Hearing Research, 61 (1992) 147-154 © 1992 Elsevier Science Publishers B.V. All rights reserved 0378-5955/92/$05.00
147
HEARES 01768
Contributions of medial geniculate body subdivisions to the middle latency response * Therese McGee, Nina Kraus, Thomas Littma.n and Trent Nicol Evoked Potentials Laboratory, Northwestern Unicersity, Et'anston, Illinois, USA (Received 21 October 1991; Revision received 18 March 1992; Accepted 21 March 1992)
Ongoing studies in our laboratory, concerned with identifying the neural pathways responsible for the auditory middle latency response (MLR), have involved analysis of surface and intracranial potentials following pharmacologic inactivation (with lidocaine) of small regions in the guinea pig brain. Previous studies indicate that MLR surface waves recorded over the temporal lobe originate from pathways anatomically distinct from those that generate MLR waves recorded c0:er the midline. The medial geniculate body (MG) contributes to both MLR responses. At issue here are the relative contributions of ventral and caudomedial subdivisions, which have been linked to primary, and non-primary auditory pathways, respectively. Ventral and caudomedial subdivisions contributed to the surface-recorded MLR in a distinctive manner. Lidocaine injections to both areas reduced the amplitude of the surface temporal response. Caudomedial injections had a much greater effect on the surface midline responses than did injections in the ventral portion. Thus, the ventral division, a part of the primary auditory pathway, contributes chiefly to the temporal response. The caudomedial portion, which may be linked to non-primary auditory pathways, contributes to both responses. Auditory middle latency response, Medial geniculate body; Auditory pathway, primary and non-primary
Introduction A growing body of evidence suggests that components of the middle latency response (MLR) arise from a complex neural generating system (Kraus and McGee, 1992, review; McGee et al., 1991). In a general sense, the system inclu~!es the auditory pathway from the midbrain to cortex. Also involved are regions such as the reticular formation and non-primary divisions of the auditory thalamo-cortical pathways, which process multimodal stimuli. In the guinea pig, MLR surface waves recorded over the temporal lobe originate from pathways anatomically distinct from those that generate MLR waves
Correspondence to: Therese McGee, Evoked Potentials Laboratory, Northwestern University~ 2299 Sheridan Road, Evanston, IL 60208, USA. * Parts of this manuscript were presented at the Midwinter Meeting of the Association for Research in Otolaryngology, St. Petersburg, FL, February, 1991, and at the Annual Convention of the American Academy of Audiology, Denver, CO, April, 1991.
Abbreviations: CP cerebral peduncle, LGd dorsal lateral geniculate nucleus, LGv ventral lateral geniculate nucleus, MGcm medial geniculate body, caudomedial division, MGd medial geniculate body, dorsal division, MGv medial geniculate body, ventral division, mRF mesencephalic reticular formation, PVG periventricular gray, RN red nucleus, SC superior colliculus, SN substantia nigra
recorded over the midline. The midline responses develop early, are labile, and are affected by lesions in the reticular formation. Temporal responses develop later, are stable and robust, and are affected by lesions of the auditory thalamo-cortical pathway. Temporal components (waves A, B, C) and midline components (waves M- t, M + , M- 2) differ in response to trauma of the temporal cortex, have different rate functions, differ in their developmental time course, and differ in response to anesthetic agents (Kraus et al., 1988). These observations have led to the hypothesis that temporal and midline responses reflect the activity of separate pathways in the MLR generating system. The division of the generating system into component parts can be used advantageously for identifying the underlying contributing influences. The present study is a continuation of an ongoing series concerned with identifying the MLR generating system in the guinea pig model. The guinea pig is a particularly appropriate model for this type of investigation because, as has been previously demonstrated, in this species the responses from different neural structures are topographically distinct. It has been demonstrated that the primary auditory pathway, including-the medial geniculate body (MG), selectively contributes to the response recorded over the temporal lobe, while non-primary attention-state dependent pathways such as the reticular formation contribute to both the response recorded from the midline and the
148
response recorded over the temporal lobe (Kraus et al., 1988; 1992). Anatomic studies in the cat (Niimi and Matsuoka, 1979; Anderson et al., 1980) have demonstrated that primary auditory cortex receives substantial neural input from the ventral subdivision of the medial geniculate (MGv), while non-primary auditory cortex is innervated by projections from more caudal and medial areas. That the nuclei of the medial geniculate body involve two systems has been a consistent finding, not only in studies of neural connections but also in studies of cell morphology (Winer and Morest, 1983) and single neuron physiologic responses (Calford, 1983). Winer and Morest (1983) cited various descriptive terms for the dichotomy of thalamic nuclei: specific vs. nonspecific, extrinsic vs. intrinsic, core vs. belt. Anderson et al. (1980) applied the distinctively auditory terms of a 'cochleotopic' system and a 'diffuse' system. In addition to the differentiation of subdivisions, Anderson et al. (1980) noted a rostro-caudal gradient in the pattern of connections, with more rostral areas being more closely associated with primary auditory cortex. This is consistent with the observation of Rodrigues-Dagaeff et al. (1989) of a rostro-caudal gradient within the MG, such that more rostral areas are more sharply tuned and selectively auditory-responsive, while caudal areas are more likely to be broadly tuned and multi-sensory. Redies et al. (1989a, b) have described in the guinea pig cortical and thalamic regions comparable to the 'cochleotopic' and 'diffuse' systems described in the eat (Anderson et al., 1980). The thalamocortical projections may be less complex in their tonotopic representation, but they appear to be very similar in the arrangement of connections from MG subdivisions to auditory cortical regions. We hypothesize that the differentiation of the temporal and midline responses is associated respectively with the differentiation of primary and non-primary divisions of the MG. The current study focuses on the role of the medial geniculate body in the generation of the MLR, specifically on how primary and non-primary subdivisions of the MG influence the components of the MLR.
Methods
Subjects Guinea pigs, weighing approximately 350 grams, were used as subjects. Animals were anesthetized with ketamine hydrochloride (100 mg/kg) and xylazine (7 mg/kg), and maintained at a body temperature of 37.5°C (_+1°). Although surgical level anesthesia severely degrades MLR amplitude, within 3 to 4 h responses regain a morphology very similar to that of an awake animal (McGee et al., 1983; Smith and
Kraus, 1987). Smaller doses (15 mg/kg of ketamine; 3 mg/kg of xylazine) were administered as needed for the rest of the experiment to maintain a depth at which the MLR was minimally affected but the pedal withdrawal response was not seen.
Electrophysiologic recording Epidural silver bead electrodes (0.5 mm diameter) were used to record the surface MLR as previously described (Kraus et al., 1988). Recordings were made over the posterior midline and from the temporal lobe contralateral to the stimulated ear. The position of the temporal electrode was approximately over the dorsal portion of primary auditory cortex, as described by Redies et al. (1989b). An electrode placed 15 mm rostral to bregma and 1 mm lateral to the sagittal suture served as a reference. A combination high-impedance depth electrode and micro-injection needle was positioned stereotaxically in the MG, as described by McGee et al. (1991). Coordinates for the MGv were 5.0 mm rostrai to the interaural line, 3.9 mm lateral to the sagittal suture, and 7.5 mm ventral to the surface of the brain. Coordinates for the MGem were 4.5 mm rostral, 3.9 mm lateral, and 7.5 mm ventral. Monaural, 100/~s click stimuli were presented at 70 dB nHL at a rate of 3.5/s through Etymotic insert earphones. Recordings were filtered from 10 to 1500 Hz. Each averaged response consisted of 200 individual responses with a 60 ms time sweep and a 3.25 ms pre-stimulus period. Neural activity was recorded simultaneously from all three channels. The ABR was readily apparent in the first seven milliseconds of the midline recording and was monitored to insure the integrity of the peripheral auditory system. At least ten successive recordings were obtained from all sites prior to the administration of lidocaine. Following this baseline period, 2/~1 of 4% lidocaine HCI in saline was injected into the MG, to reversibly inactivate axonal transmission (Hille, 1966). To control injection rate, a microinfusion pump delivered the lidocaine at a rate of 2/~1/30 s. In the ease of saline control subjects, 0.9% saline with a pH similar to the lidocaine (5.0-7.0)was used. The effect of the drug was observed simultaneously on surface potentials and on the local response, with recordings obtained at intervals of 1-2 min until the return of baseline waveform morphologies and amplitudes. At the end of the experiment, the recording location was marked with electrolytic lesions (35/~A for 10 s). Brains were cut in 17/~ coronal sections and stained with the Kliiver stain which permits visualization of cell body and fiber pathways.
Data analysis Wave amplitudes and latencies before and after lidocaine injection were measured. Amplitude was
149
measured from the preceding trough to the peak, or from the preceding peak to the trough. Thus, the M-] amplitude was measured from wave 4 of the ABR to M- I, the M + amplitude was measured from M- n to M + , etc. Latencies were measured at the midpoint of the peak or trough. All peaks and troughs occurring in a post-stimulu~z latency range of 4 to 40 ms were scored. The means and standard deviations of baseline amplitudes were computed to determine confidence intervals of significant change. A change of more than two standard deviations from the baseline mean was considered significant. To compare lidocaine effe~s across animals, amplitudes were normalized to percentages, with baseline mean defined as 100%. The percentage values were utilized in parametric statistics.
i
A
*
It
a a
C
I a
IOOuV B
W4
Results
Examples of the surface and depth MLR responses are shown in Fig. 1. MLRs recorded from the surface of the temporal cortex (top) exhibit a three component complex, with a positive component at 15 ms (A), a negative component at 27 ms (B), and a positive component at 40 ms (C). The surface midline response (Fig. 1, middle) includes the rapid peaks of the ABR, followed by a negative wave at ]0-12 ms (M-I), a positive wave at 20-25 ms (M + ), and a second negativity (M- 2) at 32-35 ms. The waveform recorded locally from the MG (Fig. 1, bottom) is characterized by a prominent positive wave at roughly 9 ms, and a negativity at 12-17 ms. A more labile broad positive wave at 25-32 ms is sometimes present as well. Lesions were localized in the ventral and caudomedial subdivisions of the M G as described by Redies et al. (1989a). Histological results indicated that ~ajections were successfully made into subdivisions of the medial geniculate in 19 animals: 12 within MGv and 7 within MGcm. Subdivisions of the MG and injection sites are illustrated in the schematic drawings of Fig. 2. The sections are arranged rostral to caudal, spanning a total distance of 2.6 mm through the nucleus. The ventral division (MGv) is characterized by small, homogeneous cells. The caudomedial division (MGcm) is a magnocellular area with numerous fiber pathways. The area just dorsal to the ventral division (MGd) has been called the "shell nucleus' by Redies et al. (!989a), and is distinguished from the MGv on functional grounds.
Effects of pharmacologic inactivation of the MG The injection of ]idocaine (2/~1) in all subdivisions of the MG correlated with the marked reduction of local activity and an immediate disruption of MLR components on the surface of the temporal cortex. Amplitude decrements and latency delays were seen in
V
M- 1
1
10uV
/
M.2
MEDIAL GENICULATE
lb
~
~
tata~ (ms)
4b
~0
~0
Fig. 1. Neural activity recorded from the epidural surface over the temporal lobe (top), the epidural surface over the midline (middle), and the intracTanial electrode in MG (bottom). Responses are plotted posi~,ivityup.
all waves. In some animals, the ABC complex was completely eliminated, in others, the ABC complex was replaced for up to 50 rain by log' amplitude 'transi.tional' waveforms, termed as such because of their marked deviation from baseline morphology. The transitional waveforms often showed a morphology similar to the midline response. Thereafter, a delayed ABC complex typically reappeared, which progressively recovered to baseline amplitude and latency. In some
150
animals, amplitudes recovered to levels greater than baseline. In contrast to the effects on the ABC complex, surfac~ midline changes were dependent on the particular subdivision in which lidocaine was injected. Injections in the caudomedial portion produced major amplitude decrements and latency shifts or completely eliminated the midline response, while injections in the ventral portion of the MG produced much less change in the midline response. Injections in MGv showed a rostro-caudal gradient, with more caudal injections having more effect on midline responses. In Fig. 3 are shown representative waveforms recorded from surface sites before and after lidocaine injection. As described above, injections in ventral Mr3 were associated with disruption of MLR activity obtained over the surface of the temporal cortex. Surface MLRs obtained at the midline were relatively unchanged (top). Injections in the caudo-medial portion of MG were associated with disruption of MLR activity obtained over both the temporal cortex and the midline (bottom).
Amplitude The time course of effects of lidocaine injection on responses recorded from surface midline and temporal sites is shown in Fig. 4 for two animals. The amplitudes of waveforms from these areas were plotted over the course of the experiment. Shown is the initial baseline period, injection at time 0, and recovery to baseline amplitude. Horizontal lines denote a range of the baseline mean + 2 standard deviations. Lidocaine injections in the ventral MG were associated with disruption of MLR activity obtained over the" surface of the temporal cortex, with much less change in the midline response (Fig. 4 top). Lidocaine injections in the caudomedial MG were associated with disruption of MLR activity obtained over both temporal and midline sites (Fig. 4 bottom). Of the 12 animals with histologically confirmed injections in the MGv, 3 showed no change in the midline waveform. Three additional animals showed no change in the amplitude of M-~ and M + , although a reduction was seen in the M- 2 amplitude. The injection sites of these six animals were typically more
Histologic Reconstruction of Injection Sites
.2
3.8
Fig. 2. Schematic of MG with histologic reconstruction of subcortical injection sites: solid dot (.). Numbers in right hand corner of each schematic denote dista~Jce (mm) in a rostral direction from the interaural line.
151 rostral and more lateral than other MGv sites. The other six animals with MCe~ injections showed no change in M-~; moderate reductions in M + and M- 2 amplitude were observed. All seven animals with injections in. MGcm showed severe reductions in M + and M- 2 amplitude post-lidocaine. Absolute response amplitude varied considerably among animals. For analysis purposes, amplitudes were expressed as percent of baseline mean. Of particular interest were the percent amplitudes during the time of maximum effect, 2-12 min post-injection. Fig. 5 shows the mean post-lidocaine amplitudes for each condition during the time of maximum effect. A three-way splitplot A N O V A was performed on these data comparing electrode site x injection site x waves within a response. All three main effects were significant (F = 7.7, P
147
HEARES 01768
Contributions of medial geniculate body subdivisions to the middle latency response * Therese McGee, Nina Kraus, Thomas Littma.n and Trent Nicol Evoked Potentials Laboratory, Northwestern Unicersity, Et'anston, Illinois, USA (Received 21 October 1991; Revision received 18 March 1992; Accepted 21 March 1992)
Ongoing studies in our laboratory, concerned with identifying the neural pathways responsible for the auditory middle latency response (MLR), have involved analysis of surface and intracranial potentials following pharmacologic inactivation (with lidocaine) of small regions in the guinea pig brain. Previous studies indicate that MLR surface waves recorded over the temporal lobe originate from pathways anatomically distinct from those that generate MLR waves recorded c0:er the midline. The medial geniculate body (MG) contributes to both MLR responses. At issue here are the relative contributions of ventral and caudomedial subdivisions, which have been linked to primary, and non-primary auditory pathways, respectively. Ventral and caudomedial subdivisions contributed to the surface-recorded MLR in a distinctive manner. Lidocaine injections to both areas reduced the amplitude of the surface temporal response. Caudomedial injections had a much greater effect on the surface midline responses than did injections in the ventral portion. Thus, the ventral division, a part of the primary auditory pathway, contributes chiefly to the temporal response. The caudomedial portion, which may be linked to non-primary auditory pathways, contributes to both responses. Auditory middle latency response, Medial geniculate body; Auditory pathway, primary and non-primary
Introduction A growing body of evidence suggests that components of the middle latency response (MLR) arise from a complex neural generating system (Kraus and McGee, 1992, review; McGee et al., 1991). In a general sense, the system inclu~!es the auditory pathway from the midbrain to cortex. Also involved are regions such as the reticular formation and non-primary divisions of the auditory thalamo-cortical pathways, which process multimodal stimuli. In the guinea pig, MLR surface waves recorded over the temporal lobe originate from pathways anatomically distinct from those that generate MLR waves
Correspondence to: Therese McGee, Evoked Potentials Laboratory, Northwestern University~ 2299 Sheridan Road, Evanston, IL 60208, USA. * Parts of this manuscript were presented at the Midwinter Meeting of the Association for Research in Otolaryngology, St. Petersburg, FL, February, 1991, and at the Annual Convention of the American Academy of Audiology, Denver, CO, April, 1991.
Abbreviations: CP cerebral peduncle, LGd dorsal lateral geniculate nucleus, LGv ventral lateral geniculate nucleus, MGcm medial geniculate body, caudomedial division, MGd medial geniculate body, dorsal division, MGv medial geniculate body, ventral division, mRF mesencephalic reticular formation, PVG periventricular gray, RN red nucleus, SC superior colliculus, SN substantia nigra
recorded over the midline. The midline responses develop early, are labile, and are affected by lesions in the reticular formation. Temporal responses develop later, are stable and robust, and are affected by lesions of the auditory thalamo-cortical pathway. Temporal components (waves A, B, C) and midline components (waves M- t, M + , M- 2) differ in response to trauma of the temporal cortex, have different rate functions, differ in their developmental time course, and differ in response to anesthetic agents (Kraus et al., 1988). These observations have led to the hypothesis that temporal and midline responses reflect the activity of separate pathways in the MLR generating system. The division of the generating system into component parts can be used advantageously for identifying the underlying contributing influences. The present study is a continuation of an ongoing series concerned with identifying the MLR generating system in the guinea pig model. The guinea pig is a particularly appropriate model for this type of investigation because, as has been previously demonstrated, in this species the responses from different neural structures are topographically distinct. It has been demonstrated that the primary auditory pathway, including-the medial geniculate body (MG), selectively contributes to the response recorded over the temporal lobe, while non-primary attention-state dependent pathways such as the reticular formation contribute to both the response recorded from the midline and the
148
response recorded over the temporal lobe (Kraus et al., 1988; 1992). Anatomic studies in the cat (Niimi and Matsuoka, 1979; Anderson et al., 1980) have demonstrated that primary auditory cortex receives substantial neural input from the ventral subdivision of the medial geniculate (MGv), while non-primary auditory cortex is innervated by projections from more caudal and medial areas. That the nuclei of the medial geniculate body involve two systems has been a consistent finding, not only in studies of neural connections but also in studies of cell morphology (Winer and Morest, 1983) and single neuron physiologic responses (Calford, 1983). Winer and Morest (1983) cited various descriptive terms for the dichotomy of thalamic nuclei: specific vs. nonspecific, extrinsic vs. intrinsic, core vs. belt. Anderson et al. (1980) applied the distinctively auditory terms of a 'cochleotopic' system and a 'diffuse' system. In addition to the differentiation of subdivisions, Anderson et al. (1980) noted a rostro-caudal gradient in the pattern of connections, with more rostral areas being more closely associated with primary auditory cortex. This is consistent with the observation of Rodrigues-Dagaeff et al. (1989) of a rostro-caudal gradient within the MG, such that more rostral areas are more sharply tuned and selectively auditory-responsive, while caudal areas are more likely to be broadly tuned and multi-sensory. Redies et al. (1989a, b) have described in the guinea pig cortical and thalamic regions comparable to the 'cochleotopic' and 'diffuse' systems described in the eat (Anderson et al., 1980). The thalamocortical projections may be less complex in their tonotopic representation, but they appear to be very similar in the arrangement of connections from MG subdivisions to auditory cortical regions. We hypothesize that the differentiation of the temporal and midline responses is associated respectively with the differentiation of primary and non-primary divisions of the MG. The current study focuses on the role of the medial geniculate body in the generation of the MLR, specifically on how primary and non-primary subdivisions of the MG influence the components of the MLR.
Methods
Subjects Guinea pigs, weighing approximately 350 grams, were used as subjects. Animals were anesthetized with ketamine hydrochloride (100 mg/kg) and xylazine (7 mg/kg), and maintained at a body temperature of 37.5°C (_+1°). Although surgical level anesthesia severely degrades MLR amplitude, within 3 to 4 h responses regain a morphology very similar to that of an awake animal (McGee et al., 1983; Smith and
Kraus, 1987). Smaller doses (15 mg/kg of ketamine; 3 mg/kg of xylazine) were administered as needed for the rest of the experiment to maintain a depth at which the MLR was minimally affected but the pedal withdrawal response was not seen.
Electrophysiologic recording Epidural silver bead electrodes (0.5 mm diameter) were used to record the surface MLR as previously described (Kraus et al., 1988). Recordings were made over the posterior midline and from the temporal lobe contralateral to the stimulated ear. The position of the temporal electrode was approximately over the dorsal portion of primary auditory cortex, as described by Redies et al. (1989b). An electrode placed 15 mm rostral to bregma and 1 mm lateral to the sagittal suture served as a reference. A combination high-impedance depth electrode and micro-injection needle was positioned stereotaxically in the MG, as described by McGee et al. (1991). Coordinates for the MGv were 5.0 mm rostrai to the interaural line, 3.9 mm lateral to the sagittal suture, and 7.5 mm ventral to the surface of the brain. Coordinates for the MGem were 4.5 mm rostral, 3.9 mm lateral, and 7.5 mm ventral. Monaural, 100/~s click stimuli were presented at 70 dB nHL at a rate of 3.5/s through Etymotic insert earphones. Recordings were filtered from 10 to 1500 Hz. Each averaged response consisted of 200 individual responses with a 60 ms time sweep and a 3.25 ms pre-stimulus period. Neural activity was recorded simultaneously from all three channels. The ABR was readily apparent in the first seven milliseconds of the midline recording and was monitored to insure the integrity of the peripheral auditory system. At least ten successive recordings were obtained from all sites prior to the administration of lidocaine. Following this baseline period, 2/~1 of 4% lidocaine HCI in saline was injected into the MG, to reversibly inactivate axonal transmission (Hille, 1966). To control injection rate, a microinfusion pump delivered the lidocaine at a rate of 2/~1/30 s. In the ease of saline control subjects, 0.9% saline with a pH similar to the lidocaine (5.0-7.0)was used. The effect of the drug was observed simultaneously on surface potentials and on the local response, with recordings obtained at intervals of 1-2 min until the return of baseline waveform morphologies and amplitudes. At the end of the experiment, the recording location was marked with electrolytic lesions (35/~A for 10 s). Brains were cut in 17/~ coronal sections and stained with the Kliiver stain which permits visualization of cell body and fiber pathways.
Data analysis Wave amplitudes and latencies before and after lidocaine injection were measured. Amplitude was
149
measured from the preceding trough to the peak, or from the preceding peak to the trough. Thus, the M-] amplitude was measured from wave 4 of the ABR to M- I, the M + amplitude was measured from M- n to M + , etc. Latencies were measured at the midpoint of the peak or trough. All peaks and troughs occurring in a post-stimulu~z latency range of 4 to 40 ms were scored. The means and standard deviations of baseline amplitudes were computed to determine confidence intervals of significant change. A change of more than two standard deviations from the baseline mean was considered significant. To compare lidocaine effe~s across animals, amplitudes were normalized to percentages, with baseline mean defined as 100%. The percentage values were utilized in parametric statistics.
i
A
*
It
a a
C
I a
IOOuV B
W4
Results
Examples of the surface and depth MLR responses are shown in Fig. 1. MLRs recorded from the surface of the temporal cortex (top) exhibit a three component complex, with a positive component at 15 ms (A), a negative component at 27 ms (B), and a positive component at 40 ms (C). The surface midline response (Fig. 1, middle) includes the rapid peaks of the ABR, followed by a negative wave at ]0-12 ms (M-I), a positive wave at 20-25 ms (M + ), and a second negativity (M- 2) at 32-35 ms. The waveform recorded locally from the MG (Fig. 1, bottom) is characterized by a prominent positive wave at roughly 9 ms, and a negativity at 12-17 ms. A more labile broad positive wave at 25-32 ms is sometimes present as well. Lesions were localized in the ventral and caudomedial subdivisions of the M G as described by Redies et al. (1989a). Histological results indicated that ~ajections were successfully made into subdivisions of the medial geniculate in 19 animals: 12 within MGv and 7 within MGcm. Subdivisions of the MG and injection sites are illustrated in the schematic drawings of Fig. 2. The sections are arranged rostral to caudal, spanning a total distance of 2.6 mm through the nucleus. The ventral division (MGv) is characterized by small, homogeneous cells. The caudomedial division (MGcm) is a magnocellular area with numerous fiber pathways. The area just dorsal to the ventral division (MGd) has been called the "shell nucleus' by Redies et al. (!989a), and is distinguished from the MGv on functional grounds.
Effects of pharmacologic inactivation of the MG The injection of ]idocaine (2/~1) in all subdivisions of the MG correlated with the marked reduction of local activity and an immediate disruption of MLR components on the surface of the temporal cortex. Amplitude decrements and latency delays were seen in
V
M- 1
1
10uV
/
M.2
MEDIAL GENICULATE
lb
~
~
tata~ (ms)
4b
~0
~0
Fig. 1. Neural activity recorded from the epidural surface over the temporal lobe (top), the epidural surface over the midline (middle), and the intracTanial electrode in MG (bottom). Responses are plotted posi~,ivityup.
all waves. In some animals, the ABC complex was completely eliminated, in others, the ABC complex was replaced for up to 50 rain by log' amplitude 'transi.tional' waveforms, termed as such because of their marked deviation from baseline morphology. The transitional waveforms often showed a morphology similar to the midline response. Thereafter, a delayed ABC complex typically reappeared, which progressively recovered to baseline amplitude and latency. In some
150
animals, amplitudes recovered to levels greater than baseline. In contrast to the effects on the ABC complex, surfac~ midline changes were dependent on the particular subdivision in which lidocaine was injected. Injections in the caudomedial portion produced major amplitude decrements and latency shifts or completely eliminated the midline response, while injections in the ventral portion of the MG produced much less change in the midline response. Injections in MGv showed a rostro-caudal gradient, with more caudal injections having more effect on midline responses. In Fig. 3 are shown representative waveforms recorded from surface sites before and after lidocaine injection. As described above, injections in ventral Mr3 were associated with disruption of MLR activity obtained over the surface of the temporal cortex. Surface MLRs obtained at the midline were relatively unchanged (top). Injections in the caudo-medial portion of MG were associated with disruption of MLR activity obtained over both the temporal cortex and the midline (bottom).
Amplitude The time course of effects of lidocaine injection on responses recorded from surface midline and temporal sites is shown in Fig. 4 for two animals. The amplitudes of waveforms from these areas were plotted over the course of the experiment. Shown is the initial baseline period, injection at time 0, and recovery to baseline amplitude. Horizontal lines denote a range of the baseline mean + 2 standard deviations. Lidocaine injections in the ventral MG were associated with disruption of MLR activity obtained over the" surface of the temporal cortex, with much less change in the midline response (Fig. 4 top). Lidocaine injections in the caudomedial MG were associated with disruption of MLR activity obtained over both temporal and midline sites (Fig. 4 bottom). Of the 12 animals with histologically confirmed injections in the MGv, 3 showed no change in the midline waveform. Three additional animals showed no change in the amplitude of M-~ and M + , although a reduction was seen in the M- 2 amplitude. The injection sites of these six animals were typically more
Histologic Reconstruction of Injection Sites
.2
3.8
Fig. 2. Schematic of MG with histologic reconstruction of subcortical injection sites: solid dot (.). Numbers in right hand corner of each schematic denote dista~Jce (mm) in a rostral direction from the interaural line.
151 rostral and more lateral than other MGv sites. The other six animals with MCe~ injections showed no change in M-~; moderate reductions in M + and M- 2 amplitude were observed. All seven animals with injections in. MGcm showed severe reductions in M + and M- 2 amplitude post-lidocaine. Absolute response amplitude varied considerably among animals. For analysis purposes, amplitudes were expressed as percent of baseline mean. Of particular interest were the percent amplitudes during the time of maximum effect, 2-12 min post-injection. Fig. 5 shows the mean post-lidocaine amplitudes for each condition during the time of maximum effect. A three-way splitplot A N O V A was performed on these data comparing electrode site x injection site x waves within a response. All three main effects were significant (F = 7.7, P
