec and a peak at about 11 msec.
Discussion
The initial cortical AEP component, N8, is a surface-negative wave that inverts in polarity across the lamina 4-5 boundary and is primarily generated by a lamina 4 sink and a deeper lamina 5 source. The timing and laminar distribution of the concurrent MUA indicate that a portion of this sink represents activity of the afferent volley in TC fibers. The initial MUA burst can be traced into the underlying white matter, indicating its axonal origin. It has an upper limit that extends into lower' lamina 3, consistent with the depths of TC fiber terminations in A1 (Pandya and Sanides 1973; Jones and Burton 1976). Timing of this MUA burst is appropriate for activation of
medial geniculate neurons and their projections within the auditory radiations (Legatt et al. 1986). The initial portion of the lamina 4 sink and the earliest MUA have identical onsets, further supporting the axonal origin of this transmembrane current flow because postsynaptic contributions to the initial sink should be delayed by the time of synaptic transmission. The finding that the initial lamina 4 current sink is partially generated by depolarization of axonal terminals disputes earlier investigations, which minimized the role these neural elements play in cortical CSD analyses (e.g., Muller-Preuss and Mitzdorf 1984; Mitzdorf 1985; Kulics and Cauller 1986). TC fibers from the medial geniculate nucleus enter A1 and branch profusely to form a dense terminal plexus (Jones and Burton 1976). We postulate that depolarization of this plexus generates a net extracellular current sink within the confines of the axon terminal ramifications. The current source necessary to balance this depolarization is drawn from more proximal portions of the axons within deeper cortical laminae, establishing
CELLULAR GENERATORS OF THE INITIAL CORTICAL AEP
199
t
f~
Fig. 3. Coronal section of the brain stained with cresyl violet, illustrating the 2 electrode tracks whose laminar profiles are shown in Fig. 1. T h e electrodes traversed A1, located on the lower bank of the sylvian fissure. Wave forms labeled " A " and " B " in Fig. 1A are recorded at the correspondingly labeled iron deposition sites. Iron deposition site " A " is located on the upper bank of the sylvian fissure, 0.5 m m above A1. Iron deposition site " B " is located near the boundary of laminae 3 and 4 in A1. Laminar profiles from Fig. 1B are recorded from the electrode penetration labeled "C." Separation between the two tracks is 0.8 mm.
a surface-negative dipole. A surface-positive c o m p o n e n t should precede N8, during the time when the axon terminals act as a source for the approaching depolarization. This c o m p o n e n t was identified in
the cortical recordings, but at a much smaller amplitude than N8, and corresponds to wave 8 of the simian brain-stem A E P that reflects activity in the TC radiations (Legatt et al. 1986). Lamina 4
200 current sinks generated by terminal portions of TC fiber tracts have also been described in monkey visual cortex (Schroeder et al. 1991). The duration of the lamina 4 sink is longer than the MUA burst in TC fibers, indicating that the initial lamina 4 sink also represents, in part, activation of cortical cells. Elements implicated are lamina 4 stellate cells, which receive monosynaptic input from TC fibers (Smith and Moskowitz 1979). Many of these cells are characterized by an asymmetrical and vertical orientation of their dendritic trees that ascend into lamina 3 (McMillen and Glaser 1982). Because stellate cells in auditory cortex receive TC afferents mainly on dendritic branches (Smith and Moskowitz 1979) concentrated in lower lamina 3 and upper lamina 4 (Jones and Burton 1976), the initial stetlate cell depolarization would lead to an open-field configuration with a sink surrounding the dendrites and a source near the cell bodies, resulting in a surface-negative potential. This cortical activation would temporally overlap the pre-synaptic responses for all but the initial 1-2 msec of the surface N8 component. Similar conclusions have been reached for the initial cortical component of the primate visual evoked response (Schroeder et al. 1991). The monkey N8 component corresponds to the initial polarity-inverting response in human auditory cortex, which is a negative wave with an onset of 8-10 msec and a peak at 13 msec (Liegeois-Chauvel et al. 1991). The short latency of this component is consistent with slightly earlier activity noted within the human medial geniculate (Velasco et al. 1982). Thus, we propose that terminal portions of TC fibers are a generator of this early cortical component in man. We also hypothesize that the initial negative component in the monkey cortical AEP is partially generated by lamina 4 stellate cells. Stellate cells with a vertical dendritic orientation also occur in middle cortical laminae of the human auditory cortex (Seldon 1981). This anatomical configuration is consistent with a vertically oriented, open-field generator situated in the thalamo-recipient zone. These data and the present findings support a role for stellate cells in the generation of the initial auditory cortical component. The intracortical N8 in monkey and N13 in man (Liegeois-Chauvel et al. 1991) appear to represent the initial segment of the earliest scalp-recorded auditory cortical component in man (Seherg and Von Cramon 1986). This scalp-recorded component is a negativity with a peak at 19 msec and an onset at about 10 reset. Inability to completely isolate the initial intracorticai negativity by scalp recordings may be due to several factors. There is temporal overlap of this early cortical activity with negative waves generated within the midbrain (Hashimoto 1982), which may obscure generator identification from scalp recordings. The 300 Hz upper cut-off frequency used for the scalp recordings may be insufficient to fully characterize the wave form of the high frequency intracortical negativity. Further work will be required to resolve this issue. We thank the excellent technical and secretarial assistance of Mona Litwak, Shirley Seto, Chester Freeman and Linda O'Donnell. Supported by Grants DC00657 from NIH, MH06723 from NIMH, and 90-31 from the McDonnell-Pew Program in Cognitive Neuroscience.
M. STEINSCHNEIDER ET AL. References
Hashimoto, I. Auditory evoked potentials from the human midbrain: slow brain stem responses. Electroenceph. clin. Neurophysiol., 1982, 53: 652-657. Jones, E.G. and Burton, H. Areal differences in the laminar distribution of thalamic afferents in cortical fields of the insular, parietal and temporal regions of primates. J. Comp. Neurol., 1976, 168: 197-248. Kulics, A.T. and Cauller, L.J. Cerebral cortical somatosensory evoked responses, multiple unit activity and current source-densities: their interrelationships and significance to somatic sensation as revealed by stimulation of the awake monkey's hand. Exp. Brain Res., 1986, 62: 46-60. Legatt, A.D., Arezzo, J.C. and Vaughan, Jr., H.G. Short-latency auditory evoked potentials in the monkey. II. Intracranial generators. Electroenceph. olin. Neurophysiol., 1986, 64: 53-73. Liegeois-Chauvel, C., Musolino, A. and Chauvel, P. Localization of the primary auditory area in man. Brain, 1991, 114: 139-153. McMillen, N.T. and Glaser, E.M. Morphology and laminar distribution of nonpyramidal neurons in the auditory cortex of the rabbit. J. Comp. Neurol., 1982, 208: 85-106. Mitzdorf, U. Current source-density method and application in cat cerebral cortex: investigations of evoked potentials and EEG phenomena. Physiol. Rev., 1985, 65: 37-100. Muller-Preuss, P. and Mitzdorf, U. Functional anatomy of the inferior colliculus and the auditory cortex: current source density analyses of click-evoked potentials. Hearing Res., 1984, 16: 133142. Pandya, D.N. and Sanides, F. Architectonic parcellation of the temporal operculum in rhesus monkey and its projection pattern. Z. Anat. Entwickl.-Gesch., 1973, 139: 127-161. Scherg, M. and Von Cramon, D. Evoked dipole source potentials of the human auditory cortex. Electroenceph. clin. Neurophysiol., 1986, 65: 344-360. Schroeder, C.E., Tenke, C.E., Givre, S.J., Arezzo, J.C. and Vaughan, Jr., H.G. Striate cortical contribution to the surface-recorded pattern-reversal VEP in the alert monkey. Vision Res., 1991, 31: 1143-1157. Seldon, H.L. Structure of human auditory cortex. I. Cytoarchitectonics and dendritic distributions. Brain Res., 1981, 229: 277-294. Smith, D.E. and Moskowitz, N. Ultrastructure of layer IV of the primary auditory cortex of the squirrel monkey. Neuroscience, 1979, 4: 349-359. Steinschneider, M., Arezzo, J.C. and Vaughan, Jr., H.G. Speech evoked activity in the auditory radiations and cortex of the awake monkey. Brain Res., 1982, 252: 353-365. Velasco, M., Velasco, F., Almanza, X. and Coats, A.C. Subcortical correlates of the auditory brain stem potentials in man: bipolar EEG and multiple unit activity and electrical stimulation. Electroenceph, clin. Neurophysiol., 1982, 53: 133-142.
