Fear conditioning-induced plasticity in auditory thalamus and cortex : To what extent is it expressed during slow-wave sleep?

Acknowledgement: We warmly thank Jean-Marc Edeline for his constructive criticisms concerning this article.

The idea that sleep is involved in memory processing has received considerable attention in the recent years, rekindling interest in a potential link between sleep and brain plasticity (see Maquet, Smith, & Stickgold, 2003). A challenge for the future is thus to identify potential mechanisms whereby neurophysiological and neurochemical characteristics of sleep states could facilitate neuronal plasticity in relevant networks (for a review of candidate mechanisms, see Benington & Frank, 2003). However, the question of sleep and learning-dependent plasticity can also be addressed at a merely phenomenological level. We started this line of research 15 years ago by investigating whether neuronal plasticity induced by learning in wakefulness could be expressed during sleep. Obviously, this approach does not allow any inference about the role of sleep in brain plasticity. However, it provides an indication of the accessibility and reactivability of memory traces during sleep. It also constitutes an indirect way to assess whether the mechanisms that underlie physiological plasticity and permit its expression are preserved and operational during sleep. Because we had accumulated converging evidence for a role of paradoxical sleep (PS; also known as REM sleep) in memory processing (reviewed in Hennevin, 2003; Hennevin & Hars, 1985; Hennevin & Leconte, 1977), we first focused on that sleep state. We examined how neurons responded to a tone presented during PS before and after that tone had acquired significance through associative learning. We found that neurons in hippocampus (Maho, Hennevin, Hars, & Poincheval, 1991), auditory thalamus (Hennevin, Maho, & Hars, 1998; Hennevin, Maho, Hars, & Dutrieux, 1993), and amygdala (Hennevin et al., 1998) exhibited enhanced tone responsiveness during PS after auditory fear conditioning. Similar results were observed in auditory thalamus after appetitive conditioning (Maho & Hennevin, 2002). Thus, these studies indicate that the PS state allows expression of learning-induced plasticity. The present experiment examined whether it is also the case for slow-wave sleep (SWS).

As in most of our previous studies, the learning paradigm was classical fear conditioning, the most commonly used model to study emotional memory and neuronal plasticity. A tone (conditioned stimulus; CS) was paired with a footshock (unconditioned stimulus; US) during waking, and it was subsequently presented, at a nonawakening intensity, during SWS episodes. Multiunit recordings were collected in the medial division of the medial geniculate nucleus (MGm), a nonlemniscal division of the auditory thalamus, and in the primary auditory cortex (ACx). Indeed, an extensive literature has established that the learned behavioral significance of an acoustic stimulus is encoded as early as the thalamocortical auditory system (reviewed in Weinberger, 2004). Both the MGm and the primary ACx display associative plasticity during fear conditioning (and other learning situations as well): Within a few pairing trials, they develop plastic changes that are highly specific, robust, and long-lasting (reviewed in Edeline, 1999; Weinberger, 1995, 2004). From our previous studies (Hennevin et al., 1993, 1998), we knew that these changes, at least those in MGm, could be expressed during PS. Here we examined whether they could also be expressed during SWS.

Male Wistar rats (Iffa-Credo, Saint Germain-sur-l'Arbresle, France) weighing 300 to 350 g at the time of surgery were used. They were housed in individual cages, with continuous access to food and water, in a temperature-controlled room (23 ± 1 °C) on a 12-hr light–dark cycle (lights on at 8 a.m.). The experiments were conducted in the afternoon. All procedures were in accordance with the European legislation (86/609/EEC) on animal experimentation.

The surgery was conducted under sodium pentobarbital anesthesia (60 mg/kg ip), supplemented as necessary during implantation. Two extraduralcortical electrodes, made of small silver spheres, were placed along the interhemispheric suture with a large frontoparietal contralateral derivation for the recording of electrocorticographic (ECoG) activity. Two silver wires were inserted into the dorsal neck muscles for the recording of electromyographic (EMG) activity. A small brass screw fixed into the frontal bone was used as the ground. The recording electrodes implanted in MGm and ACx were Teflon-insulated tungsten wires (50 μm, ≈ 0.5–1 MΩ; A-M Systems, Everett, WA); two of them were inserted into a stainless steel microtube (outer diameter = 300 μm). Relative to bregma, the target sites for electrode placements were for MGm 5.8 mm posterior, 3 mm lateral, 6 mm ventral, and for ACx 4.8 mm posterior, 6.5 mm lateral, 4 mm ventral (Paxinos & Watson, 1986). The electrodes were lowered into both structures under electrophysiological control. Responses to pure tones delivered through hollow ear bars were used to optimize the final placement of the electrodes. All electrodes were connected to three miniature sockets fixed to the skull with dental acrylic cement.

The experimental box (25 cm long × 25 cm wide × 50 cm high) was located in a sound-attenuating chamber; both had a transparent front door that allowed the rat to be seen. Counterbalanced recording cables were relayed at the top of the experimental box through a multichannel rotating connector. The top of the box was equipped with a loudspeaker (5 cm in diameter, bandpass = 20–20000 kHz). The grid floor of the box was made of stainless steel rods, 0.5 cm in diameter, spaced 1.5 cm center to center. The scrambled electrical footshock used as US was delivered through the grid floor via an isolation unit placed in the sound-attenuating chamber on the side of the experimental box.

Neuronal activity was recorded through subminiature operational amplifiers (TL074 surface-mount package, Texas Instruments, Dallas, TX; input = 15 pA) located on the rats' heads at the extremity of the recording cables. The activity was amplified (Model P511K, Grass, Quincy, MA; gain = 10000), filtered (600–10000 Hz), displayed on an oscilloscope, and sent to voltage window discriminators (two-threshold triggers with temporal windows; see Courtice, 1975). The triggering level was adjusted under oscilloscope control to select the largest spikes (signal-to-noise ratio > 2:1). The EMG activity was fed to a preamplifier (bandpass = 1–75 Hz) of a Grass polygraph and sent to a window discriminator. The levels of all window discriminators were adjusted during the familiarization period when the rat was in quiet wakefulness; they were kept the same in the different phases of the protocol.

The output pulses of all of the triggers were stored on each trial by a laboratory microcomputer during the 1-s pretone period and the 2-s-after-tone onset. The time of occurrence of each pulse was recorded by the acquisition board at a precision of 50 μs. On each trial, custom-made software provided online rasters for four channels of neuronal activity (two from ACx, two from MGm) and for EMG activity. Off-line analysis allowed construction of standard peristimulus time histograms using any selected time bin, as well as quantification of tone responses on each channel by selection of temporal windows after tone onset. Trials on which a movement was detected in the awake rat during the 1-s pretone period, as well as trials on which the sleeping rat woke up during the pretone period or the 1-s period after tone onset, were discarded from analysis.

Familiarization period

For 2 days, each rat was familiarized with the experimental conditions. On the 2nd day, ECoG and EMG activities were monitored, and tone intensity that would be used subsequently was determined as follows. The tone (8 kHz, 2 s) was presented at different intensities (delivered in a random order) during SWS episodes. Ten to 15 tone presentations were given, and the highest intensity that did not wake the rat was determined. The intensity used during all subsequent experimental phases was 15 dB below that highest intensity; it varied among rats between 50 and 60 dB SPL.

Habituation and conditioning in wakefulness

For 4 consecutive days, at the same time of day, each rat was placed into the experimental box, and ECoG and EMG activities were permanently monitored. On the 1st day, all rats were submitted to a habituation session during which the tone (8 kHz, 2 s) was presented alone; 10 tone presentations were given. For the next 3 days, rats in the conditioned group were submitted to a session of classical conditioning during which the tone (2 s in duration) was immediately followed by a footshock (0.25–0.35 mA, 0.5 s); there were 10 tone–footshock pairings per session with variable intertrial intervals (range = 2–6 min). Rats in the pseudoconditioned group were given explicitly unpaired presentations of tone and footshock; there were 10 tone presentations separated by a varying interval of 2–6 min, and footshocks occurred randomly between the tones with a delay of 1 to 5 min after the tone; two successive tones or two successive shocks occasionally occurred.

Test trials in SWS

After the habituation session and each conditioning or pseudoconditioning session, the rat was kept in the experimental box and monitored for sleep phases. Ten test tones, 100 ms in duration, were presented during SWS. Tone duration was shortened compared with that used during conditioning in order to avoid any awakening during tone presentation. Indeed, in a pilot experiment, a 2-s tone was found to awake the rats too often. Tones were presented only when ECoG activity exhibited continuous high-voltage slow waves for a period of at least 30 s. They were distributed among several SWS episodes; the shortest intertone interval was 30 s. The duration of postsession recordings varied between 1.30 and 3 hr.

As in previous experiments (Hennevin et al., 1993, 1998; Maho & Hennevin, 2002), off-line analyses of neuronal data focused on the phasic “on” reponses, first to be sure that neuronal activity in waking was not contaminated by conditioned motor responses, and second because maximal conditioned discharges in MGm and ACx were expressed in the initial tens of milliseconds of the tone. A response window of 40 ms from tone onset was chosen, because this period included the whole response evoked at tone onset during SWS. Tone-evoked discharges were computed by subtracting the number of spikes per bin occurring during the 1-s pretone period from the number of spikes per bin occurring during the tone. Phasic excitatory responses occurring in the initial tens of milliseconds after tone offset were also observed during SWS owing to the short tone duration in that state. These “off” responses were not analyzed because they were not present in all of the MGm or ACx recordings.

The output pulses generated by triggering the EMG activity were analyzed in the same way as were the neuronal data. EMG reactivity to the tone was assessed by subtracting the number of counts per bin during the 1-s pretone period from the number of counts per bin during and after (for SWS) tone presentation, for each selected temporal window.

All statistical comparisons were performed on the averaged values obtained at a given block of trials, using contrast analysis of variance (Rouanet, Bernard, & Leroux, 1990). For neuronal data, each electrode served as a subject factor. For each structure, within-group comparisons were made to assess whether the tone-evoked responses recorded during (i.e., in waking) or after (i.e., in SWS) conditioning or pseudoconditioning (considering the three sessions pooled together or each session separately) differed from those obtained during or after habituation. Between-groups comparisons were also carried out to evaluate the associative nature of the changes induced by the conditioning procedure. Similar comparisons were performed for EMG data, using each rat as a subject factor.

At the end of the experiment, the rats were given an overdose of pentobarbital (100 mg/kg ip) and were perfused transcardially with 0.9% (wt/vol) saline (100 μl) followed by 500 μl cold (4 °C) fixative (4% [wt/vol] paraformaldehyde in 0.1 M phosphate buffer). The brains were removed and postfixed in the same fixative solution for a week. They were subsequently placed in 0.1 M phosphate buffer containing 10% (wt/vol) sucrose for 1 day, then 20% (wt/vol) sucrose for 3 days, both at 4 °C. Serial 50-μm coronal sections were cut on a freezing microtome and stained with cresyl violet.

The results were derived from a total of 19 rats, of which 15 had electrodes implanted in both the MGm and the ACx and 4 had electrodes implanted in only the MGm or only the ACx. Multiunit recordings were obtained from 15 electrodes in the MGm and from 14 electrodes in the ACx in the conditioned group (n = 11); they were from 11 electrodes in the MGm and from 11 electrodes in the ACx in the pseudoconditioned group (n = 8).

Behavioral data

As can be seen in Figure 1A, EMG reactivity during the first second of tone significantly changed across the four sessions in the conditioned group, F(3, 30) = 15.23, p < .001, whereas it did not in the pseudoconditioned group, F(3, 21) = 2.16, p = .12. This between-groups difference was supported by a significant Group × Session interaction, F(3, 51) = 5.79, p < .005. The conditioned rats exhibited increased reactivity to the tone as early as the first conditioning session, F(1, 10) = 8.55, p < .025, compared with the habituation level. This increase was more pronounced at the second session than at the first one, F(1, 10) = 9.34, p = .012; it was comparable at the second and third sessions, F(1, 10) < 1. More detailed analyses, focused on the first 100-ms bins after tone onset (see Figure 2A), indicated that EMG reactivity began to augment in the 100–200 ms after tone onset (compared with habituation, p < .05 for the first conditioning session, and ps < .01 for the next two ones). It remained unchanged during the first 100 ms after tone onset: The value obtained for that bin did not change across days, F(3, 30) < 1, and it did not differ from the habituation level at any of the three conditioning sessions, F(1, 10) < 1 for each comparison.

Tone responses in MGm

The 26 sites from which recordings were collected were located in the caudal part of MGm. Most of them (13 in the conditioned group and 9 in the pseudoconditioned group) were 6 mm posterior to bregma; the other four (two in each group) were 5.6 mm posterior to bregma (Paxinos & Watson, 1986).

In the conditioned group, the “on” response evoked during the first 40 ms after tone onset changed across sessions, F(3, 42) = 10.01, p < .001. It was higher during conditioning (the three sessions pooled together) than during habituation, F(1, 14) = 15.82, p < .005. As shown in Figure 3A, the discharges occurring during the 1st two 20-ms bins were increased; so it was for those occurring during the subsequent two bins (all ps < .025). The “on” response was enhanced at the first, second, and third conditioning sessions: compared with habituation, F(1, 14) = 8.76; F(1, 14) = 14.67; and F(1, 14) = 17.16, respectively, all ps ≤ .01. As in previous experiments (Hennevin et al., 1998; Maho & Hennevin, 2002), late and sustained increases in discharges, starting more than 100 ms after tone onset, were also observed during the conditioning sessions, specially the last two ones. Given their latency, these increases were probably due to feedback from the conditioned motor responses described earlier.

None of these changes were observed in the pseudoconditioned group. The “on” response did not change across sessions, F(3, 30) < 1; it did not differ between pseudoconditioning and habituation, F(1, 10) = 1.03, ns. Significant Group × Session interactions confirmed the difference between the two groups, F(3, 72) = 3.88, p = .012, and F(1, 24) = 6.57, p = .017.

Tone responses in ACx

The 25 recording sites were located in the temporal area TE1, from 4.3 to 5.3 mm posterior to bregma, which corresponds to the primary auditory cortical field (Doron, LeDoux, & Semple, 2002; Rutkowski, Miasnikov, & Weinberger, 2003; Sally & Kelly, 1988).

In the conditioned group, the “on” response evoked during the first 40 ms of tone changed across sessions, F(3, 39) = 18.70, p < .001. It was higher during conditioning than during habituation, F(1, 13) = 45.34, p < .001. As shown in Figure 4A, the discharges were increased during the 1st two 20-ms bins after tone onset, as well as during the next two bins (all ps < .05). The “on” response was enhanced at each of the three conditioning sessions: compared with habituation, F(1, 13) = 16.55; F(1, 13) = 26.48; and F(1, 13) = 44.23, respectively, all ps < .005.

In the pseudoconditioned group, the “on” response remained unchanged: It was not modified across sessions, F(3, 30) = 1.10, ns; it was comparable during pseudoconditioning and habituation, F(1, 10) < 1. The difference between the two groups was corroborated by significant Group × Session interactions, F(3, 69) = 8.30, and F(1, 23) = 16.10; p < .001 for each comparison.

Electromyographic data

Figures 1B and 2B show the EMG data obtained during the 1st s after tone onset. For statistical analysis, this 1-s period was divided into four bins: (a) the 100 ms of tone presentation (0–100 ms), (b) the first 100 ms after tone offset (100–200 ms), (c) the subsequent 300 ms (200–500 ms), and (d) the last 500 ms (500–1,000 ms).

In the conditioned group, tone presentation elicited no significant EMG change on the habituation day: The activity recorded during each time bin did not differ from mean pretone level: higher F value, F(1, 10) = 3.20, p = .10, for the 200-to-500-ms bin, during which EMG activity was lower than pretone value. Reactivity to the tone did not change across the 4 days: higher F value, F(3, 30) = 2.25, p = .11 for the 200-to-500-ms bin, p > .21 for all the other bins. It was not significantly different after conditioning and after habituation: higher F value, F(1, 10) = 4.03, p = .073 for the 200-to-500-ms bin; F(1, 10) < 1 for all the other bins.

In the pseudoconditioned group, EMG reactivity to the neutral tone was slightly higher than in the conditioned group during the 0-to-100-ms bin, F(1, 17) = 3.93, p = .064, and during the 200-to-500-ms bin, F(1, 17) = 7.77, p = .012. This reactivity did not significantly change across days: higher F value, F(3, 21) = 2.66, p = .074, for the 200-to-500-ms bin. It was not different after pseudoconditioning and after habituation: higher F value, F(1, 7) = 2.15, ns, for the 500-to-1,000-ms bin.

Thus, no change in EMG reactivity occurred during the 100 ms of tone presentation. Further, over the whole second after tone onset, reactivity to the conditioned tone remained similar to that expressed to the neutral tone, F(1, 10) < 1, for the comparison between EMG responses after conditioning and after habituation, and it was comparable to that expressed after pseudoconditioning, F(1, 17) < 1, for the interaction between the factors of group and treatment day (habituation vs. conditioning or pseudoconditioning).

Tone responses in MGm: Group data

On the habituation day and for the whole set of 26 recordings, the evoked activity recorded during the 100 ms of the tone was reduced compared with wakefulness, F(1, 25) = 7.22, p = .012. The discharges occurring during the first 40 ms were not significantly decreased, F(1, 25) = 2.20, ns, but those occurring during each of the next three 20-ms bins were (all ps < .05). As can be seen in Figure 3B, phasic increases in discharges occurred at tone offset. They were observed after habituation, conditioning, and pseudoconditioning. As shown in Figure 5, this response was time locked to tone offset; it occurred with a latency comparable to that of the “on” response; it was of short duration, lasting only 15–20 ms. From these characteristics it is clear that these increased discharges did not reflect an arousal reaction but corresponded to an excitatory “off” response, as typically observed in auditory thalamus and cortex (Calford, 1983; He, 2002; Recanzone, 2000). This response was present in half of the MGm recordings.

Spontaneous activity recorded during the 1-s pretone period did not change across days in either group, F(3, 42) < 1, and F(3, 30) = 2.04, ns, for the conditioned and the pseudoconditioned group, respectively. This was also the case for the tone-evoked response. In the conditioned group, the “on” response evoked in the first 40 ms of tone remained unchanged across days, F(3, 42) < 1, and was comparable after conditioning and after habituation, F(1, 14) = 1.18, ns. Similarly, no change was observed after pseudoconditioning, F(3, 30) < 1, and F(1, 10) < 1. Therefore, the responses of MGm neurons in SWS did not differ after conditioning and after pseudoconditioning, F(3, 72) < 1, and F(1, 24) < 1, for the Group × Day interactions. Similar results were obtained when the whole response evoked during the 100 ms of the tone was considered, F(3, 72) < 1, and F(1, 24) < 1.

Tone responses in ACx: Group data

On the habituation day and for the whole set of 25 recordings, the evoked activity recorded during the 100 ms of the tone was reduced compared with waking, F(1, 24) = 33.42, p < .001. Evoked discharges were not significantly decreased during the first 20-ms bin, F(1, 24) = 1.15, ns, but they were during each of the subsequent three bins (all ps ≤ .014; for the fifth 20-ms bin, p = .085). As in MGm, offset responses were also observed after habituation, conditioning, and pseudoconditioning (see Figure 4B); they were present in less than half of the ACx recordings.

In the conditioned and the pseudoconditioned group, spontaneous pretone activity did not change across days, F(3, 39) < 1, and F(3, 30) < 1, respectively. The response evoked in the first 40 ms of the tone was also unchanged, F(3, 39) < 1, and F(3, 30) = 1.96, ns. It did not differ before and after conditioning, F(1, 13) = 2.33, ns, or before and after pseudoconditioning, F(1, 10) = 2.36, ns. There was therefore no between-groups difference, F(3, 69) < 1, and F(1, 23) < 1, for the Group × Day interactions. It was also the case when the whole tone duration was considered, F(3, 69) = 1.22, ns, and F(1, 23) < 1. Thus, contrasting with the results obtained during waking, group data did not show enhanced responsiveness to the tone CS during SWS, whether in MGm or in ACx.

Tone responses in MGm and ACx: Further analyses

A possibility that could account for the absence of response changes during posttraining SWS is that neuronal plasticity induced during wakefulness was not sufficiently robust to be maintained and expressed in another behavioral state. To address this issue, we compared the conditioned changes obtained in MGm during waking with those obtained in a previous experiment. In that study (Hennevin et al., 1998), using exactly the same conditioning procedure and the same recording techniques as here, we found that MGm neurons expressed increased discharges when the tone CS was presented during posttraining PS. Figure 6 shows that the distribution and magnitude of the conditioned changes expressed during conditioning were quite comparable in the two studies. This similarity suggests that the present failure to detect enhanced responding in MGm during SWS did not result from substandard plasticity induced during waking.

Nonetheless, further analyses indicated that the expression of conditioned responses in SWS did depend on the strength of the conditioned changes. First, the response changes observed in MGm during posttraining SWS were correlated with those observed during conditioning, r(15) = .691, p = .003. This contrasts with the total absence of correlation between the changes occurring during SWS and those occurring during waking in the pseudoconditioned group, r(11) = −.007. Therefore, on the basis of the scatter diagram presented in Figure 7A, MGm recordings were divided into two subsets: One included the five recordings that exhibited the largest response increases during conditioning (they corresponded to the five dots located in the upper right quadrant of Figure 7A); the other subset included the other seven recordings that also displayed discharge increases during conditioning. Three recordings were discarded because they did not show increased discharges in the first 40 ms of tone (they corresponded to the three dots located at the left end of Figure 7A). As shown in Figure 8A, the subset of five recordings that exhibited the largest conditioned responses in waking showed enhanced responsiveness during SWS: The “on” response tended to change across days, F(3, 12) = 3.09, p = .067, and it was significantly larger after conditioning than after habituation, F(1, 4) = 29.42, p < .01.

The same analyses were performed for the cortical responses. Although there was no correlation between the changes observed during SWS and during conditioning, r(14) = .041, ACx recordings were divided into two subsets: One included the six recordings that exhibited the largest response increases during conditioning (in B Figure 7B, they corresponded to the six recordings for which discharge augmentation was larger than 1 spike/20 ms); the other included the other seven recordings that also exhibited increased discharges during waking. One recording was discarded because it did not show discharge changes in the first 40 ms of tone (it corresponded to the dot located at the left end of Figure 7B). Unlike in MGm, the subset of recordings showing the largest increases during conditioning displayed no response changes during SWS (see Figure 8C). However, it must be stressed that the discharge increase during conditioning was weaker in that subset than in the corresponding MGm subset (shown in Figure 8A), F(1, 9) = 19.07, p = .0018. Indeed, the largest increase observed in ACx only reached the level of the smallest increase observed in the subset of five MGm recordings (cf. Figures 7A and 7B).

Although tone responses in MGm and ACx were enhanced during the three conditioning sessions, overall they failed to show any significant changes during SWS. This failure contrasts with previous results demonstrating that fear-conditioning-induced plasticity in MGm was expressed during PS (Hennevin et al., 1993, 1998). However, examination of individual data revealed that the recordings that displayed the largest conditioned changes exhibited enhanced responding in SWS, indicating that expression of discharge plasticity in SWS depended on the strength of learning-induced changes.

During wakefulness, increases in evoked responses were observed only in the conditioned group, not in the pseudoconditioned group, which attested to their associative nature. As these increases were already present in the first 40 ms of tone, they did not result from movements and/or behavioral feedback: Myographic activities did not change in the first 100 ms of tone, conditioned motor responses starting only in the subsequent 100 ms. Furthermore, in line with pioneering studies (Buchwald, Halas, & Schramm, 1966; Disterhoft & Stuart, 1976; Gabriel, Saltwick, & Miller, 1975; Oleson, Ashe, & Weinberger, 1975; Ryugo & Weinberger, 1978), an extensive literature using a variety of learning paradigms has conclusively established that response plasticity reliably occurs in the MGm (Edeline & Weinberger, 1992; Lennartz & Weinberger, 1992; McEchron, McCabe, Green, Llabre, & Schneiderman, 1995; O'Connor, Allison, Rosenfield, & Moore, 1997; Supple & Kapp, 1989) and in the primary ACx (Bakin, South, & Weinberger, 1996; Bakin & Weinberger, 1990; Edeline & Weinberger, 1993; Gao & Suga, 2000; Kisley & Gerstein, 2001; Ohl & Scheich, 1996). Thus, all of the aforementioned supports the conclusion that the enhanced responsiveness expressed in MGm and ACx during waking reflected physiological plasticity induced by associative processes.

Tone responses were reduced during SWS. This decrease was mainly due to suppression of the sustained activity that followed the “on” response in waking; the early response component occurring in the first 40 ms was less affected. Most of the single-unit studies that examined sleep-dependent changes in sensory systems showed that evoked activity in SWS was predominantly depressed at the thalamic level (Coenen & Vendrik, 1972; Edeline, Manunta, & Hennevin, 2000; Livingstone & Hubel, 1981; Mariotti, Formenti, & Mancia, 1989) and at the cortical level (Edeline, Dutrieux, Manunta, & Hennevin, 2001; Evarts, 1963; Gücer, 1979). However, these studies also showed that the magnitude, and even the direction, of changes varied considerably from cell to cell (see, e.g., Edeline et al., 2001; Pena, Pérez-Perera, Bouvier, & Velluti, 1999). As a result, when the discharges of several neurons are combined, as was the case in the present experiment, the observed decrease is less pronounced than what can be detected with single-unit recordings. Last, evidence from various preparations indicates that the early components of sensory responses are less sensitive to state changes than are the late components (Evarts, 1963; Funke & Eysel, 1992; Imig, Weinberger, & Westenberg, 1972).

Admittedly, testing tone responses during SWS after fear conditioning without awaking the rat was far from being simple. The arousing properties of a stimulus closely depend on its emotional salience and behavioral relevance. Thereby, sensory thresholds for awakening from sleep in general and SWS in particular are lowered after a stimulus has acquired significance through associative learning (Ciancia, Trigona-Leisinger, & Bloch, 1980; Halperin & Iorio, 1981; Siegel & Langley, 1965; Van Twyver & Garrett, 1972). This is particularly true for a tone CS signaling a threatening event such as footshock. Because it was crucial that no awakening occurred during tone presentation, tone duration in SWS was reduced to 100 ms. In addition, we used a tone of lower intensity than that used in a previous experiment (Hennevin et al., 1998) in which the tone was presented during PS (arousal thresholds are higher during PS than during SWS). This lower intensity did not decrease the potential associability of the tone CS with footshock: Behavioral and MGm neuronal responses were conditioned as rapidly in the two studies, and the magnitude of conditioned changes in MGm was also comparable.

On the basis of a systematic and careful inspection of ECoG tracings and on quantification of nuchal EMG activity, no awakening from SWS was detected during tone presentation and even over the 1-s period after tone onset. Obviously, we cannot assert that microarousals never occurred. Quantitative analysis would have possibly revealed subtle modifications in the spectral composition of the ECoG. However, the overall absence of changes in the pattern and amplitude of evoked responses in MGm and ACx, as well as in posttone firing activity, suggests that microarousal reactions to the conditioned tone, if any, were rare. Actually, the question of possible arousal effects on tone responsiveness is pertinent only for the recordings that displayed response changes in postlearning SWS, particularly the five MGm recordings that exhibited the largest discharge increases during conditioning and enhanced responses during SWS. It is improbable that the enhancement observed in that subset was due to arousal from SWS, for the following reason. In the 4 rats from which these five recordings were obtained, two other MGm recordings and five ACx recordings were simultaneously collected. If arousal from SWS had occurred, then response changes should have been observed also for those other recordings. This was not the case: Compared with habituation levels, the response changes observed after conditioning for the two MGm recordings were –0.01 and 0.38 spikes/20 ms; the mean change obtained for the five ACx recordings was 0.14 spikes/20 ms. Therefore, we can reasonably assume that the enhanced responding expressed by the five-MGm-recording subset did reflect learning-induced plasticity.

No ACx recording subset showed increased responses during SWS. This difference between ACx and MGm is probably more apparent than real. First, because the robustness of associative plasticity in ACx has been largely demonstrated (Galván & Weinberger, 2002; Weinberger, Javid, & Lepan, 1993), there is no reason to suspect that plastic changes in ACx would be less easily transferred from waking to SWS than plastic changes in MGm. Second, the possibility that active inhibitory processes acting during SWS prevented the expression of plastic changes more in ACx than in MGm cannot be excluded but has no experimental support. In fact, inhibitory processes in SWS are more prevalent at the thalamic level than at the cortical level: Compared with the global inhibition of thalamic neurons, due to the hyperpolarizing pressure from the thalamic reticular neurons, cortical neurons have a relatively rich activity during SWS and they even display, periodically, epochs of enhanced excitability (Steriade, 2001, 2003). Third, although all but one ACx recording displayed conditioning, none of them exhibited discharge increases as large as the best conditioned responses in MGm, maybe as a consequence of their lower initial values during habituation. Last, tone responding in SWS was not increased either for MGm recordings that showed conditioned changes of moderate magnitude. Thus, the most parsimonious and likely interpretation is that discharge increases induced by our conditioning procedure in ACx were not sufficiently strong to enable them to be maintained and expressed during SWS.

That neuronal plasticity is or is not transferred across physiological states depending on the strength of conditioning-induced plastic changes is per se far from surprising, of course. However, in previous studies we systematically found that response plasticity was transferred from waking to PS; this was observed for conditioned responses in the hippocampus, the amygdala, and the MGm (Hennevin et al., 1993, 1998; Maho & Hennevin, 2002; Maho et al., 1991). To check whether the effects of conditioning were equivalent in the present experiment and the previous ones, we directly compared the present results with those obtained in a previous study (Hennevin et al., 1998). Whereas response changes expressed in MGm during conditioning were similar in the two studies, significant plasticity was observed across the population during PS, but not during SWS (see Figure 6). Therefore, reasoning solely in terms of weak plasticity in wakefulness to account for the failure to reliably observe plasticity in SWS would be an oversimplification. Other intervening factors were involved. Two possibilities can be envisioned.

It might be speculated that an extinction process, due to repeated tones in the absence of the US, developed during SWS. As a result, conditioned responses would have diminished over trials and days; only the strongest of them would have been partially preserved and, thus, detected. This possibility is unlikely for several reasons. First, fear-conditioning-induced responses do not easily extinguish: For example, in awake rats, conditioned onset responses in ACx were still retained after 30 extinction trials (Quirk, Armony, & LeDoux, 1997; see also Maho, Hars, Edeline, & Hennevin, 1995). Second, the possibility that an extinction process can indeed develop during SWS has never been demonstrated (for a recent tentative experiment, see Coenen & Drinkenburg, 2002). Third, if we suppose that extinction occurred during SWS in the present study, then two additional, and questionable, assumptions have to be made to account for the fact that conditioning displayed the same time course as in our previous studies: (a) The effects of extinction did not transfer from SWS to waking (whereas the effects of conditioning would have transferred from waking to SWS), and (b) extinction did not develop during PS. Last, an extinction explanation would predict a progressive response decrement across trials in SWS; such an effect was not detected (data not shown). It would also predict a response attenuation across days; the opposite effect was observed: For the five-MGm-recording subset that exhibited conditioned responses in SWS, their responses increased across sessions (see Figure 8A).

The other possibility, much more likely, is that the physiological conditions prevailing in the thalamocortical system during SWS make the expression of response plasticity difficult. Whereas waking and PS are similar in a number of respects (Llinás & Paré, 1991), SWS is a fundamentally different state. Thalamic and cortical neurons in SWS display prolonged periods of hyperpolarization and fire rhythmically with a high proportion of spike bursts; thalamocortical circuits display coherent, synchronous oscillatory activities (Steriade, 2003). To what extent such patterns of activity could have compromised the expression of discharge plasticity in the present experiment is unknown, but this possibility cannot be excluded. State-specific changes in neuronal properties and activities are under the control of generalized modulatory systems (McCormick & Bal, 1997; Pace-Schott & Hobson, 2002). Among neuromodulators, the one that distinguishes SWS from both waking and PS is acetylcholine (ACh) which, in contrast to waking and PS, is released at low levels during SWS. These between-states differences in ACh levels were observed in many brain regions, including thalamus (Kodama & Honda, 1996; Williams, Comisarow, Day, Fibiger, & Reiner, 1994) and cortex (Jasper & Tessier, 1971; Marrosu et al., 1995). Multiple lines of evidence indicate that ACh promotes neuronal plasticity. In particular, a number of studies using iontophoretic application of ACh or stimulation or lesion of the nucleus basalis (the major source of cortical ACh) have demonstrated the involvement of ACh in the induction of sensory cortex plasticity (reviewed in Dykes, 1997; Gu, 2003; Rasmusson, 2000; Weinberger, 2003). Of interest, there is also evidence for an involvement of ACh in the expression of neuronal plasticity. Pairing whisker stimulation with iontophoretically applied ACh in somatosensory cortex yielded selective changes of response that required the presence of ACh to be expressed after pairing (Ego-Stengel, Shulz, Haidarliu, Sosnik, & Ahissar, 2001; Shulz, Sosnik, Ego, Haidarliu, & Ahissar, 2000). This ACh-dependent expression of plasticity has been viewed as a cellular analog of the state-dependent-learning phenomenon described for a long time at the behavioral level (Overton, 1964, 1985). In the present experiment, neuronal plasticity was induced while thalamic and cortical ACh levels were very high (during training they were probably above waking normal levels; see Acquas, Wilson, & Fibiger, 1996; Maho et al., 1995). In contrast, at the time of testing, that is, during SWS, ACh levels were minimal. This discordance, and more generally the dramatic functional changes that occur from waking to SWS in the thalamocortical system, could explain why thalamic and cortical plasticity induced during waking was not easily expressed during SWS.

An increasing amount of evidence supports the view that SWS facilitates the consolidation of memories by providing suitable conditions for reactivating memory traces and reinforcing neural plasticity initiated during waking (e.g., Buzsáki, 1998; Gais & Born, 2004a; Sejnowski & Destexhe, 2000; Steriade & Timofeev, 2003; Sutherland & McNaughton, 2000; Wilson, 2002). For example, in line with earlier predictions (Buzsáki, 1989; Hasselmo, 1999; see also Hasselmo & McGaughy, 2004), a low level of ACh during SWS was shown to be necessary for SWS-related declarative memory consolidation to take place in humans (Gais & Born, 2004b). By showing that learning-induced plasticity was not easily expressed during SWS, the present results could, at first glance, seem inconsistent with the spontaneous replay of memory traces reported in the aforementioned literature. However, it can be reasonably supposed that stimulus-elicited expression of associative plastic changes in thalamocortical sensory systems after fear conditioning, on the one hand, and spontaneous reactivation of ensemble activity patterns in hippocampocortical circuits after hippocampus-dependent tasks, on the other hand, might require different cerebral conditions. Various arguments give support to this possibility (see Smith, 2001; Walker & Stickgold, 2004). For example, neuroimaging studies showed that, depending on the learning task, reactivation of different brain regions occurred during different sleep states: After a procedural reaction time task, increased activity in the cuneus and premotor cortex was observed during PS, but not during SWS (Maquet et al., 2000; Peigneux et al., 2003); on the other hand, after a declarative spatial task, increased activity in the hippocampal formation was observed during SWS, but not during PS (Peigneux et al., 2004). Present results indicate that in the case of emotional learning, SWS is a state less propitious than PS for expression of associative plasticity in the thalamocortical auditory system.

Acquas, E., Wilson, C., & Fibiger, H. C. (1996). Conditioned and unconditioned stimuli increase frontal cortical and hippocampal acetylcholine release: Effects of novelty, habituation, and fear. Journal of Neuroscience, 16, 3089–3096.

Bakin, J. S., South, D. A., & Weinberger, N. M. (1996). Induction of receptive field plasticity in the auditory cortex of the guinea pig during instrumental avoidance conditioning. Behavioral Neuroscience, 110, 905–913.

Bakin, J., & Weinberger, N. M. (1990). Classical conditioning induces CS-specific receptive field plasticity in the auditory cortex of the guinea pig. Brain Research, 536, 271–286.

Benington, J. H., & Frank, M. G. (2003). Cellular and molecular connections between sleep and synaptic plasticity. Progress in Neurobiology, 69, 71–101.

Buchwald, J. S., Halas, E., & Schramm, S. (1966). Changes in cortical and subcortical unit activity during behavioral conditioning. Physiology & Behavior, 1, 11–22.

Buzsáki, G. (1989). Two-stage model of memory trace formation: A role for “noisy” brain states. Neuroscience, 31, 551–570.

Buzsáki, G. (1998). Memory consolidation during sleep: A neurophysiological perspective. Journal of Sleep Research, 7, (Suppl. 1)17–23.

Calford, M. B. (1983). The parcellation of medial geniculate body of the cat defined by the auditory response properties of single units. Journal of Neuroscience, 3, 2350–2364.

Ciancia, F., Trigona-Leisinger, M. C., & Bloch, V. (1980). Etude, chez le chat, de la réactivité aux stimulus auditifs signifiants en fonction des niveaux de vigilance [Study, in the cat, of reactivity to significant auditory stimuli as a function of vigilance state]. Archives Italiennes de Biologie, 118, 375–385.

Coenen, A. M. L., & Drinkenburg, W. H. I. M. (2002). Animal models for information processing during sleep. International Journal of Psychophysiology, 46, 163–175.

Coenen, A. M. L., & Vendrik, A. J. H. (1972). Determination of the transfer ratio of cat's geniculate neurons through quasi-intracellular recordings and the relation with the level of alertness. Experimental Brain Research, 14, 227–242.

Courtice, C. J. (1975). Action potential selection by amplitude and width. Journal of Physiology, 246, 18P–19P.

Disterhoft, J. F., & Stuart, D. K. (1976). Trial sequence of changed unit activity in auditory system of alert rat during conditioned response acquisition and extinction. Journal of Neurophysiology, 39, 266–281.

Doron, N. N., LeDoux, J. E., & Semple, M. N. (2002). Redefining the tonotopic core of rat auditory cortex: Physiological evidence for a posterior field. Journal of Comparative Neurology, 453, 345–360.

Dykes, R. W. (1997). Mechanisms controlling neuronal plasticity in somatosensory cortex. Canadian Journal of Physiological Pharmacology, 75, 535–545.

Edeline, J.-M. (1999). Learning-induced physiological plasticity in the thalamo-cortical sensory systems: Evaluation of receptive field plasticity, map changes and their potential mechanisms. Progress in Neurobiology, 57, 165–224.

Edeline, J.-M., Dutrieux, G., Manunta, Y., & Hennevin, E. (2001). Diversity of receptive field changes in auditory cortex during natural sleep. European Journal of Neuroscience, 14, 1865–1880.

Edeline, J.-M., Manunta, Y., & Hennevin, E. (2000). Auditory thalamus neurons during sleep: Changes in frequency selectivity, threshold, and receptive field size. Journal of Neurophysiology, 84, 934–952.

Edeline, J.-M., & Weinberger, N. M. (1992). Associative retuning in the thalamic source of input to the amygdala and auditory cortex: Receptive field plasticity in the medial division of the medial geniculate body. Behavioral Neuroscience, 106, 81–105.

Edeline, J.-M., & Weinberger, N. M. (1993). Receptive field plasticity in the auditory cortex during frequency discrimination training: Selective retuning independent of task difficulty. Behavioral Neuroscience, 107, 82–103.

Ego-Stengel, V., Shulz, D. E., Haidarliu, S., Sosnik, R., & Ahissar, E. (2001). Acetylcholine-dependent induction and expression of functional plasticity in the barrel cortex of adult rat. Journal of Neurophysiology, 86, 422–437.

Evarts, E. V. (1963). Photically evoked responses in visual cortex units during sleep and waking. Journal of Neurophysiology, 26, 229–248.

Funke, K., & Eysel, U. T. (1992). EEG-dependent modulation of response dynamics of cat dLGN relay cells and the contribution of corticogeniculate feedback. Brain Research, 573, 217–227.

Gabriel, M., Saltwick, S. E., & Miller, J. D. (1975, September26). Conditioning and reversal of short-latency multiple-unit responses in the rabbit medial geniculate nucleus. Science, 189, 1108–1109.

Gais, S., & Born, J. (2004a). Declarative memory consolidation: Mechanisms acting during human sleep. Learning & Memory, 11, 679–685.

Gais, S., & Born, J. (2004b). Low acetylcholine during slow-wave sleep is critical for declarative memory consolidation. Proceedings of the National Academy of Sciences, USA, 101, 2140–2144.

Galván, V. V., & Weinberger, N. M. (2002). Long-term consolidation and retention of learning-induced tuning plasticity in the auditory cortex of the guinea pig. Neurobiology of Learning and Memory, 77, 78–108.

Gao, E., & Suga, N. (2000). Experience-dependent plasticity in the auditory cortex and the inferior colliculus of bats: Role of the corticofugal system. Proceedings of the National Academy of Sciences, USA, 97, 8081–8086.

Gu, Q. (2003). Contribution of acetylcholine to visual cortex plasticity. Neurobiology of Learning and Memory, 80, 291–301.

Gücer, G. (1979). The effect of sleep upon the transmission of afferent activity in the somatic afferent system. Experimental Brain Research, 34, 287–298.

Halperin, J. M., & Iorio, L. C. (1981). Responsivity of rats to neutral and danger-signaling stimuli during sleep. Behavioral and Neural Biology, 33, 213–219.

Hasselmo, M. E. (1999). Neuromodulation: Acetylcholine and memory consolidation. Trends in Cognitive Sciences, 3, 351–359.

Hasselmo, M. E., & McGaughy, J. (2004). High acetylcholine levels set circuit dynamics for attention and encoding and low acetylcholine levels set dynamics for consolidation. Progress in Brain Research, 145, 207–231.

He, J. (2002). OFF responses in the auditory thalamus of the guinea pig. Journal of Neurophysiology, 88, 2377–2386.

Hennevin, E. (2003). Expression and modulation of memory traces during paradoxical sleep. In P.Maquet, C.Smith, & R.Stickgold (Eds.), Sleep and brain plasticity (pp. 101–116). New York: Oxford University Press.

Hennevin, E., & Hars, B. (1985). Post-learning paradoxical sleep: A critical period when new memory is reactivated?. In B. E.Will, P.Schmitt, & J. C.Dalrymple-Alford (Eds.), Brain plasticity, learning and memory (pp. 193–203). New York: Plenum Press.

Hennevin, E., & Leconte, P. (1977). Etude des relations entre le sommeil paradoxal et les processus d'acquisition [Relationships between paradoxical sleep and the acquisition process]. Physiology & Behavior, 18, 307–319.

Hennevin, E., Maho, C., & Hars, B. (1998). Neuronal plasticity induced by fear conditioning is expressed during paradoxical sleep: Evidence from simultaneous recordings in the lateral amygdala and the medial geniculate in rats. Behavioral Neuroscience, 112, 839–862.

Hennevin, E., Maho, C., Hars, B., & Dutrieux, G. (1993). Learning-induced plasticity in the medial geniculate nucleus is expressed during paradoxical sleep. Behavioral Neuroscience, 107, 1018–1030.

Imig, T. J., Weinberger, N. M., & Westenberg, I. S. (1972). Relationships among unit discharge rate, pattern, and phasic arousal in the medial geniculate nucleus of the waking cat. Experimental Neurology, 35, 337–357.

Jasper, H. H., & Tessier, J. (1971, May7). Acetylcholine liberation from cerebral cortex during paradoxical (REM) sleep. Science, 172, 601–602.

Kisley, M. A., & Gerstein, G. L. (2001). Daily variation and appetitive conditioning-induced plasticity of auditory cortex receptive fields. European Journal of Neuroscience, 13, 1993–2003.

Kodama, T., & Honda, Y. (1996). Acetylcholine releases of mesopontine PGO-on cells in the lateral geniculate nucleus in sleep–waking cycle and serotonergic regulation. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 20, 1213–1227.

Lennartz, R. C., & Weinberger, N. M. (1992). Frequency-specific receptive field plasticity in the medial geniculate body induced by Pavlovian fear conditioning is expressed in the anesthetized brain. Behavioral Neuroscience, 106, 484–497.

Livingstone, M. S., & Hubel, D. H. (1981). Effects of sleep and arousal on the processing of visual information in the cat. Nature, 291, 554–561.

Llinás, R. R., & Paré, D. (1991). Of dreaming and wakefulness. Neuroscience, 44, 521–535.

Maho, C., Hars, B., Edeline, J.-M., & Hennevin, E. (1995). Conditioned changes in the basal forebrain: Relations with learning-induced cortical plasticity. Psychobiology, 23, 10–25.

Maho, C., & Hennevin, E. (2002). Appetitive conditioning-induced plasticity is expressed during paradoxical sleep in the medial geniculate, but not in the lateral amygdala. Behavioral Neuroscience, 116, 807–823.

Maho, C., Hennevin, E., Hars, B., & Poincheval, S. (1991). Evocation in paradoxical sleep of a hippocampal conditioned cellular response acquired during waking. Psychobiology, 19, 193–205.

Maquet, P., Laureys, S., Peigneux, P., Fuchs, S., Petiau, C., Phillips, C., et al. (2000). Experience-dependent changes in cerebral activation during human REM sleep. Nature Neuroscience, 3, 831–836.

Maquet, P., Smith, C., & Stickgold, R. (2003). Sleep and brain plasticity. New York: Oxford University Press.

Mariotti, M., Formenti, A., & Mancia, M. (1989). Responses of VPL thalamic neurons to peripheral stimulation in wakefulness and sleep. Neuroscience Letters, 102, 70–75.

Marrosu, F., Portas, C., Mascia, M. S., Casu, M. A., Fa, M., Giagheddu, M., et al. (1995). Microdialysis measurement of cortical and hippocampal acetylcholine release during sleep–wake cycle in freely moving cats. Brain Research, 671, 329–332.

McCormick, D. A., & Bal, T. (1997). Sleep and arousal: Thalamocortical mechanisms. Annual Review of Neuroscience, 20, 185–215.

McEchron, M. D., McCabe, P. M., Green, E. J., Llabre, M. M., & Schneiderman, N. (1995). Simultaneous single unit recording in the medial nucleus of the medial geniculate nucleus and amygdaloid central nucleus throughout habituation, acquisition, and extinction of the rabbit's classically conditioned heart rate. Brain Research, 682, 157–166.

O'Connor, K. N., Allison, T. L., Rosenfield, M. E., & Moore, J. W. (1997). Neural activity in the medial geniculate nucleus during auditory trace conditioning. Experimental Brain Research, 113, 534–556.

Ohl, F. W., & Scheich, H. (1996). Differential frequency conditioning enhances spectral contrast sensitivity of units in auditory cortex (Field A1) of the alert Mongolian gerbil. European Journal of Neuroscience, 8, 1001–1017.

Oleson, T., Ashe, J., & Weinberger, N. M. (1975). Modification of auditory and somatosensory activity during pupillary conditioning in the paralysed cat. Journal of Neurophysiology, 38, 1114–1139.

Overton, D. A. (1964). State-dependent or “dissociated” learning produced with pentobarbital. Journal of Comparative Physiological Psychology, 57, 3–12.

Overton, D. A. (1985). Contextual stimulus effects of drugs and internal states. In P. D.Balsam & A.Tomie (Eds.), Context and learning (pp. 357–384). Hillsdale, NJ: Erlbaum.

Pace-Schott, E. F., & Hobson, J. A. (2002). The neurobiology of sleep: Genetics, cellular physiology and subcortical networks. Nature Reviews Neuroscience, 3, 591–605.

Paxinos, G., & Watson, C. (1986). The rat brain in stereotaxic coordinates (2nd ed.). San Diego, CA: Academic Press.

Peigneux, P., Laureys, S., Fuchs, S., Collette, F., Perrin, F., Reggers, J., et al. (2004). Are spatial memories strengthened in the human hippocampus during slow wave sleep?Neuron, 44, 535–545.

Peigneux, P., Laureys, S., Fuchs, S., Destrebecqz, A., Collette, F., Delbeuck, X., et al. (2003). Learned material content and acquisition level modulate cerebral reactivation during posttraining rapid-eye-movements sleep. NeuroImage, 20, 125–134.

Pena, J. L., Pérez-Perera, P. L., Bouvier, M., & Velluti, R. A. (1999). Sleep and wakefulness modulation of the neuronal firing in the auditory cortex of the guinea pig. Brain Research, 816, 463–470.

Quirk, G. J., Armony, J. L., & LeDoux, J. E. (1997). Fear conditioning enhances different temporal components of tone-evoked spike trains in auditory cortex and lateral amygdala. Neuron, 19, 613–624.

Rasmusson, D. D. (2000). The role of acetylcholine in cortical synaptic plasticity. Behavioral Brain Research, 115, 205–218.

Recanzone, G. H. (2000). Response profiles of auditory cortical neurons to tones and noise in behaving macaque monkeys. Hearing Research, 150, 104–118.

Rouanet, H., Bernard, J. M., & Leroux, B. (1990). Statistiques en sciences humaines: Analyse inductive des données. [Statistics in human sciences: Inductive analysis of constants]. Paris: Dunot.

Rutkowski, R. G., Miasnikov, A. A., & Weinberger, N. M. (2003). Characterisation of multiple physiological fields within the anatomical core of rat auditory cortex. Hearing Research, 181, 116–130.

Ryugo, D. K., & Weinberger, N. M. (1978). Differential plasticity of morphologically distinct neuron populations in the medial geniculate of the cat during classical conditioning. Behavioral Biology, 22, 275–301.

Sally, S. L., & Kelly, J. B. (1988). Organization of auditory cortex in the albino rat: Sound frequency. Journal of Neurophysiology, 59, 1627–1638.

Sejnowski, T. J., & Destexhe, A. (2000). Why do we sleep?Brain Research, 886, 208–223.

Shulz, D. E., Sosnik, R., Haidarliu, S., & Ahissar, E. (2000). A neuronal analogue of state-dependent learning. Nature, 403, 549–553.

Siegel, J., & Langley, T. D. (1965). Arousal threshold in the cat as a function of sleep phase and stimulus significance. Experientia, 21, 740–741.

Smith, C. (2001). Sleep states and memory processes in humans: Procedural versus declarative memory systems. Sleep Medicine Reviews, 5, 491–506.

Steriade, M. (2001). Active neocortical processes during quiescent sleep. Archives Italiennes de Biologie, 139, 37–51.

Steriade, M. (2003). The corticothalamic system in sleep. Frontiers in Bioscience, 8, 878–899.

Steriade, M., & Timofeev, I. (2003). Neuronal plasticity in thalamocortical networks during sleep and waking oscillations. Neuron, 37, 563–576.

Supple, W. F., & Kapp, B. S. (1989). Response characteristics of neurons in the medial component of the medial geniculate nucleus during Pavlovian differential fear conditioning in rabbits. Behavioral Neuroscience, 103, 1276–1286.

Sutherland, G. R., & McNaughton, B. (2000). Memory trace reactivation in hippocampal and neocortical neuronal ensembles. Current Opinion in Neurobiology, 10, 180–186.

Van Twyver, H., & Garrett, W. (1972). Arousal threshold in the rat determined by “meaningful” stimuli. Behavioral Biology, 7, 205–215.

Walker, M. P., & Stickgold, R. (2004). Sleep-dependent learning and memory consolidation. Neuron, 44, 121–133.

Weinberger, N. M. (1995). Dynamic regulation of receptive fields and maps in the adult sensory cortex. Annual Review of Neuroscience, 18, 129–158.

Weinberger, N. M. (2003). The nucleus basalis and memory codes: Auditory cortical plasticity and the induction of specific, associative behavioral memory. Neurobiology of Learning and Memory, 80, 268–284.

Weinberger, N. M. (2004). Specific long-term memory traces in primary auditory cortex. Nature Reviews Neuroscience, 5, 279–290.

Weinberger, N. M., Javid, R., & Lepan, B. (1993). Long-term retention of learning-induced receptive-field plasticity in the auditory cortex. Proceedings of the National Academy of Sciences, USA, 90, 2394–2398.

Williams, J. A., Comisarow, J., Day, J., Fibiger, H. C., & Reiner, P. B. (1994). State-dependent release of acetylcholine in rat thalamus measured by in vivo microdialysis. Journal of Neuroscience, 14, 5236–5242.

Wilson, M. A. (2002). Hippocampal memory formation, plasticity, and the role of sleep. Neurobiology of Learning and Memory, 78, 565–569.

Submitted: February 28, 2005 Revised: May 11, 2005 Accepted: May 23, 2005