JOURNALOF NEUROPHYSIOLOGY Vol. 68, No. 2, August 1992. Printed
in U.S.A.
The Functional Anatomy of Middle-Latency Auditory Evoked Potentials: Thalamocortical Connections SHI DI AND
DANIEL
S. BARTH
Department of Psychology, University of Colorado, Boulder, Colorado 80309-0345, and Mental Health Institute, BeEJing Medical University, Begjing, China SUMMARY
AND
CONCLUSIONS
I. An 8 X &channel microelectrode array was used to map epicortical field potentials from a 4.375 X 4.375mm2 area in the right parietotemporal neocortex of four rats. Potentials were evoked with bilaterally presented click stimuli and with electrical stimulation of the ventral and dorsal divisions of the medial geniculate body. 2. Epicortical responses to click stimuli replicated earlier findings. The responses consisted of a positive-negative biphasic waveform (P 1a and N 1) in the region of primary auditory cortex (area 41) and a positive monophasic waveform (PI b) in the region of secondary auditory cortex (area 36). Two potential patterns, one at the latency of the Nl and the other at the latency of the P 1b, were used to represent activation of cells within areas41and 36. A linear combination of these patterns was sufficient to explain from 90 to 94% of the variance of the evoked potential complex at all latencies. 3. In the same animals, epicortical responses to electrical stimulation of the ventral and dorsal divisions of the medial geniculate body were also localized to areas41 and 36, respectively. A linear combination of potential patterns from these separate stimulation conditions was sufficient to explain from 80 to 93% of the variance of the original click-evoked potential complex at all latencies. 4. These data provide functional evidence for anatomically defined topographical thalamocortical projections to primary and secondary auditory cortex. They suggest that short-latency cortical evoked potentials ( 1O-60 ms poststimulus) are dominated by parallel thalamocortical activation of areas41 and 36.
INTRODUCTION
In the rat, click stimuli evoke a series of extracellular potentials whose sequential amplitude peaks are thought to reflect the activation of both subcortical and cortical populations of cells as afferent information proceeds from the cochlea to the cerebral cortex (Barth and Di 1990, 199 1; Bhargava et al. 1978; Borbely and Hall 1970; Hall and Borbely 1970; Iwasa and Potsic 1982; Knight et al. 1985; Shaw 1988; Teas and Kiang 1964). Small amplitude peaks occurring before - 10 ms poststimulus, the brain stem auditory evoked potentials (BAEP) have been extensively studied in the rat (Chen and Chen 199 1; Church et al. 1984; Funai and Funasaka 1983; Jewett and Roman0 1972; Shaw 1987, 1988) and probably reflect a nearly sequential relay of information through associated brain stem nuclei. In contrast, the neurogenesis of amplitude peaks occurring > 10 ms poststimulus, the middle-latency auditory evoked potentials (MAEP), is not well understood (Bhargava et al. 1978; Borbely and Hall 1970; Hall and Borbely 1970; Iwasa and Potsic 1982; Knight et al. 1985; Mourek et al. 1967).
Although it is generally assumed that these amplitude peaks are the result of afferent information arriving at the primary auditory cortex, the MAEP recorded topographically on the surface of rat temporal cortex is not a simple waveform restricted to a single region but instead appears as a complex waveform; the latency and polarity of amplitude peaks vary systematically with small changes in the surface location of the recording electrode. Such a complex pattern suggests that, unlike the BAEP, each amplitude peak of the MAEP may not simply represent sequential or serial activation of subregions within auditory cortex but may also reflect the parallel processing of auditory information at the cortical level. Recent studies in our laboratory (Barth and Di 1990, 199 1) have demonstrated that by using high-resolution epicortical potential mapping in combination with numerical methods of spatiotemporal analysis, it is possible to identify and separate putative neural generators of the MAEP complex in the region of auditory cortex in the rat. Thus the temporal components of the MAEP may be given functional meaning by relating them to anatomically distinct regions of auditory cortex responsible for their neurogenesis. This work has suggested that the earliest components of the MAEP may be produced by parallel activation of two adjacent regions of auditory cortex. One region is distinguished by an early positive-negative sharp wave complex that corresponds approximately with the location and extent of primary auditory cortex [area 41 (Krieg 1946)]. The other region is distinguished by a longer-latency positive sharp wave that corresponds approximately with the location of secondary auditory cortex [area 36 (Krieg 1946)]. Although these results are based on multivariate time series analysis and are therefore not necessarily related to underlying physiology, the hypothesis that both areas 41 and 36 are activated in parallel during the MAEP complex receives anatomic support from the demonstrated presence of separate thalamocortical fiber paths projecting from the ventral and dorsal divisions of the medial geniculate ( MGv and MGd) to areas 41 and 36, respectively (Patterson 1977 ) . The object of the present study was to test this hypothesis by comparing the click-evoked epicortical MAEP complex with potentials evoked by direct electrical stimulation of the MGv and MGd. MATERIALS
AND
METHODS
Four adult Sprague-Dawley rats (280-300 g) were anesthetized with a combination of ketamine H,Cl (66 mg/ kg) and xylazine ( 13 mg/kg) followed by atropine sulfate (0.6 mg/kg). During
OO22-3077192 $2.00 Copyright 0 1992 The American Physiological Society
425
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
S. DI AND D. S. BARTH subsequent experimentation, anesthetic levels were maintained so that the cornea1 reflex could barely be elicited. After the initial anesthesia, animals were placed in a stereotaxic device with the incisor bar set at -3.3 mm. A unilateral craniotomy exposed a wide region over the parietotemporal cortex in the right hemisphere from bregma to lambda, extending 12 mm lateral to midline. The dura was removed and the cortical surface covered with saline throughout the experiment. Normal body temperature was maintained with the use of a regulated heating pad. Auditory click stimuli were presented bilaterally using high-frequency ribbon speakers (model EAS- lOTH400A) mounted on the stereotaxic instrument 3 cm lateral to each ear and driven with computer-controlled, monophasic square-wave pulses ( 0.3 ms; 0.5 V; AMP1 Master 8 stimulator) that were delivered with an interstimulus interval (ISI) of 500 ms. Because solid ear bars were in place during all experiments, it is assumed that the sound entered the cochlea through both bone conduction and through conduction from the ear bars. Electrical stimuli (biphasic pulses; 0.1 ms per phase; 200 ,uA; AMP1 Master 8 stimulator with ISO-flex constant current source; IS1 = 500 ms) were delivered to medial geniculate nuclei using stainless steel microconcentric bipolar electrodes (25 pm ID and 150 pm OD, respectively; flush tip; Rhodes Medical Instruments model MCE- 100). The concentric electrode generated radially symmetric currents at the stimulation site, resulting in very little electrical artifact. Stimulus artifact was further decreased by alternating the polarity of biphasic pulses on a trial-to-trial basis. The electrode was inserted 5.5 mm caudal to bregma and 3.3 mm lateral to midline in the MGd or MGv (Paxinos and Watson 1982). The deepest point of an electrode track was marked with a focal lesion by passing a DC current (300 PA, 10 s). Electrode positions were histologically verified by examining formalin-fixed, 40 pm frozen sections stained with cresyl violet. Epicortical potentials were mapped using an array of 64 stainless steel electrodes (tip diameter 100 vm) configured in an 8 X 8 electrode matrix with interelectrode distance of 625 pm, covering a 4.375 X 4.375-mm2 area. The array was constructed with a concave shape to fit the curvature of posterior and lateral temporal cortex. The exact location of the array in relation to local vasculature and stereotaxic coordinates was determined in each animal. Recordings were referred to a Ag/AgCl ball electrode mounted in a burrhole drilled in the front bone. Potentials from the 64 channels were simultaneously amplified (Grass 12A5 amplifiers; O.l100 Hz bandpass), digitally sampled ( 1,000 Hz), and averaged (n = 50). Topographical analysis was performed to model the spatial and temporal pattern of epicortical potentials associated with the short-latency sharp waves of MAEP in areas 41 and 36 (Barth and Di 1990, 199 1) . The spatial patterns of the MAEP at the latency of the first positive amplitude peak in area 36 and the first negative amplitude peak in area 41 were identified. These latencies were chosen because they best reflected the unique potential distributions of the separate areas. At most other latencies of the MAEP complex, areas 41 and 36 were simultaneously active. Multiple linear regression was then used to derive a weighted combination of these two selected spatial patterns that most closely fit the composite potential distribution at other latencies of the MAEP, yielding a pair of weights at each latency of the sharp wave complex. A physical interpretation of this model is that the two spatial patterns represent topographical distributions of potentials that are produced at a moment in time when only area 41 or 36 are active in isolation. At latencies of the MAEP where this is the case, the regression weights are nonzero for one pattern and zero for the other. However, at most latencies of the MAEP, where both areas 41 and 36 are active, but to differing extents, the regression weights reflect the separate contributions of each pattern to the evoked response. Because the potential patterns are normalized before analysis, the regression weights also reflect the absolute am-
plitude and polarity of potentials in areas 41 and 36 throughout the time course of the MAEP. These results were compared with those obtained by modeling the same physiologically evoked data using two epicortical potential patterns derived separately from electrical stimulation of MGd and MGv. RESULTS
Figure IA shows the location of the 8 X &channel electrode array (0) in relation to primary (area 41; darker shadow) and secondary (areas 36 and 20; lighter shadow) auditory cortex. The location and spatial extent of these regions was estimated from previous histological studies (Barth and Di 199 1; Krieg 1946; Patterson 1977; Zilles 1990). The epicortical MAEP had a total spatial distribution covering -4 X 6 mm2 of lateral cortex, determined from multiple placements of the electrode array. The A
B -
-
x
FIG. 1. A: middle-latency auditory evoked potentials (MAEP) were mapped in the region of primary (area 41, dark shaded) and secondary (areas 36 and 20, light shaded) auditory cortex as estimated from stereotaxic and vascular landmarks. An 8 X 8 electrode array (interelectrode distance of 625 pm) covered a 4.375 x 4.375mm2 area (0). B: typical MAEP complex (shown here for rat 1) consisted of a positive-negative biphasic wave in the approximate region of area 41 (bottom right box) and a positive monophasic wave at area 36 (top left box). C: monophasic wave of area 36 (- - -) was consistently of 5- 10 ms longer poststimulus latency than the response from area 41 (--). The amplitude peaks are labeled P 1a, P 1b, and N 1, reflecting polarity and sequence of occurrence. Calibrations: horizontal = 75 and 25 ms (B and C); vertical = .45 and .2 mV (B and C).
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
THALAMOCORTICAL
CONNECTIONS
X
X
FK+. 2. A: normalized isopotential maps reflect the spatial distribution of potentials at the latencies of the Pl b (top) and Nl (bottom) of rut 1. Only the top 50% amplitude peaks of the amplitude distributions are shown. In this model, the maps are taken to represent spatial distributions of cortical generators of the P 1b and N 1. They were centered over areas 36 and 41, respectively. B: regression weights reflect the relative activity of subpopulations of cells represented by each spatial pattern at all latencies of the middle-latency auditory evoked potential (MAEP) complex. C: spatial and temporal activity of each putative generator is reconstructed by multiplying the spatial pattern of potentials for each generator by its temporal pattern of regression weights. The sum of these 2 reconstructions closely approximates the original MAEP complex.
MAEP for a single placement of the array centered on auditory cortex is shown for rut I in Fig. 1B. Consistent with previous studies (Barth and Di 1990, 199 1) , at rostra1 and ventral recording sites covering area 41 the MAEP was
AND THE MAEP
427
characterized by a biphasic sharp wave (~ Fig. 1C) . Amplitude peaks of this wave are labeled P 1a and N 1 to reflect their polarity and sequence of occurrence. With more cauda1 and dorsal electrode locations covering area 36, the MAEP consisted primarily of an initial sharp positive wave (- - - Fig. 1C) labeled PI b. In all animals, the latency of the Plb in area 36 was 5-10 ms longer than the Plain area 41. The latency at which the Nl returned to baseline defined the end of the sharp wave complex of the MAEP. All analysis was restricted to data segments between the moment of stimulation and the end of sharp wave complex, which varied from 40 to 60 ms poststimulus across animals. Figure 2A shows topographic maps of normalized epicortical potentials of the P 1b and N 1 components in rat 1. These patterns were used to best approximate the spatial distribution of potentials uniquely associated with areas 36 and 41. Regression weights reflecting the relative contribution of each spatial pattern at all latencies of the MAEP complex are shown in Fig. 2 B. A spatial and temporal model of the separate components was reconstructed by multiplying the topographic maps by their corresponding regression weights (Fig. 2C). The reconstructions reflect unique activity in area 36 (Fig. 2C, top) and area 41 (Fig. 2C, middle) during the MAEP. The sum of these reconstructions (Fig. ZC, bottom) closely approximated the original MAEP complex in this animal, accounting for 9 1% of the variance. Similar reconstructions accounted for 92,94, and 90% of the variance in the other three animals studied. After recording of the click-evoked MAEP, epicortical potentials were evoked by electrical stimulation of first the MGd and then MGv (Figs. 3 and 4 ) . Stimulation sites were histologically verified and were in the center of these two nuclei (Fig. 3). In all animals, the electrically evoked response differed from the MAEP in that it began at a shorter poststimulus latency, was of a simple biphasic (positivenegative) morphology at most areas of the recording array, and was of largest amplitude at electrode sites over the ap-
electrode track
FZ. 3. Formalin-fixed 40-pm frozen section stained with cresyl violet (5.5 mm caudal to bregma) shows sites used for electrical stimulation of the medial geniculate body (MGB). A: electrode was inserted 5.5 mm caudal to bregma and 3.3 mm lateral to midline. Band C: locations of the dorsal (light shading) and ventral (dark shading) divisions of the MGB (Paxinos and Watson 1982). D: deepest point of an electrode track was marked with a focal lesion by passing a DC current.
lmm Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
428
S. DI AND D. S. BARTH
-
evoked responses were used to model the original MAEP complex, the regression weights (Fig. 5B, dark traces) were similar in amplitude, latency, and morphology to those computed using the original maps for the P 1a and N 1 (Fig. A
-
Rat 1
FIG. 4. A : electrical stimulation was performed separately in the dorsal (light shaded) and ventral (dark shaded) medial geniculate body (MGB). B: dorsal stimulation resulted in an epicortical potential distribution centered in the approximate region of area 36. The initial positive component of the evoked response was most focal. The superimposed isopotential map represents the top 50% amplitude peaks of this component. C: same as B, but depicting the result of ventral stimulation, with an epicortical potential distribution centered in area 41. In both stimulation conditions, the later negative component was less focal and typically extended to both areas 36 and 41. Calibration: horizontal = 75 ms; vertical = .45 mV.
proximate region of areas 36 or 41 during stimulation of the MGd (Fig. 4B) or MGv (Fig. 4C), respectively. In Fig. 4, B and C, isocontour maps reflecting the peak 50% amplitude distribution of the earliest positive component of the electrically evoked response are superimposed on the response complex. Here it may seen that the early positive component was the most focal and most closely associated with the locations of areas 36 and 41. Although the second large negative wave was also approximately centered on the same region, it was more widely distributed throughout both areas 41 and 36. Figure 5A shows topographical maps of the first positive amplitude peak of epicortical potentials evoked by stimulation of the MGd (top map for each animal) and MGv (bottom map for each animal) across animals. As in Fig. 4, B and C, these isocontour maps represent the peak 50% amplitude distribution of the chosen components. The isocontour maps are superimposed on shaded regions reflecting the location and distribution of the 50% amplitude for the PI b and Nl of the MAEP complex. In all animals, there was a close correspondence between the potential maps derived from selective stimulation of the medial geniculate nucleus and maps derived from the PI b and Nl of the MAEP. Stimulation of the MGd produced maximum responses in the region of area 36, similar to the spatial distribution of the Plb, and the response to MGv stimulation was of greatest amplitude near area 41, similar to the distribution of the Nl. When maps derived from thalamically
Rat 3
I
I
I
Rat 4
FIG. 5. A : for each of the 4 animals studied, isopotential maps for the top 50% amplitude peak of the initial positive component of the epicortical potential (solid contour lines) evoked with dorsal medial geniculate (top maps for each rat) and ventral medial geniculate (bottom maps for each rat) stimulation are superimposed on the 50% amplitude distribution (shaded area) for the PI b (top maps for each rat) and N 1 (bottom maps for each rat) of the click-evoked response. B: regression weights, reflecting the contribution of each spatial potential pattern to the click-evoked response at all latencies, are shown for 2 modeling conditions. Light traces are the weights resulting when maps for the Pla and N 1 are used in the model, and dark traces are the weights resulting when maps for dorsal and ventral geniculate stimulation are used to model the same data.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
THALAMOCORTICAL
CONNECTIONS
AND THE MAEP
429
5B, light traces). The regression model using maps derived from MGv and MGd stimulation accounted for 8 1, 8893, and 80% of the variance of the MAEP complex in the separate animals.
topographical pattern of the surface responses evoked separately from areas 36 and 41 during electrical stimulation are quite similar to those derived from the MAEP complex evoked by physiological stimulation. It has now been well established that at least the projections of MGv to area 41 are spatially organized on the basis of sound frequency DISCUSSION ( Sally and Kelly 1988 ) and binaural response properties (Kelly and Sally 1988 ) . Thus the spatial distribution of reThe auditory cortex of the rat consists of several distinct cytoarchitectonic zones situated in the posterolateral neo- sponses in this area depend directly on the nature of the of the cortex. On the basis of cytoarchitectural differences and sound stimulus. In contrast, electrical stimulation patterns of thalamocortical connections, Patterson ( 1977 ) MGv at the current strengths used in this study would be subdivided the auditory cortex into an area characterized expected to excite most of the cells in the nucleus, producby a proliferation of granular cells in layer IV, which he ing a widespread response at the cortical surface. A likely termed “core cortex,” and a surrounding, less granular re- explanation for the similarity between physiologically and electrically evoked responses recorded here is that the click -gion principally on the lateral and caudal borders, termed stimulus is of sufficiently wide band to excite a majority of “belt cortex.” The core-belt regions correspond approximately with primary-secondary cortex or area 41-36 and the frequency specific cells in the MGv, thus approximating the effect of direct electrical stimulation. 20 of Kreig ( 1946) or Tel-Te2 and Te3 of Zilles ( 1990). Regression analysis demonstrates that either set of potenUsing the method of Fink and Heimer ( 1967 ) for staining tial patterns in areas 41 and 36, electrically or physiologidegenerating axons and terminals, and retrograde axonal transport of horseradish peroxide (HRP), it has been demcally evoked, may be used in a weighted linear combination onstrated that the ventral division of the medial geniculate to explain most of the variance of the short-latency MAEP body (MGB) projects almost exclusively to core cortex, the complex. This result has several implications for the manshell nuclei consisting of the dorsal division, the caudal di- ner in which auditory cortex is activated during the MAEP. vision, the ventrolateral nucleus, and the marginal zone First, it suggests that activation does not progressively project topographically to the belt cortex, and the medial spread throughout either cortical region. This event would division projects diffusely to core cortex and possibly re- require that the topographical pattern of epicortical potengions of belt cortex as well. The present data suggest that tials change continuously as a function of time. Initial examthese anatomic distinctions between core and belt cortex ination of the MAEP complex might imply such a continuare directly reflected in the neurogenesis of spatially over- ous spread of activation. The positive wave of the MAEP lapping short-latency components of the MAEP complex. peaks earliest in area 41 (Pla) and appears to peak at proThe first cortical response to either auditory click stimulagressively longer latencies with electrode locations outside tion or direct electrical stimulation of the MGB is a brief of area 41, in the direction of area 36 (PI b). However, our positive epicortical potential. Previous laminar studies of results indicate that this apparently continuous spread of the MAEP in rat area 41 have suggested that this early posi- activity may be adequately modeled by the asynchronous activation of cells of fixed spatial distributions within areas tivity is the result of deep depolarization of the supragranular pyramidal cells, resulting in a passive current source and 41 and 36. Thus the waveform and latency of the MAEP at positive extracellular potentials at the cortical surface a given electrode site is determined by a linear combination (Barth and Di 1990). The initial positive component of the of potentials from these two stationary regions. Second, evoked potential therefore probably reflects the initial excit- these results support a model in which the earliest phases of atory response of the middle cortical layers, mediated either information processing in primary and secondary auditory directly through monosynaptic connections of thalamocorcortex are conducted essentially in parallel. Although the tical fibers at proximal regions of the supragranular pyramiPla of area 41 peaks earlier than the PI b of area 36, the dal cells or indirectly through similar excitatory connec- initial activation of both regions is simultaneous. The retions on layer IV stellate cells. This conclusion is in good sults of MGv and MGd stimulation indicate that areas 41 agreement with the known terminations of fibers from the and 36 may be activated independently through parallel dorsal and ventral divisions of the MGB in layer IV and the thalamocortical projections. These same projections are lower part of layer III (Kelly 1990; Patterson 1977; Ryugo probably responsible for the simultaneous activation of and Killackey 1974; Vaughan 1983 ) . Spatial distributions areas 41 and 36 at the onset of the MAEP. Finally, the Pla of the early positive response therefore most likely repre- and Nl have the same spatial distribution at the cortical sent the location and extent of populations of cells within surface, suggesting that both components are generated by auditory cortex receiving direct thalamocortical afferent the same, or at least completely overlapping, populations of fibers. cells within area 41. This conclusion is in concordance with Epicortical distributions of the early positive response to earlier laminar studies of the MAEP in area 41 (Barth and electrical stimulation of the MGB conform closely to the Di 1990) suggesting that the Pla and Nl primarily reflect parallel connections between this nucleus and core and belt sequential activation of supragranular and infragranular cortex. With MGv stimulation, the cortical response is con- cells along the vertical axis. centrated in the rostra1 and ventral region of the recording Our conclusion that the earliest-latency components of array, in the approximate location of area 41. Stimulation the MAEP complex are produced by parallel activation of of the MGd results in a systematic shift of the surface posi- stationary cell populations within areas 41 and 36 should tive response to a more caudal and dorsal location of the not imply that information processing within these two rearray, centered on the approximate region of area 36. The gions is conducted separately, in the absence of influence Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
430
S. DI AND D. S. BARTH
from interconnections between the parallel pathways. Recent studies using anatomic tracing techniques have demonstrated extensive corticocortical connections between areas 41 and 36 (Kolb 1990). Although differences in the polarity, morphology, and timing of epicortical responses evoked by physiological compared with electrical stimulation may be due in large part to differences in the stimulus properties, they may also provide at least indirect evidence for the influence of these intracortical connections on the neurogenesis of the MAEP. The epicortical response to electrical stimulation of either the MGd or MGv is typically biphasic, with an initial brief positive component followed by a longer and larger-amplitude negative component. The early positive component is spatially constrained and centered over either area 41 or 36. However, the later negative component is more widely distributed, often spreading to both areas 41 and 36 regardless of whether MGd or MGv is stimulated. In contrast, the negative wave (N 1) of the physiologically evoked MAEP is consistently restricted to area 41 and is absent or greatly attenuated in area 36. The spread of negativity during thalamic stimulation may represent the influence of horizontal corticocortical fibers unmasked by activation of area 41 or 36 in isolation. Similarly, in no animals was it possible to evoke a monophasic positive response characteristic of the Pl b in area 36 using thalamic stimulation. The absence of a distinct biphasic response in area 36 during the MAEP may be due to the influence of intracortical connections with area 41. Finally, during stimulation of either MGv or MGd, the initial positive potential evoked at the cortex was of the same latency. If the MAEP were due only to parallel thalamocortical activation of areas 41 and 36, the amplitude peaks of the Pla and P 1b would also be expected to be of the same latency. This expectation is based on the assumption that the MGv and MGd are activated simultaneously during physiological stimulation, an assumption that has not been verified. The substantial delay between these components during physiological stimulation may reflect the influence of intracortical connections when areas 41 and 36 are simultaneously active. However, evidence for intracortical interactions between parallel pathways during the MAEP is indirect and may not be clearly disassociated from possible concurrent interactions within the MGB. In contrast to the extensive intracortical connections between areas 41 and 36, intrathalamic connections between the MGv and MGd in the rat have not been reported. Furthermore, the present demonstration that selective stimulation of either nucleus evokes distinct potentials at the cortex suggests that interactions between the MGv and MGd may not have a strong influence on the MAEP. Yet it cannot be stated with certainty that electrical stimulation in this study was restricted to the MGv or MGd with no overlap. The coaxial design of the stimulating electrode constrained the current density to a small area and resulted in an abrupt transition from no evoked response to a large response at the dorsal and ventral borders of the MGB. Within the MGB, no abrupt transition in the epicortical evoked potential pattern was found as the stimulating electrode was progressively lowered through the MGd and MGv. The final stimulating locations used were in the approximate center of each subdivision. The spread of stimu-
lating current between the MGv and MGd cannot be ruled out and may have contributed to both the spread of negativity between areas 41 and 36 and the inability to evoke a monophasic positive response in area 36 during electrical stimulation. These data, derived from topographical recording of extracellular potentials evoked by physiological or electrical stimulation of the MGB, provide functional evidence for anatomically defined topographical thalamocortical projections. In contrast to the BAEP, which is thought to be produced by the serial activation of nuclei within the brain stem, the present results suggest that the click-evoked MAEP complex is produced by parallel asynchronous activity in areas 41 and 36. Although this activity is simultaneously initiated through parallel thalamocortical projections from the MGv and MGd, the morphology, timing, and polarity of the epicortical MAEP may also be influenced by intracortical connections between areas 41 and 36. Further studies are now needed, using simultaneous stimulation of MGv and MGd and lesions of primary and secondary auditory cortex, to further elucidate the contribution of thalamocortical and corticocortical connections to the spatial and temporal distribution of the MAEP complex during auditory information processing. This research was supported by United States Public Health Service Grant l-R0 1-NS22575, National Science Foundation Grants BNS-8657764 and IBN 9 119525, Whitaker Foundation Grant S880620, and a Grants in Aid from the Graduate School Council on Research and Creative Work at the University of Colorado at Boulder. Address for reprint requests: D. S. Barth, Dept. of Psychology, Campus Box 345, University of Colorado, Boulder, CO 80309-0345. Received 9 January 1992; accepted in final form 8 April 1992. REFERENCES D. S. AND DI, S. Three dimensional analysis of auditory evoked potentials in rat neocortex. J. NeurophysioZ. 64: 1527- 1536, 1990. BARTH, D. S. AND DI, S. The functional anatomy of auditory evoked potentials in rat neocortex. Brain Res. 565: 109-l 15, 199 1. BHARGAVA, V. &SALAMY, A., ANDMCKEAN, C.M.Effectsofcholinergic drugs on auditory evoked responses (CER) of rat cortex. Neuropharmacology 17: 1009-1013, 1978. BORB~LY, A. A. AND HALL, R. D. Effects of pentobarbitone and chlorpromazine on acoustically evoked potentials in the rat. Neuropharmacology 9: 575-586, 1970. CHEN, T. J. AND CHEN, S. S. Generator study of brainstem auditory evoked potentials by radiofrequency lesion methods in rats. Exp. Brain Res. 85: 537-542, 1991. CHURCH, M. W., WILLIAMS, H. L., AND HOLLOWAY, J. A. Brain-stem auditory evoked potentials in the rat: effects of gender, stimulus characteristics and ethanol sedation. Electroencephalogr. Clin. Neurophysiol. 59: 328-339, 1984. FINK, R. P. AND HEIMER, L. Two methods for selective silver impregnation of degenerating axons and their synaptic endings in the central nervous system. Brain Res. 4: 369-374, 1967. FUNAI, H. AND FUNASAKA, S. Experimental study on the effect of inferior colliculus lesions upon auditory brainstem response. Audiology 22: 9-19, 1983. HALL, R. D. AND BORB~LY, A. A. Acoustically evoked potentials in the rat during sleep and waking. Exp. Brain Res. 11: 93-l 10, 1970. IWASA, H. AND POTSIC, W. P. Maturational change of early, middle, and late components of the auditory evoked responses in rats. Otolaryngol. Head Neck Surg. 90: 95-102, 1982. JEWETT, D. L. AND ROMANO, M. N. Neonatal development of auditory system potentials averaged from the scalp of the rat cat. Brain Res. 36: 101-l 15. 1972. BARTH,
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
THALAMOCORTICAL
CONNECTIONS
J. B. Rat auditory cortex. In: The Cerebral Cortex of the Rat, edited by B. Kolb and R. C. Tees. Cambridge, MA: MIT Press, 1990, p. 38 l405. KNIGHT, R. T., BRAILOWSKY, S., SCABINI, D., AND SIMPSON, G. V. Surface auditory evoked potentials in the unrestrained rat: component definition. Electroencephalogr. Clin. Neurophysiol. 6 1: 430-439, 1985. KOLB, B. Organization of the neocortex of the rat. In: The Cerebral Cortex of the Rat, edited by B. Kolb and R. C. Tees. Cambridge, MA: MIT Press, 1990, p. 2 l-34. KRIEG, W. J. S. Connections of the cerebral cortex. I. The albino rat. Topography of the cortical areas. J. Camp. Neural. 84: 221-275, 1946. MOUREK, J., HIMWICH, W. A., MYSLIVECEK, J., AND CALLISON, D. A. The role of nutrition in the development of evoked cortical responses in rat. Brain Res. 6: 24 1-25 1, 1967. PATTERSON, H. An Anterograde Degeneration and Retrograde Axonal Transport Study of the Cortical Projections of the Rat Medial Geniculate Body (doctoral dissertation). Boston, MA: Boston Univ. Press, 1977. PAXINOS, G. AND WATSON, C. The Rat Brain in Stereotaxic Coordinates. New York: Academic, 1982. KELLY,
AND THE MAEP
431
D. K. AND KILLACKEY, H. P. Differential telencephalic projections of the medial and ventral divisions of the medial geniculate body of the rat. Brain Res. 82: 173-177, 1974. SALLY, S. L. AND KELLY, J. B. Organization of auditory cortex in the albino rat: sound frequency. J. Neurophysiol. 59: 1627- 1638, 1988a. SALLY, S. L. AND KELLY, J. B. Organization of auditory cortex in the albino rat: Binaural response properties. J. Neurophysiol. 59: 17561769, 1988b. SHAW, N. A. Effects of low pass filtering on the brainstem auditory evoked potential in the rat. Exp. Brain Res. 65: 686-690, 1987. SHAW, N. A. The auditory evoked potential in rat-a review. Prog. Neurobiol. 31: 19-45, 1988. TEAS, D. C. AND KIANG, N. Y. Evoked responses from the auditory cortex. Exp. Neural. 10: 91-119, 1964. VAUGHAN, D. W. Thalamic and callosal connections of the rat auditory cortex. Brain Res. 260: 18 1-189, 1983. ZILLES, K. Anatomy of the neocortex: cytoarchitecture and myeloarchitecture. In: The Cerebral Cortex of the Rat, edited by B. Kolb and R. C. Tees. Cambridge, MA: MIT Press, 1990, p. 77-l 12. RYUGO,
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
in U.S.A.
The Functional Anatomy of Middle-Latency Auditory Evoked Potentials: Thalamocortical Connections SHI DI AND
DANIEL
S. BARTH
Department of Psychology, University of Colorado, Boulder, Colorado 80309-0345, and Mental Health Institute, BeEJing Medical University, Begjing, China SUMMARY
AND
CONCLUSIONS
I. An 8 X &channel microelectrode array was used to map epicortical field potentials from a 4.375 X 4.375mm2 area in the right parietotemporal neocortex of four rats. Potentials were evoked with bilaterally presented click stimuli and with electrical stimulation of the ventral and dorsal divisions of the medial geniculate body. 2. Epicortical responses to click stimuli replicated earlier findings. The responses consisted of a positive-negative biphasic waveform (P 1a and N 1) in the region of primary auditory cortex (area 41) and a positive monophasic waveform (PI b) in the region of secondary auditory cortex (area 36). Two potential patterns, one at the latency of the Nl and the other at the latency of the P 1b, were used to represent activation of cells within areas41and 36. A linear combination of these patterns was sufficient to explain from 90 to 94% of the variance of the evoked potential complex at all latencies. 3. In the same animals, epicortical responses to electrical stimulation of the ventral and dorsal divisions of the medial geniculate body were also localized to areas41 and 36, respectively. A linear combination of potential patterns from these separate stimulation conditions was sufficient to explain from 80 to 93% of the variance of the original click-evoked potential complex at all latencies. 4. These data provide functional evidence for anatomically defined topographical thalamocortical projections to primary and secondary auditory cortex. They suggest that short-latency cortical evoked potentials ( 1O-60 ms poststimulus) are dominated by parallel thalamocortical activation of areas41 and 36.
INTRODUCTION
In the rat, click stimuli evoke a series of extracellular potentials whose sequential amplitude peaks are thought to reflect the activation of both subcortical and cortical populations of cells as afferent information proceeds from the cochlea to the cerebral cortex (Barth and Di 1990, 199 1; Bhargava et al. 1978; Borbely and Hall 1970; Hall and Borbely 1970; Iwasa and Potsic 1982; Knight et al. 1985; Shaw 1988; Teas and Kiang 1964). Small amplitude peaks occurring before - 10 ms poststimulus, the brain stem auditory evoked potentials (BAEP) have been extensively studied in the rat (Chen and Chen 199 1; Church et al. 1984; Funai and Funasaka 1983; Jewett and Roman0 1972; Shaw 1987, 1988) and probably reflect a nearly sequential relay of information through associated brain stem nuclei. In contrast, the neurogenesis of amplitude peaks occurring > 10 ms poststimulus, the middle-latency auditory evoked potentials (MAEP), is not well understood (Bhargava et al. 1978; Borbely and Hall 1970; Hall and Borbely 1970; Iwasa and Potsic 1982; Knight et al. 1985; Mourek et al. 1967).
Although it is generally assumed that these amplitude peaks are the result of afferent information arriving at the primary auditory cortex, the MAEP recorded topographically on the surface of rat temporal cortex is not a simple waveform restricted to a single region but instead appears as a complex waveform; the latency and polarity of amplitude peaks vary systematically with small changes in the surface location of the recording electrode. Such a complex pattern suggests that, unlike the BAEP, each amplitude peak of the MAEP may not simply represent sequential or serial activation of subregions within auditory cortex but may also reflect the parallel processing of auditory information at the cortical level. Recent studies in our laboratory (Barth and Di 1990, 199 1) have demonstrated that by using high-resolution epicortical potential mapping in combination with numerical methods of spatiotemporal analysis, it is possible to identify and separate putative neural generators of the MAEP complex in the region of auditory cortex in the rat. Thus the temporal components of the MAEP may be given functional meaning by relating them to anatomically distinct regions of auditory cortex responsible for their neurogenesis. This work has suggested that the earliest components of the MAEP may be produced by parallel activation of two adjacent regions of auditory cortex. One region is distinguished by an early positive-negative sharp wave complex that corresponds approximately with the location and extent of primary auditory cortex [area 41 (Krieg 1946)]. The other region is distinguished by a longer-latency positive sharp wave that corresponds approximately with the location of secondary auditory cortex [area 36 (Krieg 1946)]. Although these results are based on multivariate time series analysis and are therefore not necessarily related to underlying physiology, the hypothesis that both areas 41 and 36 are activated in parallel during the MAEP complex receives anatomic support from the demonstrated presence of separate thalamocortical fiber paths projecting from the ventral and dorsal divisions of the medial geniculate ( MGv and MGd) to areas 41 and 36, respectively (Patterson 1977 ) . The object of the present study was to test this hypothesis by comparing the click-evoked epicortical MAEP complex with potentials evoked by direct electrical stimulation of the MGv and MGd. MATERIALS
AND
METHODS
Four adult Sprague-Dawley rats (280-300 g) were anesthetized with a combination of ketamine H,Cl (66 mg/ kg) and xylazine ( 13 mg/kg) followed by atropine sulfate (0.6 mg/kg). During
OO22-3077192 $2.00 Copyright 0 1992 The American Physiological Society
425
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
S. DI AND D. S. BARTH subsequent experimentation, anesthetic levels were maintained so that the cornea1 reflex could barely be elicited. After the initial anesthesia, animals were placed in a stereotaxic device with the incisor bar set at -3.3 mm. A unilateral craniotomy exposed a wide region over the parietotemporal cortex in the right hemisphere from bregma to lambda, extending 12 mm lateral to midline. The dura was removed and the cortical surface covered with saline throughout the experiment. Normal body temperature was maintained with the use of a regulated heating pad. Auditory click stimuli were presented bilaterally using high-frequency ribbon speakers (model EAS- lOTH400A) mounted on the stereotaxic instrument 3 cm lateral to each ear and driven with computer-controlled, monophasic square-wave pulses ( 0.3 ms; 0.5 V; AMP1 Master 8 stimulator) that were delivered with an interstimulus interval (ISI) of 500 ms. Because solid ear bars were in place during all experiments, it is assumed that the sound entered the cochlea through both bone conduction and through conduction from the ear bars. Electrical stimuli (biphasic pulses; 0.1 ms per phase; 200 ,uA; AMP1 Master 8 stimulator with ISO-flex constant current source; IS1 = 500 ms) were delivered to medial geniculate nuclei using stainless steel microconcentric bipolar electrodes (25 pm ID and 150 pm OD, respectively; flush tip; Rhodes Medical Instruments model MCE- 100). The concentric electrode generated radially symmetric currents at the stimulation site, resulting in very little electrical artifact. Stimulus artifact was further decreased by alternating the polarity of biphasic pulses on a trial-to-trial basis. The electrode was inserted 5.5 mm caudal to bregma and 3.3 mm lateral to midline in the MGd or MGv (Paxinos and Watson 1982). The deepest point of an electrode track was marked with a focal lesion by passing a DC current (300 PA, 10 s). Electrode positions were histologically verified by examining formalin-fixed, 40 pm frozen sections stained with cresyl violet. Epicortical potentials were mapped using an array of 64 stainless steel electrodes (tip diameter 100 vm) configured in an 8 X 8 electrode matrix with interelectrode distance of 625 pm, covering a 4.375 X 4.375-mm2 area. The array was constructed with a concave shape to fit the curvature of posterior and lateral temporal cortex. The exact location of the array in relation to local vasculature and stereotaxic coordinates was determined in each animal. Recordings were referred to a Ag/AgCl ball electrode mounted in a burrhole drilled in the front bone. Potentials from the 64 channels were simultaneously amplified (Grass 12A5 amplifiers; O.l100 Hz bandpass), digitally sampled ( 1,000 Hz), and averaged (n = 50). Topographical analysis was performed to model the spatial and temporal pattern of epicortical potentials associated with the short-latency sharp waves of MAEP in areas 41 and 36 (Barth and Di 1990, 199 1) . The spatial patterns of the MAEP at the latency of the first positive amplitude peak in area 36 and the first negative amplitude peak in area 41 were identified. These latencies were chosen because they best reflected the unique potential distributions of the separate areas. At most other latencies of the MAEP complex, areas 41 and 36 were simultaneously active. Multiple linear regression was then used to derive a weighted combination of these two selected spatial patterns that most closely fit the composite potential distribution at other latencies of the MAEP, yielding a pair of weights at each latency of the sharp wave complex. A physical interpretation of this model is that the two spatial patterns represent topographical distributions of potentials that are produced at a moment in time when only area 41 or 36 are active in isolation. At latencies of the MAEP where this is the case, the regression weights are nonzero for one pattern and zero for the other. However, at most latencies of the MAEP, where both areas 41 and 36 are active, but to differing extents, the regression weights reflect the separate contributions of each pattern to the evoked response. Because the potential patterns are normalized before analysis, the regression weights also reflect the absolute am-
plitude and polarity of potentials in areas 41 and 36 throughout the time course of the MAEP. These results were compared with those obtained by modeling the same physiologically evoked data using two epicortical potential patterns derived separately from electrical stimulation of MGd and MGv. RESULTS
Figure IA shows the location of the 8 X &channel electrode array (0) in relation to primary (area 41; darker shadow) and secondary (areas 36 and 20; lighter shadow) auditory cortex. The location and spatial extent of these regions was estimated from previous histological studies (Barth and Di 199 1; Krieg 1946; Patterson 1977; Zilles 1990). The epicortical MAEP had a total spatial distribution covering -4 X 6 mm2 of lateral cortex, determined from multiple placements of the electrode array. The A
B -
-
x
FIG. 1. A: middle-latency auditory evoked potentials (MAEP) were mapped in the region of primary (area 41, dark shaded) and secondary (areas 36 and 20, light shaded) auditory cortex as estimated from stereotaxic and vascular landmarks. An 8 X 8 electrode array (interelectrode distance of 625 pm) covered a 4.375 x 4.375mm2 area (0). B: typical MAEP complex (shown here for rat 1) consisted of a positive-negative biphasic wave in the approximate region of area 41 (bottom right box) and a positive monophasic wave at area 36 (top left box). C: monophasic wave of area 36 (- - -) was consistently of 5- 10 ms longer poststimulus latency than the response from area 41 (--). The amplitude peaks are labeled P 1a, P 1b, and N 1, reflecting polarity and sequence of occurrence. Calibrations: horizontal = 75 and 25 ms (B and C); vertical = .45 and .2 mV (B and C).
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
THALAMOCORTICAL
CONNECTIONS
X
X
FK+. 2. A: normalized isopotential maps reflect the spatial distribution of potentials at the latencies of the Pl b (top) and Nl (bottom) of rut 1. Only the top 50% amplitude peaks of the amplitude distributions are shown. In this model, the maps are taken to represent spatial distributions of cortical generators of the P 1b and N 1. They were centered over areas 36 and 41, respectively. B: regression weights reflect the relative activity of subpopulations of cells represented by each spatial pattern at all latencies of the middle-latency auditory evoked potential (MAEP) complex. C: spatial and temporal activity of each putative generator is reconstructed by multiplying the spatial pattern of potentials for each generator by its temporal pattern of regression weights. The sum of these 2 reconstructions closely approximates the original MAEP complex.
MAEP for a single placement of the array centered on auditory cortex is shown for rut I in Fig. 1B. Consistent with previous studies (Barth and Di 1990, 199 1) , at rostra1 and ventral recording sites covering area 41 the MAEP was
AND THE MAEP
427
characterized by a biphasic sharp wave (~ Fig. 1C) . Amplitude peaks of this wave are labeled P 1a and N 1 to reflect their polarity and sequence of occurrence. With more cauda1 and dorsal electrode locations covering area 36, the MAEP consisted primarily of an initial sharp positive wave (- - - Fig. 1C) labeled PI b. In all animals, the latency of the Plb in area 36 was 5-10 ms longer than the Plain area 41. The latency at which the Nl returned to baseline defined the end of the sharp wave complex of the MAEP. All analysis was restricted to data segments between the moment of stimulation and the end of sharp wave complex, which varied from 40 to 60 ms poststimulus across animals. Figure 2A shows topographic maps of normalized epicortical potentials of the P 1b and N 1 components in rat 1. These patterns were used to best approximate the spatial distribution of potentials uniquely associated with areas 36 and 41. Regression weights reflecting the relative contribution of each spatial pattern at all latencies of the MAEP complex are shown in Fig. 2 B. A spatial and temporal model of the separate components was reconstructed by multiplying the topographic maps by their corresponding regression weights (Fig. 2C). The reconstructions reflect unique activity in area 36 (Fig. 2C, top) and area 41 (Fig. 2C, middle) during the MAEP. The sum of these reconstructions (Fig. ZC, bottom) closely approximated the original MAEP complex in this animal, accounting for 9 1% of the variance. Similar reconstructions accounted for 92,94, and 90% of the variance in the other three animals studied. After recording of the click-evoked MAEP, epicortical potentials were evoked by electrical stimulation of first the MGd and then MGv (Figs. 3 and 4 ) . Stimulation sites were histologically verified and were in the center of these two nuclei (Fig. 3). In all animals, the electrically evoked response differed from the MAEP in that it began at a shorter poststimulus latency, was of a simple biphasic (positivenegative) morphology at most areas of the recording array, and was of largest amplitude at electrode sites over the ap-
electrode track
FZ. 3. Formalin-fixed 40-pm frozen section stained with cresyl violet (5.5 mm caudal to bregma) shows sites used for electrical stimulation of the medial geniculate body (MGB). A: electrode was inserted 5.5 mm caudal to bregma and 3.3 mm lateral to midline. Band C: locations of the dorsal (light shading) and ventral (dark shading) divisions of the MGB (Paxinos and Watson 1982). D: deepest point of an electrode track was marked with a focal lesion by passing a DC current.
lmm Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
428
S. DI AND D. S. BARTH
-
evoked responses were used to model the original MAEP complex, the regression weights (Fig. 5B, dark traces) were similar in amplitude, latency, and morphology to those computed using the original maps for the P 1a and N 1 (Fig. A
-
Rat 1
FIG. 4. A : electrical stimulation was performed separately in the dorsal (light shaded) and ventral (dark shaded) medial geniculate body (MGB). B: dorsal stimulation resulted in an epicortical potential distribution centered in the approximate region of area 36. The initial positive component of the evoked response was most focal. The superimposed isopotential map represents the top 50% amplitude peaks of this component. C: same as B, but depicting the result of ventral stimulation, with an epicortical potential distribution centered in area 41. In both stimulation conditions, the later negative component was less focal and typically extended to both areas 36 and 41. Calibration: horizontal = 75 ms; vertical = .45 mV.
proximate region of areas 36 or 41 during stimulation of the MGd (Fig. 4B) or MGv (Fig. 4C), respectively. In Fig. 4, B and C, isocontour maps reflecting the peak 50% amplitude distribution of the earliest positive component of the electrically evoked response are superimposed on the response complex. Here it may seen that the early positive component was the most focal and most closely associated with the locations of areas 36 and 41. Although the second large negative wave was also approximately centered on the same region, it was more widely distributed throughout both areas 41 and 36. Figure 5A shows topographical maps of the first positive amplitude peak of epicortical potentials evoked by stimulation of the MGd (top map for each animal) and MGv (bottom map for each animal) across animals. As in Fig. 4, B and C, these isocontour maps represent the peak 50% amplitude distribution of the chosen components. The isocontour maps are superimposed on shaded regions reflecting the location and distribution of the 50% amplitude for the PI b and Nl of the MAEP complex. In all animals, there was a close correspondence between the potential maps derived from selective stimulation of the medial geniculate nucleus and maps derived from the PI b and Nl of the MAEP. Stimulation of the MGd produced maximum responses in the region of area 36, similar to the spatial distribution of the Plb, and the response to MGv stimulation was of greatest amplitude near area 41, similar to the distribution of the Nl. When maps derived from thalamically
Rat 3
I
I
I
Rat 4
FIG. 5. A : for each of the 4 animals studied, isopotential maps for the top 50% amplitude peak of the initial positive component of the epicortical potential (solid contour lines) evoked with dorsal medial geniculate (top maps for each rat) and ventral medial geniculate (bottom maps for each rat) stimulation are superimposed on the 50% amplitude distribution (shaded area) for the PI b (top maps for each rat) and N 1 (bottom maps for each rat) of the click-evoked response. B: regression weights, reflecting the contribution of each spatial potential pattern to the click-evoked response at all latencies, are shown for 2 modeling conditions. Light traces are the weights resulting when maps for the Pla and N 1 are used in the model, and dark traces are the weights resulting when maps for dorsal and ventral geniculate stimulation are used to model the same data.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
THALAMOCORTICAL
CONNECTIONS
AND THE MAEP
429
5B, light traces). The regression model using maps derived from MGv and MGd stimulation accounted for 8 1, 8893, and 80% of the variance of the MAEP complex in the separate animals.
topographical pattern of the surface responses evoked separately from areas 36 and 41 during electrical stimulation are quite similar to those derived from the MAEP complex evoked by physiological stimulation. It has now been well established that at least the projections of MGv to area 41 are spatially organized on the basis of sound frequency DISCUSSION ( Sally and Kelly 1988 ) and binaural response properties (Kelly and Sally 1988 ) . Thus the spatial distribution of reThe auditory cortex of the rat consists of several distinct cytoarchitectonic zones situated in the posterolateral neo- sponses in this area depend directly on the nature of the of the cortex. On the basis of cytoarchitectural differences and sound stimulus. In contrast, electrical stimulation patterns of thalamocortical connections, Patterson ( 1977 ) MGv at the current strengths used in this study would be subdivided the auditory cortex into an area characterized expected to excite most of the cells in the nucleus, producby a proliferation of granular cells in layer IV, which he ing a widespread response at the cortical surface. A likely termed “core cortex,” and a surrounding, less granular re- explanation for the similarity between physiologically and electrically evoked responses recorded here is that the click -gion principally on the lateral and caudal borders, termed stimulus is of sufficiently wide band to excite a majority of “belt cortex.” The core-belt regions correspond approximately with primary-secondary cortex or area 41-36 and the frequency specific cells in the MGv, thus approximating the effect of direct electrical stimulation. 20 of Kreig ( 1946) or Tel-Te2 and Te3 of Zilles ( 1990). Regression analysis demonstrates that either set of potenUsing the method of Fink and Heimer ( 1967 ) for staining tial patterns in areas 41 and 36, electrically or physiologidegenerating axons and terminals, and retrograde axonal transport of horseradish peroxide (HRP), it has been demcally evoked, may be used in a weighted linear combination onstrated that the ventral division of the medial geniculate to explain most of the variance of the short-latency MAEP body (MGB) projects almost exclusively to core cortex, the complex. This result has several implications for the manshell nuclei consisting of the dorsal division, the caudal di- ner in which auditory cortex is activated during the MAEP. vision, the ventrolateral nucleus, and the marginal zone First, it suggests that activation does not progressively project topographically to the belt cortex, and the medial spread throughout either cortical region. This event would division projects diffusely to core cortex and possibly re- require that the topographical pattern of epicortical potengions of belt cortex as well. The present data suggest that tials change continuously as a function of time. Initial examthese anatomic distinctions between core and belt cortex ination of the MAEP complex might imply such a continuare directly reflected in the neurogenesis of spatially over- ous spread of activation. The positive wave of the MAEP lapping short-latency components of the MAEP complex. peaks earliest in area 41 (Pla) and appears to peak at proThe first cortical response to either auditory click stimulagressively longer latencies with electrode locations outside tion or direct electrical stimulation of the MGB is a brief of area 41, in the direction of area 36 (PI b). However, our positive epicortical potential. Previous laminar studies of results indicate that this apparently continuous spread of the MAEP in rat area 41 have suggested that this early posi- activity may be adequately modeled by the asynchronous activation of cells of fixed spatial distributions within areas tivity is the result of deep depolarization of the supragranular pyramidal cells, resulting in a passive current source and 41 and 36. Thus the waveform and latency of the MAEP at positive extracellular potentials at the cortical surface a given electrode site is determined by a linear combination (Barth and Di 1990). The initial positive component of the of potentials from these two stationary regions. Second, evoked potential therefore probably reflects the initial excit- these results support a model in which the earliest phases of atory response of the middle cortical layers, mediated either information processing in primary and secondary auditory directly through monosynaptic connections of thalamocorcortex are conducted essentially in parallel. Although the tical fibers at proximal regions of the supragranular pyramiPla of area 41 peaks earlier than the PI b of area 36, the dal cells or indirectly through similar excitatory connec- initial activation of both regions is simultaneous. The retions on layer IV stellate cells. This conclusion is in good sults of MGv and MGd stimulation indicate that areas 41 agreement with the known terminations of fibers from the and 36 may be activated independently through parallel dorsal and ventral divisions of the MGB in layer IV and the thalamocortical projections. These same projections are lower part of layer III (Kelly 1990; Patterson 1977; Ryugo probably responsible for the simultaneous activation of and Killackey 1974; Vaughan 1983 ) . Spatial distributions areas 41 and 36 at the onset of the MAEP. Finally, the Pla of the early positive response therefore most likely repre- and Nl have the same spatial distribution at the cortical sent the location and extent of populations of cells within surface, suggesting that both components are generated by auditory cortex receiving direct thalamocortical afferent the same, or at least completely overlapping, populations of fibers. cells within area 41. This conclusion is in concordance with Epicortical distributions of the early positive response to earlier laminar studies of the MAEP in area 41 (Barth and electrical stimulation of the MGB conform closely to the Di 1990) suggesting that the Pla and Nl primarily reflect parallel connections between this nucleus and core and belt sequential activation of supragranular and infragranular cortex. With MGv stimulation, the cortical response is con- cells along the vertical axis. centrated in the rostra1 and ventral region of the recording Our conclusion that the earliest-latency components of array, in the approximate location of area 41. Stimulation the MAEP complex are produced by parallel activation of of the MGd results in a systematic shift of the surface posi- stationary cell populations within areas 41 and 36 should tive response to a more caudal and dorsal location of the not imply that information processing within these two rearray, centered on the approximate region of area 36. The gions is conducted separately, in the absence of influence Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
430
S. DI AND D. S. BARTH
from interconnections between the parallel pathways. Recent studies using anatomic tracing techniques have demonstrated extensive corticocortical connections between areas 41 and 36 (Kolb 1990). Although differences in the polarity, morphology, and timing of epicortical responses evoked by physiological compared with electrical stimulation may be due in large part to differences in the stimulus properties, they may also provide at least indirect evidence for the influence of these intracortical connections on the neurogenesis of the MAEP. The epicortical response to electrical stimulation of either the MGd or MGv is typically biphasic, with an initial brief positive component followed by a longer and larger-amplitude negative component. The early positive component is spatially constrained and centered over either area 41 or 36. However, the later negative component is more widely distributed, often spreading to both areas 41 and 36 regardless of whether MGd or MGv is stimulated. In contrast, the negative wave (N 1) of the physiologically evoked MAEP is consistently restricted to area 41 and is absent or greatly attenuated in area 36. The spread of negativity during thalamic stimulation may represent the influence of horizontal corticocortical fibers unmasked by activation of area 41 or 36 in isolation. Similarly, in no animals was it possible to evoke a monophasic positive response characteristic of the Pl b in area 36 using thalamic stimulation. The absence of a distinct biphasic response in area 36 during the MAEP may be due to the influence of intracortical connections with area 41. Finally, during stimulation of either MGv or MGd, the initial positive potential evoked at the cortex was of the same latency. If the MAEP were due only to parallel thalamocortical activation of areas 41 and 36, the amplitude peaks of the Pla and P 1b would also be expected to be of the same latency. This expectation is based on the assumption that the MGv and MGd are activated simultaneously during physiological stimulation, an assumption that has not been verified. The substantial delay between these components during physiological stimulation may reflect the influence of intracortical connections when areas 41 and 36 are simultaneously active. However, evidence for intracortical interactions between parallel pathways during the MAEP is indirect and may not be clearly disassociated from possible concurrent interactions within the MGB. In contrast to the extensive intracortical connections between areas 41 and 36, intrathalamic connections between the MGv and MGd in the rat have not been reported. Furthermore, the present demonstration that selective stimulation of either nucleus evokes distinct potentials at the cortex suggests that interactions between the MGv and MGd may not have a strong influence on the MAEP. Yet it cannot be stated with certainty that electrical stimulation in this study was restricted to the MGv or MGd with no overlap. The coaxial design of the stimulating electrode constrained the current density to a small area and resulted in an abrupt transition from no evoked response to a large response at the dorsal and ventral borders of the MGB. Within the MGB, no abrupt transition in the epicortical evoked potential pattern was found as the stimulating electrode was progressively lowered through the MGd and MGv. The final stimulating locations used were in the approximate center of each subdivision. The spread of stimu-
lating current between the MGv and MGd cannot be ruled out and may have contributed to both the spread of negativity between areas 41 and 36 and the inability to evoke a monophasic positive response in area 36 during electrical stimulation. These data, derived from topographical recording of extracellular potentials evoked by physiological or electrical stimulation of the MGB, provide functional evidence for anatomically defined topographical thalamocortical projections. In contrast to the BAEP, which is thought to be produced by the serial activation of nuclei within the brain stem, the present results suggest that the click-evoked MAEP complex is produced by parallel asynchronous activity in areas 41 and 36. Although this activity is simultaneously initiated through parallel thalamocortical projections from the MGv and MGd, the morphology, timing, and polarity of the epicortical MAEP may also be influenced by intracortical connections between areas 41 and 36. Further studies are now needed, using simultaneous stimulation of MGv and MGd and lesions of primary and secondary auditory cortex, to further elucidate the contribution of thalamocortical and corticocortical connections to the spatial and temporal distribution of the MAEP complex during auditory information processing. This research was supported by United States Public Health Service Grant l-R0 1-NS22575, National Science Foundation Grants BNS-8657764 and IBN 9 119525, Whitaker Foundation Grant S880620, and a Grants in Aid from the Graduate School Council on Research and Creative Work at the University of Colorado at Boulder. Address for reprint requests: D. S. Barth, Dept. of Psychology, Campus Box 345, University of Colorado, Boulder, CO 80309-0345. Received 9 January 1992; accepted in final form 8 April 1992. REFERENCES D. S. AND DI, S. Three dimensional analysis of auditory evoked potentials in rat neocortex. J. NeurophysioZ. 64: 1527- 1536, 1990. BARTH, D. S. AND DI, S. The functional anatomy of auditory evoked potentials in rat neocortex. Brain Res. 565: 109-l 15, 199 1. BHARGAVA, V. &SALAMY, A., ANDMCKEAN, C.M.Effectsofcholinergic drugs on auditory evoked responses (CER) of rat cortex. Neuropharmacology 17: 1009-1013, 1978. BORB~LY, A. A. AND HALL, R. D. Effects of pentobarbitone and chlorpromazine on acoustically evoked potentials in the rat. Neuropharmacology 9: 575-586, 1970. CHEN, T. J. AND CHEN, S. S. Generator study of brainstem auditory evoked potentials by radiofrequency lesion methods in rats. Exp. Brain Res. 85: 537-542, 1991. CHURCH, M. W., WILLIAMS, H. L., AND HOLLOWAY, J. A. Brain-stem auditory evoked potentials in the rat: effects of gender, stimulus characteristics and ethanol sedation. Electroencephalogr. Clin. Neurophysiol. 59: 328-339, 1984. FINK, R. P. AND HEIMER, L. Two methods for selective silver impregnation of degenerating axons and their synaptic endings in the central nervous system. Brain Res. 4: 369-374, 1967. FUNAI, H. AND FUNASAKA, S. Experimental study on the effect of inferior colliculus lesions upon auditory brainstem response. Audiology 22: 9-19, 1983. HALL, R. D. AND BORB~LY, A. A. Acoustically evoked potentials in the rat during sleep and waking. Exp. Brain Res. 11: 93-l 10, 1970. IWASA, H. AND POTSIC, W. P. Maturational change of early, middle, and late components of the auditory evoked responses in rats. Otolaryngol. Head Neck Surg. 90: 95-102, 1982. JEWETT, D. L. AND ROMANO, M. N. Neonatal development of auditory system potentials averaged from the scalp of the rat cat. Brain Res. 36: 101-l 15. 1972. BARTH,
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
THALAMOCORTICAL
CONNECTIONS
J. B. Rat auditory cortex. In: The Cerebral Cortex of the Rat, edited by B. Kolb and R. C. Tees. Cambridge, MA: MIT Press, 1990, p. 38 l405. KNIGHT, R. T., BRAILOWSKY, S., SCABINI, D., AND SIMPSON, G. V. Surface auditory evoked potentials in the unrestrained rat: component definition. Electroencephalogr. Clin. Neurophysiol. 6 1: 430-439, 1985. KOLB, B. Organization of the neocortex of the rat. In: The Cerebral Cortex of the Rat, edited by B. Kolb and R. C. Tees. Cambridge, MA: MIT Press, 1990, p. 2 l-34. KRIEG, W. J. S. Connections of the cerebral cortex. I. The albino rat. Topography of the cortical areas. J. Camp. Neural. 84: 221-275, 1946. MOUREK, J., HIMWICH, W. A., MYSLIVECEK, J., AND CALLISON, D. A. The role of nutrition in the development of evoked cortical responses in rat. Brain Res. 6: 24 1-25 1, 1967. PATTERSON, H. An Anterograde Degeneration and Retrograde Axonal Transport Study of the Cortical Projections of the Rat Medial Geniculate Body (doctoral dissertation). Boston, MA: Boston Univ. Press, 1977. PAXINOS, G. AND WATSON, C. The Rat Brain in Stereotaxic Coordinates. New York: Academic, 1982. KELLY,
AND THE MAEP
431
D. K. AND KILLACKEY, H. P. Differential telencephalic projections of the medial and ventral divisions of the medial geniculate body of the rat. Brain Res. 82: 173-177, 1974. SALLY, S. L. AND KELLY, J. B. Organization of auditory cortex in the albino rat: sound frequency. J. Neurophysiol. 59: 1627- 1638, 1988a. SALLY, S. L. AND KELLY, J. B. Organization of auditory cortex in the albino rat: Binaural response properties. J. Neurophysiol. 59: 17561769, 1988b. SHAW, N. A. Effects of low pass filtering on the brainstem auditory evoked potential in the rat. Exp. Brain Res. 65: 686-690, 1987. SHAW, N. A. The auditory evoked potential in rat-a review. Prog. Neurobiol. 31: 19-45, 1988. TEAS, D. C. AND KIANG, N. Y. Evoked responses from the auditory cortex. Exp. Neural. 10: 91-119, 1964. VAUGHAN, D. W. Thalamic and callosal connections of the rat auditory cortex. Brain Res. 260: 18 1-189, 1983. ZILLES, K. Anatomy of the neocortex: cytoarchitecture and myeloarchitecture. In: The Cerebral Cortex of the Rat, edited by B. Kolb and R. C. Tees. Cambridge, MA: MIT Press, 1990, p. 77-l 12. RYUGO,
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 18, 2019.
