Experimental Brain Research

, Volume 153, Issue 4, pp 573–578

Bursts in the medial geniculate body: a comparison between anesthetized and unanesthetized states in guinea pig


  • Aurélie Massaux
    • Laboratoire de Neurobiologie de l'Apprentissage de la Mémoire et de la Communication (NAMC), CNRS UMR 8620Université Paris-Sud
    • Laboratoire de Neurobiologie de l'Apprentissage de la Mémoire et de la Communication (NAMC), CNRS UMR 8620Université Paris-Sud
Research Article

DOI: 10.1007/s00221-003-1516-3

Cite this article as:
Massaux, A. & Edeline, J. Exp Brain Res (2003) 153: 573. doi:10.1007/s00221-003-1516-3


Thalamic high frequency bursts have long been described under anesthesia and during slow-wave sleep (SWS). More recently, studies in the lateral geniculate nucleus have pointed out that they are also present during waking (W). Here, we compared the bursts recorded in the medial geniculate body of guinea pigs under anesthesia or during periods of W and SWS. The tuning of single units was tested between threshold and 80 dB SPL in two conditions: (1) in restrained, undrugged, non-sleep-deprived guinea pigs (n=101 cells) and (2) under pentobarbital anesthesia (n=53 cells). Off-line analyses allowed us to distinguish single action potentials (APs) from bursts. A burst was defined as a group of APs with an interspike interval ≤4 ms, preceded by a silent period ≥100 ms. We found that auditory thalamus bursts occur in synchronized electroencephalogram states (SWS and anesthesia), but also during W. Although the burst characteristics did not differ among the three states, group data showed that the proportion of bursts within spike trains was the greatest under anesthesia. This observation resulted from two types of effects: (1) the percentage of non-bursting cells was lowest under anesthesia and (2) some cells under anesthesia exhibited up to 90% of bursts, whereas during W or SWS the highest proportion of bursts did not exceed 40%. The presence of these bursts is discussed with regards to the known fluctuations of membrane potential which occur in these various states.


Auditory thalamusSingle unitSlow-wave sleepWakingAnesthesia


It has long been known that thalamic cells exhibit two distinct modes of discharge (Steriade and Llinas 1988). Since the 1960s, studies performed in several sensory modalities reported that the tonic mode prevails during waking (W) and paradoxical sleep, whereas the burst mode is prevalent during slow-wave sleep (SWS; Bizzy 1966; Sakakura 1968; Mukhametov et al. 1970; Benoit and Chataignier 1973).

The burst mode of thalamic neurons has recently received more attention since a few studies reported the presence of bursts during W (Guido and Weyand 1995; Ramcharan et al. 2000; Fanselow et al. 2001; Swadlow et al. 2001; Weyand et al. 2001) and proposed that bursts act as a "wake-up-call" in the thalamocortical system of unanesthetized animals (Sherman and Guillery 1996; Sherman 2001).

However, the burst mode is also prominent under general anesthesia, a state traditionally used to describe the functional properties of thalamic relay cells. Therefore, before attributing a role to the burst mode, it is necessary to compare the characteristics of bursts collected during anesthesia, W, and SWS. The properties of bursts produced by medial geniculate body (MGB) neurons during each of these three states are described below.

Materials and methods

All procedures were performed in conformity with national (JO887-848) and European (86/609/EEC) legislations on animal experimentation, and have been previously described (Edeline et al. 1999, 2000).

Chronic recordings

For eighteen animals (380–450 g), electrodes were implanted during initial surgery performed under anesthesia (atropine sulfate 0.08 mg/kg, diazepam (Valium) 8 mg/kg, pentobarbital sodium 20 mg/kg; see Evans 1979). All injections were intraperitoneal. To monitor the states of vigilance, electrodes were inserted: (1) between bone and dura (silver-ball electrodes), (2) in the hippocampus (bipolar electrode), and (3) in the dorsal neck muscles (silver wires). An array of five tungsten microelectrodes (1 MΩ, spaced 200–300 μm rostrocaudally) was lowered in the MGB under electrophysiological control. A pedestal in dental acrylic cement including two cylindrical threaded tubes was built to allow atraumatic fixation of the animal's head during the recording sessions. Three days after surgery, animals were adapted to restrained conditions in an acoustically isolated chamber (IAC model AC2). For this, they were laid in a hammock with the head fixed for increasing periods of time (2–6 h/day). They were accustomed to hearing sequences of pure tones, while they spontaneously shifted from W to SWS. No animal was sleep-deprived before the recording sessions. The signals derived from each electrode were checked daily.

Acute recordings

For six animals (400–450 g), the recording session took place under pentobarbital anesthesia (atropine sulfate 0.08 mg/kg, diazepam (Valium) 8 mg/kg, pentobarbital sodium 20 mg/kg; see Evans 1979). Pentobarbital was supplemented regularly (about 4 mg/kg every hour) to maintain an areflexive state. All injections were intraperitoneal. A hole was drilled in the skull above the MGB and single units were recorded with tungsten microelectrodes (5–10 MΩ) or glass micropipettes (10 MΩ).

Experimental procedures

The signal derived from the electrodes was amplified (bandpass: 0.6–10 kHz; gain: 5000) and sent in parallel to an audio monitor and to a voltage window discriminator. As no waveform sorting system was used, only one clear single unit waveform was isolated (signal/noise ratio=5/1) and continuously displayed on the screen of a digital oscilloscope with the corresponding TTL pulses generated by the discriminator. These TTL pulses were sent to the acquisition board of a Pentium II computer; their time of occurrence was known with a 50-μs resolution. Single unit waveforms were digitized (50 kHz sampling rate, see Fig. 1).
Fig. 1A–C.

Examples of single unit waveforms and percentages of bursts during waking (W), slow-wave sleep (SWS), and under pentobarbital anesthesia (Pento.). A Waveforms (A.1 and A.3, 30 sweeps, 50 kHz sampling rate) of two single units recorded under anesthesia within the same penetration, at 6800 and 7820 μm under pia, respectively. Bursts of these two cells are shown in (A.2) and (A.4), respectively. Note that the last spike of the burst can be apart from the preceding ones. B, C Group data: the percentage of bursts obtained under anesthesia was greater than those obtained during W or SWS, both in spontaneous (B) and evoked (C) activity

Pure tone frequencies were generated by a remotely controlled wave analyzer and attenuated by a passive programmable attenuator; both were controlled by a computer via an IEEE bus. An earphone was placed contralaterally close to the ear canal. Stimuli consisted of ascending sequences of eleven isointensity tones (20–80 dB, 100 ms duration, 1 s intertone interval) repeated ten times. Responses to sequences of clicks (1 ms duration, 1 s interclick interval) were also tested in anesthetized conditions.

Data analysis

For undrugged animals, only data obtained in unambiguous states of vigilance (W or SWS) were analyzed. In drugged and undrugged animals, only the responses at the intensity eliciting the strongest evoked response will be considered here.

To quantify the firing rates, the following time windows were used. For each single unit, the spontaneous firing rate was quantified as the discharge rate during the 100 ms preceding each tone, i.e., over a total period of 11 s (100 ms × 11 frequencies × 10 repetitions). The evoked firing rate was defined as the discharge rate averaged across the 11 frequencies used to determine the tuning, i.e., over a total period of 11 s (100 ms duration tone × 11 frequencies × 10 repetitions).

To detect the bursts, the following time windows were used. The period of spontaneous activity was defined as the period beginning 200 ms after tone onset and lasting up to the end of the intertone interval, i.e., over a total period of 88 s (800 ms × 11 frequencies × 10 repetitions). The period of evoked activity differed for cells displaying "on" and "off" responses. For on responses, this period was defined as the 100 ms of tone duration, whereas it was defined as the 100 ms following the end of the tone for off responses. In both cases, these periods extended over 11 s (100 ms × 11 frequencies × 10 repetitions).

During off-line analyses, bursts were isolated from single action potentials (APs) on the basis of criteria established using intracellular recordings (Lu et al. 1992). Groups of APs were defined as being part of a burst only when the interspike interval (ISI) was ≤4 ms and when there was a preceding silent period of at least 100 ms. Such criteria have been applied to extracellular recordings in anesthetized or awake animals (Guido et al. 1992; Guido and Weyand 1995; Ramcharan et al. 2000). Our experimental observations led us to consider, as in a previous study (McCarley et al. 1983), that when there were more than two APs in a burst, the last AP could be separated from the preceding one by up to 10 ms (see Fig. 1). The analyses software used to look for bursts applied, as a first step, the ISIs criteria and, only as a second step, the 100-ms silent period criterion. So, groups of adjacent APs found in a 100-ms evoked activity period could be defined as bursts when the software detected a period of 100 ms silence before the beginning of the evoked activity period (i.e., during spontaneous activity period for on responses, or during tone presentation for off responses).

The percentage of bursts for each single unit was calculated as follows:

[(number of bursts)/(number of bursts + number of single APs)]×100.

Paired t-tests were carried out to compare the variables obtained in W with those in SWS. Unpaired t-tests were used to compare the variables obtained under anesthesia with those obtained during W or SWS. The distributions of the percentages of bursts over the cell population were compared using χ2 tests. The significance level was P≤0.05.


At the end of the experiment, the animals received a lethal dose of pentobarbital (200 mg/kg). In the case of the chronic recordings, the animals were perfused intracardially with 0.9% saline (200 ml) followed by 2000 ml of fixative solution (4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4). In the case of the acute recordings, the brains were removed from the skull and placed in the fixative solution for 2 weeks. In both cases, the brains were subsequently placed in a 30% sucrose solution for 3–4 days, then coronal sections were cut on a freezing microtome (50 μm thickness) and counterstained with cresyl violet. The sections were examined under several microscopic magnifications, and the location of the recording sites was reconstructed using a camera lucida. Examination of the histological material was always done "blind" of the electrophysiological results.


Results are based on recordings derived from 101 single units tested in W and in SWS and from 53 units recorded under pentobarbital anesthesia. The mean spontaneous firing rate was higher during W than during SWS (6.1 vs 4.1 spikes/s, P=0.0008; median 4.3 vs 2.4 spikes/s) and also higher during W than under anesthesia (3.5 spikes/s, P=0.003; median=1.7 spikes/s). The mean spontaneous firing rate did not differ between SWS and anesthetized conditions (P=0.39). The mean evoked firing rate was higher during W than during SWS (17.4 vs 10.2 spikes/s, P<0.0001; median 12.2 vs 6.2 spikes/s), but similar to that under anesthesia (18.2 spikes/s, P=0.76; median=12.5 spikes/s). The mean evoked firing rate was lower during SWS than under anesthesia (P=0.0007).

Proportions and internal structure of bursts

Figure 1B, C refers to group data; the percentages of bursts provided below are actually percentages of bursts averaged across all single units. During the spontaneous activity (Fig. 1B), the percentage of bursts was higher during SWS than during W (6.3% vs 2.7%, P<0.0001; median: 4.0% vs 2.0%). In anesthetized conditions, this percentage (14.8%; median 9%) was higher than those obtained during W and during SWS (P<0.0001 for both comparisons). Similar results were observed for the evoked activity (Fig. 1C): it increased from W (4.5%) to SWS (8.7%) to pentobarbital (16.2%; P<0.001 for all comparisons; median: 2% in W, 8% in SWS, and 10% under pentobarbital). The percentage of bursts was lower in spontaneous than in evoked activity during W (P=0.002) and SWS (P=0.002), but not under anesthesia (P=0.27).

During spontaneous activity, the mean number of APs per burst was lower during W than during SWS (2.2 vs 2.5 APs, P<0.0001), whereas it did not differ between anesthesia and W (2.3 APs, P=0.13) nor between anesthesia and SWS (P=0.11). During evoked activity, no differences were found between W, SWS, and anesthesia (2.4, 2.5, and 2.5 APs, respectively, all P>0.09). Under anesthesia the mean number of APs was lower during spontaneous than during evoked activity (P=0.001).

The burst duration was similar between spontaneous and evoked activity (all P>0.07), as well as between the three states of vigilance (spontaneous activity: W 3.9 ms, SWS 4.2 ms, pentobarbital 4.2 ms, all P>0.09; evoked activity: W 4.4 ms, SWS 4.3 ms, pentobarbital 4.5 ms, all P>0.45).

During spontaneous activity, the bursts' internal frequency did not differ between W and SWS (347 vs 357 Hz, P=0.19), nor between W and anesthesia (329 Hz, P=0.23), whereas it was higher during SWS than under anesthesia (P=0.02). During evoked activity, no differences were found between the three states (W: 350 Hz, SWS: 358 Hz, pentobarbital: 341 Hz, all P>0.25). For each state, there was no difference between spontaneous and evoked activity (all P>0.3).

Across all states, both during spontaneous and evoked activity, the bursts were mainly composed of two or three APs (about 70% and 20% of the bursts, respectively). There were also bursts consisting of at least four to six or more APs, but these bursts represented only low proportions. Both during spontaneous and evoked activity, the proportions of bursts made of two to less than six APs did not differ between anesthesia versus W or SWS (all P>0.06). Under anesthesia, but not during W or SWS, the percentage of bursts made of two APs was greater during spontaneous than during evoked activity (P=0.008).

Distributions of bursts percentages over the cells population

During spontaneous and evoked activity, a certain percentage of cells did not exhibit bursts. In spontaneous activity, there were no differences between states (W: 18% of cells; SWS: 18%; anesthesia: 17%; all χ2<1). In evoked activity, the percentage of non-bursting cells during W was greater than those obtained during SWS or anesthesia (W: 35% of cells; SWS: 23%; anesthesia:13%; P=0.06 for W vs SWS and for SWS vs anesthesia; P=0.0003 for W vs anesthesia).

During spontaneous activity (Fig. 2A), the distributions of burst percentages differed between W and SWS (χ2=19.3, P=0.002), between W and pentobarbital (χ2=50.2, P<0.0001), and also between SWS and pentobarbital (χ2=21.0, P=0.0008). During evoked activity (Fig. 2B), the distributions of burst percentages differed between W and SWS (χ2=16.9, P=0.0007), between W and pentobarbital (χ2=30.1, P<0.0001), and also between SWS and pentobarbital (χ2=16.9, P=0.02).
Fig. 2A, B.

Distributions of bursts percentages over the cells population during W, SWS, and under pentobarbital anesthesia (Pento.). Both in spontaneous activity (A) and in evoked activity (B), no cells had a percentage of bursts above 30% during W, in comparison to 2–3% of the recorded cells during SWS and 19–21% of the recorded cells under anesthesia

Both during W and SWS, the distributions of burst percentages during spontaneous activity differed from those obtained during evoked activity (W: χ2=13.3, P=0.001; SWS: χ2=11.9, P<0.008). This was not the case during anesthesia (χ2=4.2, P=0.76).

Location of the recordings in the divisions of the auditory thalamus

Examination of the histological material obtained from the undrugged animals indicated that the locations of the implanted electrodes were distributed in all MGB divisions. The values of the burst percentages obtained in the various regions of the auditory thalamus are presented in Table 1. In each division, the percentage of bursts was higher during periods of evoked activity than during periods of spontaneous activity, both in W and SWS. In each division, the percentages of bursts obtained during SWS were higher than those obtained during W. In each division, the highest bursts percentage was found during the periods of evoked activity collected in SWS. An ANOVA revealed that there was an effect of the recording site for the burst percentage obtained in W, both during spontaneous (F(3,97)=3.64, P=0.015) and during evoked activity (F(3,97)=2.60, P=0.056); the highest percentage of bursts was found in MGv. Surprisingly, in SWS there was no difference between divisions, both during spontaneous and evoked activity (F(3,97)<1, NS in both cases).
Table 1.

Percentages of bursts (mean ± SD) obtained during periods of spontaneous or evoked activity in the different divisions of the auditory thalamus while the animals were in stable periods of waking or slow-wave sleep (SWS)










Percentage of spontaneous bursts





Percentage of evoked bursts






Percentage of spontaneous bursts





Percentage of evoked bursts





In several cases, the electrode tracks obtained from the animals under pentobarbital anesthesia could not be reconstructed, preventing us from determining precisely the location of the recordings with regard to the MGB divisions. Thus, no statement regarding spatial distribution could be made on this set of data.


On unambiguously isolated extracellular recordings (S/N=5/1), we applied the criteria defined from in vivo intracellular recordings in the visual thalamus (Lu et al. 1992). But do auditory thalamic cells exhibit the same ionic currents as visual relay cells? In vitro intracellular recordings in auditory thalamus demonstrated that neurons displayed bursts riding a low-threshold Ca+ spike (LTS; Hu 1995; Tennigkeit et al. 1996; Bartlett and Smith 1999). The intraburst frequency and, therefore, the ISI are similar in the visual (Lu et al. 1992: 286–429 Hz) and in the auditory (Bartlett and Smith 1999: 328–365 Hz) thalamus. Only the preceding period of silence requested before an LTS differs (≥100 ms in visual thalamus and 200 ms in auditory thalamus), but this might be the consequence of a species and/or methodological differences (Lu et al. 1992: cat in vivo vs Hu 1995: rat in vitro). The criteria used here (ISI ≤4 ms and preceding silent period ≥100 ms) are in fact conservative estimates and minimize the amount of detected bursts. Therefore, some authors used more liberal criteria such as an extended ISI ≥6 ms and a preceding period of quiescence ≥50 ms (Ramcharan et al. 2000; Fanselow et al. 2001; Weyand et al. 2001). Obviously, if in auditory thalamus, an LTS requires about 200 ms to deinactivate (Hu 1995), we might have overestimated the percentages of bursts. But, as there is no evidence for the need of a 200-ms silence period in vivo, applying more traditional criteria seems appropriate, at least as a first approach.

Conflicting results have been previously reported in the literature concerning potential differences of the bursts percentages across the various divisions of the auditory thalamus. Initially, in vitro studies indicated that, in rats, 95% of MGd cells exhibited burst responses, whereas 50–70% of the MGv cells exhibited single- or dual-spike discharges in response to stimulation of the brachium of the inferior colliculus (Hu et al. 1994). However, the same laboratory showed that, in the rat MGv, 118/186 cells (63%) exhibited bursts and only 68/186 (37%) responded with only single- or dual-spike (Mooney et al. 1995 p 840). Subsequent in vitro studies could not find differences in the proportion of cells displaying the It current, a major actor underlying burst generation in thalamic relay cells (Bartlett and Smith 1999). A recent in vivo study pointed out that bursts can be detected more frequently from MGd recordings than from MGv recordings: bursts were present in 77% (33/43) of cells in MGd versus 17% (5/29) of cells in MGv (He and Hu 2002). Besides the use of a different anesthetic, the major factor that could explain the difference between this study and our results is probably the criteria used to define bursts in extracellular in vivo recordings. The criteria adopted by He and Hu (2002) were that: (1) bursting neurons should exhibit responses made of two to six spikes occurring in 15 ms in at least 50% of the trials and that (2) single-spike neurons should show no more than two spikes in a 15-ms period in at least 75% of the trials. In their study, cells exhibiting frequent two-spike responses could not be easily classified. Here, as well as in all previous studies performed in the visual thalamus (Guido et al. 1992; Lu et al. 1992; Guido and Weyand 1995; Ramcharan et al. 2000; Weyand et al. 2001), two-spike responses were treated as burst responses, and therefore, cells displaying two-spike responses were considered as bursting cells.

The present study demonstrates that the various bursts characteristics (number of spikes per burst, burst duration, intraburst frequency) do not differ between W, SWS, and under pentobarbital anesthesia. However, under anesthesia, the mean value of the percentage of spontaneous or evoked bursts is approximately twofold that observed during SWS and fourfold that observed during W. Moreover, comparing the distributions of bursts over the cell population under anesthesia versus W or SWS, reveals that: (1) the percentage of non-bursting cells is lower and (2) there are more cells exhibiting high percentages of bursts.

A pioneer study described that, compared with W, thalamic cells are slightly hyperpolarized during SWS (Hirsch et al. 1983). Actually, more recent intracellular studies performed in undrugged animals (Steriade et al. 2001) demonstrated that the membrane potential of cortical neurons is permanently fluctuating over time, whatever the state of vigilance. During W, the membrane potential undergoes small fluctuations around a mean value, whereas during SWS it transiently, but frequently, reaches hyperpolarized states. If similar effects occur in thalamic cells, they should promote the generation of LTS-related bursts during SWS. The hyperpolarizing effect of anesthetics (Nicoll and Madison 1982) could have two consequences. First, the fluctuations of the membrane potential could take place at more hyperpolarized levels during anesthesia than during SWS. Second, the epochs spent at hyperpolarized states might be of longer duration under anesthesia than during SWS. Thus, in both cases, cells under the influence of anesthetics might have a greater probability to generate LTSs crowned by bursts of conventional APs than cells recorded during periods of SWS. Moreover, our results also reveal the presence of bursts in the waking state. In fact, only 35% of non-bursting cells were found in evoked activity during W (18% in spontaneous activity), indicating that the other 65% of cells (or 82%) were bursting, at least partly. This means that cells recorded during W could have displayed LTSs, as did cells during SWS or under anesthesia. An explanation for this result could also rely on fluctuations of the membrane potential. Although neurons are relatively depolarized during W, they might go through transient hyperpolarized phases where LTSs could be generated. In the future it would be interesting to investigate the origins of such membrane potential fluctuations.


We wish to thank Gérard Dutrieux for outstanding computer programming, Jennifer Horwood for improving the language of the manuscript, and Elizabeth Hennevin for helpful advice throughout this study. A.M. was supported by a fellowship from the Ministère de la Recherche.

Copyright information

© Springer-Verlag 2003