INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The response properties of cortical frequency-tuned neurons and the cochleotopic (frequency) map of the primary auditory cortex (AC) can be changed by repetitive acoustic stimulation (Chowdhury and Suga 2000; Ma and Suga 2001a), auditory fear conditioning (Bakin et al. 1996; Diamond and Weinberger 1986, 1989; Gao and Suga 2000; Ji et al. 2001; Ohl and Scheich 1996; Weinberger et al. 1993), learning of an auditory discrimination task (Edeline and Weinberger 1993; Recanzone et al. 1993), focal electrical stimulation of AC (Chowdhury and Suga 2000; Ma and Suga 2001a; Sakai and Suga 2001, 2002), or electric stimulation of the basal forebrain during acoustic stimulation (Bakin and Weinberger 1996; Kilgard and Merzenich 1998a). The response properties of collicular frequency-tuned neurons and the cochleotopic map of the inferior colliculus (IC) can also be changed by repetitive acoustic stimulation (Gao and Suga 1998; Ma and Suga 2001a; Yan and Suga 1998), auditory fear conditioning (Gao and Suga 1998, 2000; Ji et al. 2001), or focal electric stimulation of the AC (Jen et al. 1998; Ma and Suga 2001a; Yan and Ehret 2001; Yan and Suga 1998; Zhang and Suga 2000; Zhou and Jen 2000). Focal cortical electric stimulation also modulates the response properties of collicular neurons tuned to echo delays (Yan and Suga 1996) and sound durations (Ma and Suga 2001a) or sound direction (Jen et al. 1998) and shifts their echo-delay- or duration-tuning curves. Therefore the corticofugal system modulates the functional organization of the IC not only in the frequency domain but also in the time domain. In the big brown bat, Eptesicus fuscus, the cortical and collicular changes (plasticity) both have been found to be greatly due to the corticofugal system (Gao and Suga 2000; Ji et al. 2001). Our present paper deals with both collicular and cortical plasticity in the big brown bat so that papers on collicular and cortical plasticity of the big brown bat are mainly reviewed in the following text.

In the big brown bat, the corticofugal auditory system shifts the best frequencies (BFs) of collicular neurons, together with their frequency-tuning curves, toward the frequency of a repetitively delivered acoustic stimulus (Gao and Suga 1998; Ma and Suga 2001a; Yan and Suga 1998), the frequency of a conditioned sound (Gao and Suga 1998, 2000; Ji et al. 2001), or the BF of electrically stimulated cortical neurons (Ma and Suga 2001a; Yan and Suga 1998). BF shift results in reorganization of the frequency map of the IC. Such "centripetal" BF shifts are basically the same regardless of the means that evoked them. This means that focal electric stimulation of the AC activates the essential portion of the neural mechanism for plasticity of the central auditory system and that it can be an appropriate method for the exploration of the plasticity (Suga et al. 2000).

Collicular and cortical BF shifts, which otherwise would be evoked by trace conditioning with acoustic stimuli followed by electric leg stimulation, are abolished by inactivation of the somatosensory cortex during conditioning (Gao and Suga 1998, 2000). Electric stimulation of the somatosensory cortex (ES_st) after electric stimulation of the AC (ES_at), mimicking trace conditioning, augments the collicular and cortical BF shifts. However, ES_st prior to ES_at, mimicking backward conditioning, does not augment the BF shifts (Ma and Suga 2001a). Therefore the somatosensory cortex is one of the essential portions for the plasticity caused by conditioning, and the sequence of stimulation of the two cortical areas is important for evoking the plasticity.

Cholinergic nerve fibers originating from the basal forebrain control the acetylcholine level in the cortex and play an important role in cortical plasticity as reviewed by Buonomano and Merzenich (1998), Rasmusson (2000), and Sarter and Bruno (2000). In the AC of the guinea pig, BF shift is caused by a tone burst paired with electric stimulation of the cholinergic basal forebrain but not by the tone burst or the electric stimulation alone. The BF shift is similar to that caused by behavioral learning (Bakin et al. 1996; Bjordahl et al. 1998). In the cat's AC, massive progressive reorganization of the cochleotopic map is evoked by electric stimulation of the basal forebrain paired with a tone burst (Kilgard and Merzenich 1998a). In the big brown bat, electric stimulation of the basal forebrain augments the collicular and cortical BF shifts evoked by a train of acoustic stimuli or by electric stimulation of the AC (Ma and Suga 2001a).

Gao and Suga (1998, 2000) proposed the following working hypothesis of collicular and cortical plasticity, incorporating their findings with part of the hypothesis proposed by Weinberger and his coworkers (1990). That is, the central auditory system has an intrinsic mechanism for the reorganization of the central auditory system based on the activity of the AC and the corticofugal system. When a behaviorally irrelevant acoustic stimulus is repetitively delivered, the central auditory system shows a small short-term plasticity. However, when it is paired with electric leg stimulation, the auditory and somatosensory signals ascend from the stimulated sensory cells to the auditory and somatosensory cortices, respectively, and then to the amygdala through the association cortex. These signals are probably associated in the amygdala, which is essential for evoking conditioned behavioral response. Then the acoustic stimulus becomes behaviorally relevant to the animal. The amygdala sends the "associated" signal to the cholinergic basal forebrain, which increases the cortical acetylcholine level. Therefore the plasticity in the AC and IC due to the activity of the AC and the corticofugal system is augmented.

Ji et al. (2001) applied acetylcholine or atropine to the AC or IC to examine their effect on the collicular and cortical BF shifts evoked by conditioning and obtained data indicating that acetylcholine augments both the cortical and collicular BF shifts and that the cortical BF shift depends on the cortical neural net, corticofugal system (feedback loops) and cortical acetylcholine (ACh) level. Their data support Gao and Suga's hypothesis. A 30-min-long conditioning session with acoustic stimulation followed by electric leg stimulation evoked a long-term cortical BF shift (Gao and Suga 2000) and a short-term collicular BF shift (Gao and Suga 1998, 2000). Therefore there is a possibility that the augmentation of the ES_at-evoked BF shifts by ES_st is due to the activation of the basal forebrain through the pathway from the somatosensory cortex to the association cortex, then to the amygdala, and finally to the basal forebrain, as hypothesized by Gao and Suga (1998, 2000). The aim of our present research was to further test the validity of their hypothesis. We studied whether electric stimulation of the basal forebrain and/or the somatosensory cortex augments the collicular and cortical BF shifts evoked by ES_at. We also studied the effect of a bilateral lesion of the basal forebrain on the collicular and cortical BF shifts evoked by ES_at and also on the augmentation of the collicular and cortical BF shifts by ES_st paired with ES_at. We have obtained data that support Gao and Suga's hypothesis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials, surgery, recording of neural activity, acoustic stimulation, electric brain stimulation, data acquisition, and data processing were basically the same as those described in Ma and Suga (2001a). Therefore only the essential portion of the methods are summarized in the following text. Eleven adult big brown bats (18-24 g body wt) were used for the present experiments. Under neuroleptanalgesia (Innovar 4.08 mg/kg body wt), a 1.5-cm-long metal post was glued on the dorsal surface of the bat's skull. The physiological experiment was started 3-4 days after the surgery. The awake animal was placed in a polyethylene-foam body-mold and was hung at the center of a soundproof room that was maintained at 31°C. The bats used were neither anesthetized nor tranquilized. The temperature, monitored with a thermistor placed between the bat and body mold, was 37°C. The metal post mounted on the skull was fixed on a metal rod with set screws to immobilize the animal's head, and the bat's head was adjusted to face directly at a loudspeaker located 74 cm away. Holes 50-100 µm in diameter were made in the skull covering the AC, IC, primary somatosensory cortex, or dorsal to the basal forebrain. Tungsten-wire electrodes for recording action potentials or for electrically stimulating neurons were inserted into the brain through these holes (see following text). The bats were monitored on a video monitor screen during the experiments. The protocol for this research was approved by the animal studies committee of Washington University.

Acoustic stimulation

Acoustic stimuli were 20-ms-long tone bursts with a 0.5-ms rise-decay time. They were generated by a voltage-controlled oscillator and an electronic switch, and were delivered at a rate of 5/s with a leaf tweeter. The frequency and amplitude of the tone bursts were varied manually or computer-controlled. The amplitude was calibrated with a Bruel & Kjael microphone and was expressed in dB SPL.

The frequency-tuning curve of a single collicular or cortical neuron was first measured manually. Then the amplitude of a tone burst was fixed at 10 dB above minimum threshold of the neuron, and a computer-controlled frequency scan was delivered. The frequency scan consisted of 21 time blocks. In the first 20 blocks, frequency was changed in 0.3- or 0.5-kHz steps, and in the 21st (last) block, no stimulus was presented to count background discharges. The duration of each block was 200 ms, so that the duration of the frequency scan was 4,200 ms. An identical frequency scan was repeated 50 times, and the response of a single neuron to the scan was displayed as an array of peristimulus time (PST) histograms or PST cumulative (PSTC) histograms.

Electric stimulation of the primary auditory cortex

Electric stimulation (ES_ar) were delivered to the AC through a pair of tungsten-wire electrodes, the tips of which were 6-8 µm in diameter and were separated by 150 µm, one proximal to the other. These electrodes were used first to record auditory responses of cortical neurons at depths of 200-900 µm, i.e., at cortical layers III-VI, then to measure the BF and minimum threshold of these neurons, and finally to electrically stimulate them. A 6.2-ms-long train of four monophasic electric pulses (100 nA, 0.2-ms duration, 2.0-ms interval). The train of electric pulses was repetitively delivered at a rate of 10/s for 2-90 min (hereafter, ES_ar). These stimulus parameters were chosen in the previous studies on corticofugal modulation of bat's auditory neurons (Chowdhury and Suga 2000; Ma and Suga 2001a,b; Yan and Suga 1998) because the bat emits biosonar pulses at a rate of ~10/s in the search phase of echolocation (Griffin 1962). The electric pulses were estimated to stimulate neurons within a 60-µm radius around the electrode tip (Yan and Suga 1996). Therefore electric stimulation of the AC was quite focal. The bat showed no behavioral response at all to such a weak electric stimulation delivered to the AC.

Electric stimulation of the primary somatosensory cortex

To mimic trace conditioning with a train of tone pulses followed by an electric leg stimulation, a train of electrical stimuli delivered to the AC (hereafter ES_at) was followed by an electric stimulation of the ipsilateral primary somatosensory cortex (hereafter ES_st) with a 1.0-s gap. The somatosensory cortex was localized by referring to the somatotopic map studied by Krubitzer and Calford (1992) and by recording neural responses to touch stimuli before inserting a pair of electrodes for electric stimulation (Fig. 1A). The ES_at was 1.0-s long and consisted of 33 trains. Each train was 6.2-ms long and consisted of four monophasic electric pulses, as in ES_ar. ES_at was delivered twice per minute for 30 min. ES_st was 50-ms long and consisted of 20 0.2-ms-long, 100-µA electric pulses. It was also delivered twice per minute for 30 min. ES_at + ES_st was 1.0 s ES_at + 1.0 s gap + 50 ms ES_st. To mimic backward conditioning, ES_st was delivered 1.0 s before ES_at. Because the bat emits biosonar pulses at a rate of 30-40/s at the middle of the approach phase of echolocation and because the mid-approach phase is most likely to be followed by the terminal phase and the contact with an insect (Griffin 1962), the preceding parameters of ES_at and ES_at + ES_st were chosen in our previous studies on the plasticity of the auditory system caused by electric stimulation (Ma and Suga 2001a) mimicking fear conditioning (Gao and Suga 1998, 2000). Therefore these parameters were also used in our present studies.

View larger version (66K): [in this window] [in a new window]   Fig. 1. Dorsolateral view of the brain of the big brown bat (A), frontal sections across the basal forebrain (B), electroencephalograph (EEG, C), and action potentials selected by a time-amplitude window discriminator (D). A: the auditory (AC) and somatosensory (SI) cortices are indicated (- - -). ×, locations where electric stimulation was applied. B: an electrode penetration and an electrolitic lesion at the tip of the stimulating electrodes. CP, caudate putamen; GP, globus pallidus. C: EEG recorded from the auditory cortex is desynchronized by a 30-s-long electric stimulation of the basal forebrain. D: action potentials selected with a time-amplitude window discriminator. Top: an action potential stored at the beginning of data acquisition (template). Bottom 5 traces: action potentials recorded every 1 h. , an amplitude level and a time delayed window for selecting action potentials.

Because ES_at and ES_st stimulated small groups of neurons around the tips of the stimulus electrodes, the spatial and temporal pattern of neural activity evoked by these direct electric stimulation might be quite different from that evoked by auditory fear conditioning. However, it was tested whether ES_st augmented both the collicular and cortical changes evoked by ES_at.

Electric stimulation and lesion of the basal forebrain

The photomicrographs of frontal sections across the forebrain of the big brown bat were very similar to those of the mouse. A pair of tungsten-wire electrodes fabricated in the same way as those used for ES_at and ES_ar was inserted dorsoventrally into the basal forebrain ipsilateral to the electrically stimulated AC, referring the atlas of the mouse brain (Slotnick and Leonard 1975), and a 0.2-ms-long, 100-µA electric pulse was repetitively delivered to the basal forebrain at a rate of 100/s over 15 or 30 min through these electrodes. The electrodes were implanted where the electric stimulation evoked the largest augmentation of the collicular and cortical BF shifts evoked by electric stimulation of the AC. Because it has been known that the electric stimulation of the basal forebrain paired with acoustic stimulation evokes plasticity of the AC (Bakin and Weinberger 1996; Kilgard and Merzenich 1998a,b), this procedure was one of the three criteria to verify that the electrodes were placed in the basal forebrain.

Because it has been known that electric stimulation of the basal forebrain produces desynchronization of an electroencephalogram (EEG) (Bjordahl et al. 1998), we anesthetized the animal with urethan (1.5 g/kg ip) after completing the plasticity experiment, recorded a synchronized EEG (1- to 300-Hz band-pass), and observed desynchronization of the EEG that was induced by 0.2-ms-long, 150-µA electric pulses delivered at a rate of 100/s for 30 s (Fig. 1C). This observation was the second criterion to verify the stimulation site. The third verification of the stimulation site was anatomical. The basal forebrain was lesioned by a 10-s-long 500-µA monophasic electric current. The animal was then killed by pentobarbital sodium (100 mg/kg) and was perfused with a formalin-saline solution. Its brain was frozen-sectioned 50 µm thick and Nissl stained. The electrolytic lesion was ~200 µm in radius and was located at the basal forebrain (Fig. 1B).

Because electrical stimulation of the basal forebrain (ES_br) as well as ES_ar stimulated a small group of neurons at and around the tips of the stimulus electrodes, the spatial and temporal pattern of neural activity evoked by ES_br might be quite different from that evoked by auditory fear conditioning. However, it was tested whether ES_br augmented both the collicular and cortical changes evoked by ES_ar.

Data acquisition and processing

The auditory responses of cortical neurons were recorded with a tungsten-wire microelectrode (6-8 µm tip diameter) at depths between 200 and 600 µm. The auditory responses of collicular neurons were recorded with a tungsten-wire microelectrode at depths between 200 and 2,000 µm in the central nucleus of the IC ipsilateral to the electrically stimulated AC. The central nucleus of the IC is big and shows a simple and systematic tonotopic organization (Casseday and Covey 1992). The dorsal surface of the IC is directly visible through the skull. In dorsoventral electrode penetrations through the dorsal surface, the electrode passed across the central nucleus of the IC. Therefore BFs of neurons systematically became higher as expected from the tonotopic map (e.g., Yan and Suga 1998).

Action potentials originating from a single neuron were selected from those originating from a few neurons with a time-amplitude-window discriminator (BAK Electric, model DIS-1). The responses of a single neuron to tone bursts in the frequency scan repeated 50 times were recorded before, during, and after electric stimulation of the AC, somatosensory cortex, and/or basal forebrain. The auditory responses of a single neuron to acoustic stimuli were displayed on the computer monitor as arrays of PST or PSTC histograms during the experiments. The waveform of an action potential was stored on a digital storage oscilloscope at the beginning of the data acquisition and was used as a template (Fig. 1D, top). Action potentials discharged by the neuron were continuously monitored together with the template on the screen of the digital storage oscilloscope during data acquisition: before, during, and after the electric and/or acoustic stimulation. Data acquisition was continued as far as action potentials visually matched the template (Fig. 1D, bottom 5 traces). Data were stored on a hard drive of a personal computer and were used for off-line analysis. In a 1-day experiment that lasted ~8 h, a single collicular or cortical neuron was usually studied.

Off-line data processing included plotting the frequency-response curves (the arrays of PST or PSTC histograms displaying the responses of a collicular or a cortical neuron to 50 identical frequency scans) obtained before, during, and after the electric and/or acoustic stimulation (see Ma and Suga 2001a). The BF was determined as the frequency to which the neuron showed the largest response. A recovery time was measured as the time interval between the end of the electric stimulation and the time for a 50 or 100% (full) recovery. A full recovery time was defined as the time when a recovery curve crossed a point 50 Hz below the control BF. For a statistical analysis of the data, means ± SE of maximum BF shifts and recovery times at 50 and/or 100% recovery points were calculated. A t-test was used to test the difference between the BFs obtained before and after the electric and/or acoustic stimulation and to test the difference between the responses of collicular and cortical neurons.

Stimulus parameters were as follows: AS_t, train of acoustic stimuli (10-ms-long tone bursts at 50 dB SPL; 33 tone bursts/s for 30 min); ES_ar, repetitive electrical stimulation of the auditory cortex (4 0.2-ms-long, 100-nA electric pulses/6.2-ms-long train; 10 train/s for 15, 30, or 60 min); ES_at, train of electric stimulation of the auditory cortex (4 0.2-ms-long, 100-nA electric pulses/6.2-ms-long train; 33 train/s for 1 s; 2 ES_at/min); ES_br, repetitive electric stimulation of the basal forebrain (0.2-ms-long, 100-µA electric pulses; 100 pulses/s for 15 or 30 min); ES_st, train of electrical stimulation of the somatosensory cortex (20 0.2-ms-long, 100-µA electric pulses/50-ms-long train; 2 ES_at/min for 30 min); ES_at + ES_st, ES_st was delivered 1.0 s after ES_at to mimic auditory fear conditioning; ES_st + ES_at, ES_st was delivered 1.0 s before ES_at to mimic backward conditioning; ES_at + ES_br, a 15 (or 30)-min-long session of ES_at and 15 (or 30)-min-long session of ES_br were delivered at the same time; and ES_st + ES_at + ES_br, a 30-min-long ES_at + ES_st session and a 30-min-long ES_br session were delivered at the same time.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

BF shift (plasticity) evoked by electric stimulation of the AC (ES_ar or ES_at) was studied for 144 collicular and 78 cortical auditory neurons. The effect of ES_br on the BF shift evoked by ES_ar or ES_at was studied for 133 collicular and 100 cortical neurons. The effects of ES_br alone and ES_st alone were studied on 26 collicular and 26 cortical neurons, respectively. ES_br or ES_st alone caused no BF shift of the collicular and cortical auditory neurons studied. In some of the preceding neurons, collicular and cortical BF shifts were studied after bilateral lesion of the basal forebrain.

Effect of ES_br on plasticity evoked by ES_ar

When a 15-min-long ES_ar was delivered, the BFs of collicular neurons within 10 kHz of the BF of stimulated cortical neurons (hereafter, stimulated cortical BF) mostly shifted toward the stimulated cortical BF (Fig. 2A, ). The largest negative BF shift was 1.5 kHz, which occurred at 4.5 kHz above the stimulated cortical BF. The largest positive BF shift was 0.5 kHz, which occurred at 15 kHz above and 4 kHz below the stimulated cortical BF. BF shifts were "centripetal" for BF differences between 5 and 13 kHz, but were "centrifugal" at 15 kHz above the stimulated cortical BF. When a 15-min-long ES_ar session was delivered at the same time as a 15-min-long ES_br session (hereafter, ES_ar + ES_br), the collicular BF shifts evoked by ES_ar were augmented (Fig. 2A, ). The maximum negative BF shift became 2.0 kHz, which occurred at 4-7 kHz above the stimulated cortical BF. The largest positive BF shift was 0.8 kHz, which occurred at 5 kHz below the stimulated cortical BF. (Note that in ES_ar + ES_br, the length of a ES_ar session was always the same as that of a ES_br session and that ES_ar consisted of 10 trains of electric pulses/s, whereas ES_br consisted of 100 electric pulses/s, so that every 10 electric pulses of ES_br coincided with 1 train in ES_ar.)

View larger version (28K): [in this window] [in a new window]   Fig. 2. Changes in the best frequencies (BFs) of collicular (IC) and cortical neurons (AC) evoked by repetitive electric stimulation of the auditory cortex (ESar) or ESar paired with repetitive electric stimulation of the basal forebrain (ESbr) for 15 min (A) or 60 min (B-D). In A and B, BF shifts of collicular neurons are plotted as a function of the differences in BF between recorded collicular and stimulated cortical neurons. BF shifts evoked by ESar () were augmented by ESbr (). Because there were many identical data points, they are slightly shifted to be individually identifiable. C and D: the time courses of the development of the BF shifts of collicular and cortical neurons evoked by ESar () or ESar + ESbr (), respectively. The vertical bars represent SEs. The BFs of stimulated cortical neurons ranged between 16 and 65 kHz (31.6 ± 1.5 kHz), and the BFs of recorded neurons were 14 and 68 kHz (30.3 ± 2.5 kHz). Different symbols indicate different stimulus conditions. N, the number of neurons studied in a given stimulus condition.

When a 60-min-long ES_ar was delivered, collicular BF shifts became larger than those evoked by the 15-min-long ES_ar (Fig. 2B, ). The largest negative and positive BF shifts were, respectively, 2.0 and 0.5 kHz, which occurred, respectively, at 5-6 kHz above the stimulated cortical BF and at ~4 kHz below and 14 kHz above it. For a 60-min-long ES_ar + ES_br, the BF shifts evoked by ES_ar were augmented (Fig. 2B, ). The largest negative shifts became 2.5 kHz, which occurred at 6-7 kHz above the stimulated cortical BF, and the largest positive BF shifts were 1.0 kHz, which occurred at both 3.0 kHz below and 15 kHz above the stimulated cortical BF. As shown in Fig. 2, A and B, the BF shifts were predominantly centripetal and asymmetrical, i.e., BF shifts toward the stimulated cortical BF were much larger on the high-frequency side of the stimulated cortical BF than on the low-frequency side.

In our data, 58% (23/40) of collicular neurons studied showed BF shifts for the 15-min-long ES_ar alone; 89% (32/36) for 15-min-long ES_ar + ES_br; 69% (24/35) for the 60-min-long ES_ar alone; and 73% (24/33) for 60-min ES_ar + ES_br. A BF shift depends on differences in both BF (Chowdhury and Suga 2000; Ma and Suga 2001a; Sakai and Suga 2001; Yan and Suga 1998; Zhang and Suga 2000) and distance along iso-BF lines (Sakai and Suga 2002) between stimulated and recorded neurons. Therefore the preceding percentages of BF-shifted neurons calculated according to the suggestion of the referees of our present paper, ignoring the relationship in BF between recorded and stimulated neurons, were meaningless. As shown in Fig. 2, A and B, a percentage of neurons showing a BF shift varied with BF differences. For 4- to 7-kHz BF differences, 84% of neurons (31/37) showed BF shifts for ES_ar alone and 92% (26/28) showed BF shifts for ES_ar + ES_br. Because BF shifts were largest at and around 5 kHz BF difference as previously reported by Ma and Suga (2001a) and because an increase in BF shift caused by ES_br was significant at the 4- to 7-kHz BF differences (Table 1), the further measurements of BF shifts in the present studies were performed for BF differences between 3.6 and 6.9 kHz (4.5 ± 0.06 kHz BF difference, n = 182).

The collicular BF shift evoked by the 60-min-long ES_ar developed during the ES_ar and reached a plateau of 1.01 ± 0.06 kHz (n = 15) ~30 min after the onset of the ES_ar (Fig. 2C, ). For 60-min-long ES_ar + ES_br, the BF shift became larger and reached a plateau of 1.50 ± 0.05 kHz (n = 16) ~60 min after the onset of the ES_ar + ES_br (Fig. 2C, ). The 0.5-kHz difference in plateau was statistically significant (P < 0.05). The cortical BF shift also developed during the 60-min-long ES_ar and reached a plateau of 1.12 ± 0.04 kHz (n = 13) ~60 min after the ES_ar (Fig. 2D, ). For the 60-min-long ES_ar + ES_br, it became larger and reached a plateau of 1.70 ± 0.08 kHz (n = 15) 60 min after the ES_ar + ES_br (Fig. 2D, ). The 0.6-kHz BF difference in plateau was statistically significant (P < 0.05).

The collicular and cortical BF shifts monotonically returned (recovered) to the BF in the control condition (hereafter, control BF) after the 15- or 30-min-long ES_ar (Fig. 3, A and B,  and ), as previously reported by Ma and Suga (2001a). For the 15- or 30-min-long ES_ar + ES_br, the BF shifts became larger in magnitude and longer in recovery than those evoked by the ES_ar alone (Table 2). Figure 3A shows the recovery curves of the collicular BF shifts evoked by the 15- or 30-min ES_ar with or without ES_br. For the 15-min-long ES_ar alone, the collicular BF shift was 0.82 ± 0.08 kHz (n = 15), and the shifted BF recovered to 50% of the control BF 32 ± 3.3 min after the ES_ar and to the control BF 57 ± 4.6 min after the ES_ar (Fig. 3A, ). For the 15-min-long ES_ar + ES_br, the collicular BF shift became larger in amount, 1.18 ± 0.07 kHz (n = 16), and longer in recovery time: 88 ± 5.6 min for 50% recovery and 153 ± 8.2 min for 100% recovery (Fig. 3A, ). For the 30-min-long ES_ar alone, the collicular BF shift was 1.10 ± 0.05 kHz (n = 18), the recovery time was 70 ± 4.9 min for 50% and 166 ± 9.6 min for 100% recovery (Fig. 3A, ). For the 30-min-long ES_ar + ES_br, the collicular BF shift became larger in magnitude, 1.32 ± 0.06 kHz (n = 14) and longer in recovery time: 101 ± 9.6 min for 50% and 176 ± 9.9 min for 100% recovery (Fig. 3A, ). All these changes, except the 100% recovery time for the 30-min-long ES_ar + ES_br, were statistically significant. The effect of ES_br was larger on the BF shifts evoked by the 15-min-long ES_ar than on those evoked by the 30-min-long ES_ar (Table 2).

View larger version (32K): [in this window] [in a new window]   Fig. 3. Recovery curves of the collicular (A) and cortical BF shifts (B) evoked by 15 ( and )- or 30-min-long ( and ) ESar or ESar + ESbr. Different symbols indicate different stimulus conditions, as indicated in the figure. N, the number of neurons studied in a given stimulus condition. ESar + ESbr evoked BF shifts ( and ) that were larger and longer-lasting than those evoked by ESar alone ( and ).  and , ESar and ESbr, respectively. The length of the bars indicates the duration of the stimulus, which was either 15- or 30-min-long. The location of the bars indicates the time for the delivery of ESar or ESbr. The end of ESar or ESbr is marked as 0.0 h. The stimulated cortical BFs ranged between 20 and 60 kHz (36.5 ± 1.5 kHz, n = 51) for collicular studies and between 22 and 65 kHz (36.8 ± 1.9 kHz, n = 52) for cortical studies. The BF difference between the stimulated cortical neurons and the recorded collicular or cortical neurons was 4.2 ± 0.08 or 5.1 ± 0.10 kHz, respectively.

Figure 3B shows the recovery curves of the BF shifts of cortical neurons evoked by ES_ar or ES_ar + ES_br that was either 15 or 30 min long. These cortical recovery curves were similar to the collicular ones. However, the cortical BF shifts at the peak tended to last longer than the collicular ones. Accordingly, the 50% recovery time of the cortical BF shift was significantly longer than that of the collicular one: 47 ± 2.9 min for the 15-min-long ES_ar and 108 ± 8.9 min for the 30-min-long ES_ar. Lengthening of the duration of ES_ar from 15 to 30 min had a larger effect on the 50% recovery time of the cortical BF shift than on that of the collicular BF shifts. The difference in 50% recovery time between the collicular and cortical BF shifts for a given duration of ES_ar was statistically significant (P < 0.01). For the 15- or 30-min-long ES_ar + ES_br, the cortical BF shift became larger: 1.28 ± 0.09 kHz (n = 14; P < 0.05; Fig. 3B, ) or 1.41 ± 0.10 kHz (n = 15; P = 0.07; Fig. 3B, ), respectively. The 50% recovery time significantly increased (Table 2).

Compared with the 15-min-long ES_ar + ES_br (i.e., simultaneous delivery of the ES_ar and ES_br sessions), the 15-min-long ES_br delivered immediately before or after the 15-min-long ES_ar evoked much smaller changes of the collicular and cortical BF shifts in amount and recovery time (Fig. 4). These small changes were statistically insignificant (P > 0.05) compared with the BF shifts evoked by the ES_ar alone but significant compared with the BF shifts evoked by the ES_ar + ES_br (Table 3). When the 30-min-long ES_br was delivered prior to and during the 15-min-long ES_ar, the collicular and cortical BF shifts (Fig. 4, A and B, ) tended to be slightly larger and longer-lasting than those evoked by the 15-min-long ES_ar + ES_br (Fig. 4, A and B, ). However, the difference in BF shift was small and statistically insignificant in both magnitude and recovery time.

View larger version (31K): [in this window] [in a new window]   Fig. 4. Effect of ESbr delivered prior to, during, and/or after ESar on the collicular (A) and cortical BF shifts (B) evoked by a 15-min-long ESar. A simultaneous delivery of ESar and ESbr was most effective in evoking a large and long-lasting BF shift. The stimulated cortical BFs ranged between 19 and 60 kHz (35.5 ± 2.5 kHz) for collicular studies and between 23 and 62 kHz (36.2 ± 3.5 kHz) for cortical studies. The BF difference between the stimulated cortical neurons and the recorded collicular or cortical neurons was 4.2 ± 0.04 or 5.0 ± 0.15 kHz, respectively. See Fig. 3 for abbreviations and other explanations.

The collicular and cortical BF shifts evoked by the 15-min-long ES_ar, respectively, recovered by 50% in 32 ± 3.3 and 47 ± 2.9 min and by 100% in 57 ± 4.6 and 71 ± 5.1 min after ES_ar (Fig. 3, A and B, ). When the 15-min-long ES_ar was delivered three times with a 1-h time interval, the collicular and cortical BF shifts for the third ES_ar were the same in amount as those evoked by the first ES_ar but showed a prominent lengthening in recovery time. That is, 50 and 100% recovery times after the third ES_ar were, respectively, 85 ± 5.5 and 121 ± 6.2 min for the collicular BF shift (Fig. 5, ) and 96 ± 5.1 and 150 ± 7.6 min for the cortical BF shift (Fig. 5, ). When the 15-min-long ES_ar + ES_br was repeated three times, the collicular and cortical BF shifts for the third pair was the same in amount as those evoked by the first pair but showed a recovery time much longer than that for the first pair (Figs. 5,  and , and 3, A and B, ). That is, for the third ES_ar + ES_br, the collicular BF shift plateaued for ~80 min and then recovered by 50% in ~60 min and by 100% in ~110 min (Fig. 5, ), whereas the cortical BF shift plateaued for ~160 min and then recovered by 50% in ~90 min and by 100% in ~200 min (Fig. 5, ). The effect of ES_br was stronger on the cortical BF shift than on the collicular one.

View larger version (21K): [in this window] [in a new window]   Fig. 5. Collicular ( and ) and cortical BF shifts ( and ) evoked by a 15-min-long ESar ( and ) or ESar + ESbr ( and ) delivered 3 times with a 60-min interval. The stimulated cortical BFs ranged between 16 and 64 kHz (32.3 ± 3.5 kHz, n = 29). The BF difference between the stimulated cortical neurons and the recorded collicular or cortical neurons was 4.5 ± 0.05 or 6.8 ± 0.11 kHz, respectively. See Fig. 3 for abbreviations and other explanations.

Effect of basal forebrain lesion on plasticity evoked by ES_ar

When the 15- or 30-min-long ES_ar was delivered after a bilateral lesion of the basal forebrain, the collicular BF shift evoked by ES_ar tended to become slightly smaller than that evoked by the ES_ar without the lesion. However, this decrease was statistically insignificant, P > 0.05 (Fig. 6A,  vs. ;  vs. ; also see Table 4). On the other hand, the cortical BF shift evoked by the 15- or 30-min-long ES_ar after the basal forebrain lesion became smaller and shorter-lasting (Fig. 6B,  vs.  and  vs. ). The decrease in the amount of the BF shift was statistically insignificant (P > 0.05). However, the decrease in the 50 or 100% recovery time of the BF shift was statistically significant for the 30-min-long ES_ar (P < 0.05) but not for the 15-min-long ES_ar (Fig. 6B and Table 4). The preceding data obtained through electrical stimulation of the AC with a basal forebrain lesion indicate that the cortical BF shift evoked by ES_ar was slightly augmented by the basal forebrain, but the collicular BF shift was not.

View larger version (29K): [in this window] [in a new window]   Fig. 6. Effect of a bilateral lesion of the basal nucleus of the forebrain (BN) on the collicular (A) and cortical BF shifts (B) to be evoked by the 15 ( and )- or 30-min-long ESar ( and ). The lesion of the BN had little effect on the collicular BF shifts (P > 0.05) but a significant effect on the cortical BF shifts (P < 0.05). "No BN" means the lesion of the BN. The stimulated cortical BF ranged between 20 and 60 kHz (34.5 ± 1.8 kHz, n = 46) for collicular studies and between 20 and 70 kHz (38.8 ± 1.5 kHz, n = 42) for cortical studies. The BF difference between the stimulated cortical neurons and the recorded collicular or cortical neurons was 4.0 ± 0.09 or 4.8 ± 0.16 kHz, respectively. See Fig. 3 for abbreviations and other explanations.

Effect of ES_st on plasticity evoked by ES_at or ES_at + ES_br

To mimic fear conditioning by Gao and Suga (1998, 2000), the AC, somatosensory cortex and basal forebrain were electrically stimulated for 30 min. A short train of electric stimulation of the somatosensory cortex (ES_st) following ES_at with an 1.0-s delay significantly augmented the collicular and cortical BF shifts evoked by ES_at (Fig. 7,  vs.  in A and  vs.  in B), as previously reported by Ma and Suga (2001a). The recovery curve of the collicular BF shift for the ES_at + ES_st (Fig. 7A, ) was very similar to that for a 30-min-long conditioning session obtained by Gao and Suga (2000) (Fig. 7A, - - -). However, the recovery curve of the cortical BF shift for the ES_at + ES_st (Fig. 7B, ) was quite different from that for the conditioning obtained by Gao and Suga (2000) (Fig. 7B, - - -). When a 30-min-long ES_br session was delivered together with a 30-min-long session of ES_at + ES_st, the collicular and cortical BF shifts evoked by the ES_at + ES_st were further augmented in magnitude and lengthened in recovery time. The magnitude of the collicular BF shift was 0.78 ± 0.11 kHz (n = 15) for ES_at alone, 1.08 ± 0.08 kHz (n = 15) for ES_at + ES_st, and 1.32 ± 0.08 kHz (n = 12) for ES_at + ES_st + ES_br (Fig. 7A). The recovery curve of the collicular BF shift for ES_at + ES_st + ES_br showed a plateau lasting ~90 min and a gradual recovery after the plateau (Fig. 7A, ). The magnitude of the cortical BF shift was 0.79 ± 0.12 kHz (n = 12) for ES_at alone, 1.27 ± 0.07 kHz (n = 12) for ES_at + ES_st, and 1.52 ± 0.09 kHz (n = 11) for ES_at + ES_st + ES_br (Fig. 7B). The recovery of the cortical BF shift for ES_at + ES_st + ES_br was much longer than that of the collicular BF shift. Namely, the cortical BF shift showed a plateau lasting ~210 min and then started to recover (Fig. 7B, ). This cortical recovery curve was quite different from that for the conditioning which showed a plateau lasting >360 min (Fig. 7B, - - -) (Gao and Suga 2000).

View larger version (32K): [in this window] [in a new window]   Fig. 7. Collicular (A) and cortical BF shifts (B) evoked by the trains of electric stimulation of the auditory cortex (ESat;  and ) were augmented by the trains of electric stimulation of the somatosensory cortex (ESst) ( and ). They were further augmented by electric stimulation of the basal forebrain (ESbr). The , , and , respectively, represent ESat, ESst, and ESbr. - - - in A and B, respectively: the recovery curves of the collicular and cortical BF shifts evoked by auditory conditioning (Gao and Suga 2000). The stimulated cortical BFs ranged between 20 and 60 kHz (35.6 ± 2.1 kHz, n = 22) for collicular studies and between 20 and 71 kHz (36.8 ± 2.4.5 kHz, n = 21) for cortical studies. The BF difference between the stimulated cortical neurons and the recorded collicular or cortical neurons was 3.9 ± 0.05 or 4.5 ± 0.08 kHz, respectively. See Fig. 3 for abbreviations and other explanations.

As described in the preceding text, ES_st augmented the collicular and cortical BF shifts evoked by ES_at (Figs. 7 and 8,  in A and  in B). ES_br augmented the BF shifts evoked by ES_at + ES_st (Fig. 7,  in A and B). When the 30-min-long ES_at + ES_st was delivered to the animal after the basal forebrain lesion, the collicular BF shift evoked by the ES_at + ES_st tended to be smaller and shorter-lasting than that without the lesion (Fig. 8A,  vs. ). It became the same in magnitude and recovery as that evoked by ES_at alone. The 50% recovery time was 108 ± 6.5, 133 ± 8.8, and 148 ± 9.8 min for the ES_at alone, ES_at + ES_st with the lesion, and ES_at + ES_st without the lesion, respectively. The difference in 50% recovery time was statistically insignificant (P > 0.05). On the other hand, the cortical BF shift evoked by the ES_at + ES_st after the lesion of the basal forebrain was significantly smaller in magnitude (1.02 ± 0.02 vs. 1.25 ± 0.03 kHz; P < 0.05) and shorter in recovery than that evoked by the ES_at + ES_st without the lesion (123 ± 8.9 vs. 210 ± 11.2 min for 50% and 230 ± 10.6 vs. 325 ± 11.2 min for 100% recovery; P < 0.05) (Fig. 8B, filled triangles vs. filled squares). These data indicate that the augmentation of the cortical BF shift evoked by the ES_st was not due to subcortical interaction between the auditory and somatosensory systems but the basal forebrain activated by ES_at + ES_st.

View larger version (26K): [in this window] [in a new window]   Fig. 8. Effect of a bilateral lesion of the basal forebrain on the augmentation of the collicular (A) and cortical BF shifts (B) evoked by ESat + ESst. The BF shifts evoked by ESat ( or ) were augmented by ESst ( or ). However, the augmentation to be evoked was abolished when the basal forebrain (BN) was bilaterally lesioned (). The stimulated cortical BF's ranged between 20 and 60 kHz (35.6 ± 3.5 kHz, n = 38) for collicular studies and between 20 and 71 kHz (36.8 ± 2.9 kHz, n = 31) for cortical studies. The BF difference between the stimulated cortical neurons and the recorded collicular or cortical neurons was 4.5 ± 0.05 or 4.1 ± 0.07 kHz, respectively. See Fig. 3 for abbreviations and other explanations.

One may consider that the augmentation of the BF shifts by ES_br and/or ES_st was not related to the BF shifts evoked by a train of acoustic stimuli (AS_t) because ES_at was unnatural. Therefore the effects of ES_br or ES_st on the collicular and cortical BF shifts evoked by AS_t were studied. For a 30-min-long AS_t, 20 of the 23 collicular neurons studied and 18 of the 24 cortical neurons studied showed small and short-lasting BF shifts (Fig. 9, A,  and , and B, - - -; Table 5). When AS_t was paired with a 30-min-long ES_br, the collicular and cortical BF shifts became larger and longer-lasting (Fig. 9A,  and ; Table 5). These changes are statistically significant (P < 0.05). The effect of ES_br was larger on the cortical neurons than on the collicular neurons (P < 0.05).

View larger version (29K): [in this window] [in a new window]   Fig. 9. Collicular () and cortical BF shifts () evoked by the trains of acoustic stimuli (ASt) were augmented by ESbr ( and  in A) or ESst ( and  in B). In B, the augmentation of the collicular and cortical BF shifts to be evoked by ESst ( and ) was abolished by a bilateral lesion of BN ( and ). The horizontal dotted bar represents ASt. See Figs. 3 and 7 for abbreviations and other explanations.

When AS_t was paired with ES_st, mimicking trace conditioning, the collicular and cortical BF shifts also became larger and longer-lasting (Fig. 9B,  and ; Table 5). When the AS_t + ES_st was delivered to the animal after the basal forebrain lesion, the augmentation of the collicular and cortical BF shifts that otherwise would be evoked by the AS_t + ES_st was hardly evoked (Fig. 9B,  and ; Table 5). ES_br and ES_st augmented the collicular and cortical BF shifts evoked by AS_t as well as those evoked by ES_at, and the augmentation by ES_st was evoked via the basal forebrain.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Electric stimulation (ES_at, ES_st and/or ES_br) versus auditory fear conditioning

ES_ar alone evokes collicular and cortical BF shifts as previously reported (Chowdhury and Suga 2000; Ma and Suga 2001a; Sakai and Suga 2001, 2002; Yan and Suga 1998; Zhang and Suga 2000). A lesion of the basal forebrain had no effect on the ES_ar-evoked collicular BF shift, but a small effect on the ES_ar-evoked cortical BF shift. These observations indicate that the auditory system has an intrinsic mechanism to evoke BF shifts and that the cortical BF shift is augmented by the basal forebrain. It has been found that the cortical ACh level is increased by acoustic stimuli that perhaps activate the basal forebrain (Hemsworth and Mitchell 1969; Neal et al. 1968). This finding matches with our observation that the ES_ar-evoked cortical BF shift partially depended on the activity of the basal forebrain.

As described in METHODS, the spatial and temporal pattern of neural activity evoked by electric stimulation of the AC, somatosensory cortex, and basal forebrain might be quite different from that evoked by conditioning. However, the collicular and cortical BF shifts evoked by ES_at were both augmented by ES_st and/or ES_br, as expected by Gao and Suga's hypothesis (2000). A lesion of the basal forebrain prevented the augmentation of the ES_at-evoked collicular and cortical BF shifts that would otherwise be evoked by ES_st. These observations indicate that the somatosensory system does not directly evoke the augmentation but through the basal forebrain.

Our present data support Gao and Suga's hypothesis (1998, 2000), as discussed later in detail. However, there were differences between the BF shift evoked by the electric stimulation and conditioning. The collicular BF shifts evoked by the 30-min-long ES_at + ES_st or ES_at + ES_br was very similar in amount and recovery time to that caused by a 30-min-long conditioning session, whereas the collicular BF shift evoked by the 30-min-long ES_at + ES_st + ES_br was slightly larger in amount and longer in recovery time than that caused by the conditioning. The cortical BF shift evoked by electric stimulation was quite different in time course from that evoked by the conditioning, although the amount of the BF shift was similar to one another. Namely, cortical BF shift slowly develops, reaches a plateau ~180 min after the conditioning and shows no sign of recovery even 360 min after the conditioning, whereas the cortical BF shift evoked by electric stimulation was largest at the end of the stimulation and stayed large for ~30 min for the 30-min-long ES_at + ES_br or ES_at + ES_st, and for ~180 min for the 30-min-long ES_at + ES_st + ES_br. It always showed a recovery after the plateau. The cortical BF shift was always longer in recovery time than the collicular BF shift in identical stimulus conditions. However, the 30-min-long ES_ar + ES_st + ES_br did not evoke a long-lasting cortical BF shift as the 30-min-long auditory fear conditioning did (Gao and Suga 2000; Ji and Suga 2001).

The time course of the cortical BF shift caused by the conditioning is presumably due to a slow increase in a cortical ACh level and maintenance of the increased ACh level. As reported in our present paper, ES_st augmented the collicular and cortical BF shifts evoked by ES_at, the development of this augmentation was prevented by a lesion of the basal forebrain, and ES_br augmented the BF shifts evoked by ES_at. Ji et al. (2001) demonstrated that an ACh application to the AC makes the cortical BF shift long-term but not the collicular BF shift. Therefore it is clear that ACh plays an important role in determining the time course of the cortical BF shift. Direct electric stimulation of the somatosensory cortex and/or the basal forebrain together with the electric stimulation of the AC perhaps evoked a rapid increase in a cortical ACh level but did not maintain the increased ACh level to produce a long-lasting cortical BF shift. It may be predicted that a long-lasting ES_br produces a long-lasting cortical BF shift.

It has been suggested that the lateral amygdala is the place of the plasticity directly related to memory storage of fear conditioning and that short- and long-term memories caused by fear conditioning are, respectively, related to early and late phases of long-term potentiation (a review by Schafe et al. 2001). It has been known that auditory fear conditioning causes the long-term BF shifts of cortical auditory neurons (Bakin et al. 1996; Diamond and Weinberger 1986, 1989; Gao and Suga 2000; Ji et al. 2001; Ohl and Scheich 1996; Weinberger et al. 1993) and that electric stimulation of the cholinergic basal forebrain paired with acoustic stimuli evokes the long-lasting BF shifts of cortical auditory neurons (Bakin and Weinberger 1996; Kilgard and Merzenich 1998a). ACh apparently plays an important role in causing long-term BF shifts. However, it has not yet been known in what way ACh receptors are involved in producing long-term potentiation, although it has been suggested that cGMP may act as a second messenger for producing ACh-evoked depolarization (Woody et al. 1978). It is most likely that this depolarization lasts long and causes the cellular and molecular changes well studied by Kandel and his coworkers (review by Kandel 2001). In our present experiments, the basal forebrain was presumably not activated long enough by electric stimulation to produce the long-term BF shift of cortical neurons.

Our present data support Gao and Suga's hypothesis

Gao and Suga (1998, 2000) proposed a working hypothesis of the neural pathways for the plasticity of the IC and AC. Their hypothesis contains the following four key statements, which are supported by our current data.

STATEMENT 1. The central auditory system has an intrinsic mechanism for BF shift. The collicular BF shift is evoked by the corticofugal system working together with the cortical neural net.

STATEMENT 2. The cortical BF shift evoked by acoustic stimulation is augmented by the excitation of the somatosensory cortex following the excitation of the AC. This augmentation is presumably due to the excitation of the cholinergic basal forebrain that is evoked through the pathway from the somatosensory cortex to the association cortex, then to the amygdala and to the basal forebrain. The augmentation does not depend on the integration of auditory and somatosensory signals in the subcortical auditory nuclei.

STATEMENT 3. The cortical BF shift augmented by the cholinergic basal forebrain in turn augments the collicular BF shift through the corticofugal system.

STATEMENT 4. The cortical BF shift depends on both the subcortical BF shift and an increase in cortical ACh level.