Abstract

In mature cortex, activation of the cholinergic system induces oscillatory network activity and facilitates synaptic plasticity. We used an in vitro preparation of the intact cerebral cortex and cortical slices of the neonatal rat to study carbachol (CCh, ≥30 μM)-induced network oscillations during the early postnatal period. Multi-site extracellular recordings revealed CCh-induced transient beta oscillations with an average duration of 4.6 ± 0.2 s, amplitude of 123 ± 7.4 mV and frequency of 17.7 ± 0.5 Hz. These oscillations propagated uniformly at 0.5–1.5 mm/s over the cortex and were reversibly blocked by tetrodotoxin (TTX) and atropine, indicating that they depended on action potentials and activation of muscarinic receptors. The activity was not blocked by bicuculline methiodide or gabazine, but was reversibly abolished by kynurenic acid, indicating that activation of glutamate receptors, but not GABA-A receptors, was required. CNQX caused a significant decrease in the power of the Fourier frequency spectrum of the CCh-induced oscillations and CPP or MK-801 completely blocked the activity, indicating a contribution of AMPA/kainate receptors and an essential role of NMDA receptors. Oscillations were synchronized between sites separated horizontally by ∼1 mm and for delays of 2–8 ms. Synchronized activity between neighboring recording sites was very stable over repeated applications of CCh. Whole-cell recordings from morphologically identified pyramidal neurons in the intact cortex revealed a close temporal correlation between CCh-induced membrane oscillations and local field potential recordings. In contrast, CCh-induced oscillations recorded in coronal neocortical slices were smaller in amplitude and frequency, suggesting that a widespread network of intracortical axonal connections is required for their generation.

Introduction

Ascending modulatory systems originating in various brain regions strongly influence the functional state of the cerebral cortex (Steriade et al., 1993a). The cholinergic system, arising mainly from the basal forebrain (McKinney et al., 1983), projects densely to the cortex and modulates (Steriade et al., 1991, 1993b) or even induces (Buhl et al., 1998; Fisahn et al., 1998; Beierlein et al., 2000) prominent network oscillations. In the mature hippocampus, cholinergically induced oscillations combined with a few single electrical stimuli induce a long-term potentiation or depression of excitatory synaptic transmission, depending on the timing of the stimulation (Huerta and Lisman, 1993, 1995). Cholinergically induced oscillations may also stabilize or weaken synaptic interactions in the developing cerebral cortex during early postnatal periods. In the mammalian retina, spontaneous waves of action potentials requiring synaptic activation of cholinergic receptors on ganglion cells are present well before the onset of vision (Feller et al., 1996) and it has been postulated that this spatio-temporal activity pattern may refine the neural circuitry during early ontogenesis (Wong, 1993). A cholinergic projection from the basal forebrain to the neocortex (Dinopoulos et al., 1989; Calarco and Robertson, 1995; Mechawar and Descarries, 2001), cholinergic receptors on cortical neurons (Buwalda et al., 1995; Hohmann et al., 1995) and depolarizing membrane responses to application of the cholinergic agonist carbachol (CCh) (Reece and Schwartzkroin, 1991; Avignone and Cherubini, 1999) have been demonstrated in the rat already in the first postnatal week. During this developmental period, CCh also causes a transient rise in the intracellular calcium concentration by activation of muscarinic receptors, which leads to stimulation of a variety of second messenger systems and release of calcium from intracellular stores (Bonner, 1989; Yuste and Katz, 1991). CCh-induced calcium signals propagate as a wave of correlated neural activity over several millimetres across the cerebral cortex of the newborn rat (Peinado, 2000). Recently, Garaschuk et al. (Garaschuk et al., 2000) described spontaneous large-scale oscillatory calcium waves propagating along the longitudinal axis of the newborn rat cortex. These early network oscillations are able to synchronize neuronal domains over large cortical regions and may be important in driving the refinement of the initially crude connectivity in an activity-dependent manner (Goodman and Shatz, 1993; Katz and Shatz, 1996). However, the spatio-temporal properties of such synchronization processes in the developing cerebral cortex are poorly understood. We addressed this issue by performing multisite extracellular recordings and whole-cell recordings from morphologically identified pyramidal neurons in a novel in vitro preparation of the intact cerebral cortex of the neonatal rat. We were especially interested in the following questions.

  1. Which type of oscillatory activity can be induced by CCh application?

  2. Does the CCh-induced activity require action-potential-dependent synaptic transmission and which receptors are involved?

  3. What are the spatio-temporal characteristics of the CCh-induced oscillations?

  4. Do neighboring oscillatory networks show activity-dependent modifications in their coupling strength according to the Hebbian rules?

  5. What is the cellular correlate of the network oscillations?

Material and Methods

Tissue Preparation

All experiments were conducted in accordance with national laws for the use of animals in research and approved by the local ethical committee. Neonatal Wistar rats [postnatal day (P) 0–7; day of birth = P0] were deeply anesthetized by hypothermia and decapitated. The brain was rapidly removed and transferred to oxygenated (95% O2/5% CO2), ice-cold (2–5°C) artificial cerebrospinal fluid (ACSF) of the following composition (in mM): 124 NaCl; 5 KCl; 1.6 CaCl2; 1 MgSO4; 26 NaHCO3; 1.25 NaH2PO4; and 10 glucose (pH 7.4). The subsequent preparation was performed in oxygenated, ice-cold ACSF and lasted 10–15 min. The two hemispheres were separated by a scalpel cut through the midline and all subcortical structures, including the hippocampus, were removed in one hemisphere. The pial membranes were carefully removed and the intact cerebral cortex of one hemisphere was transferred into a conventional fully submerged chamber superfused with ACSF at 32–33°C at a rate of 8–10 ml/min (Khalilov et al., 1997; Luhmann et al., 2000a). The tissue was placed on a plastic mesh with the medial side down and fixed to the silgard bottom using three entomological needles.

Conventional 400–600 mm thick neocortical slices from P5–7 rats were prepared as described previously (Kilb and Luhmann 2000; Luhmann et al., 2000b). In brief, coronal slices including the fronto-parietal cortex were prepared in oxygenated, ice-cold ACSF on a vibroslicer (TPI, St Louis, MO) and transferred to an incubation chamber containing ACSF at 32–33°C. After an incubation period of at least 1 h, slices were transferred to a recording chamber (volume <2 ml) mounted on the fixed stage of an Axioskop microscope (Zeiss, Jena, Germany), where they were continuously superfused at a rate of 2 ml/min with ACSF at 32–33°C.

Multi-site Extracellular and Whole-cell Recordings

Extracellular multi-site recordings were performed with five tungsten 4–5 MΩ microelectrodes (FHC, Bowdoinham, ME, USA) positioned in different cortical areas or in fronto-parietal cortex along a row in rostrocaudal direction with a an electrode tip separation of ∼0.5 mm. Signals were AC recorded with extracellular amplifiers, low-pass filtered at 3 kHz, stored and analyzed with an eight-channel PC-based software program (WinTida, Heka, Lambrecht, Germany). CCh (Sigma-Aldrich, Steinheim, Germany) was added to the bathing solution in a concentration of 1–200 mM. The CCh-induced oscillations were analyzed in the extra-cellular recordings in their onset latency (interval between CCh washing and occurrence of first oscillatory activity), duration (interval between beginning and end of one complete oscillatory cycle) and maximal amplitude (voltage between the positive- and negative-going peaks). Furthermore, fast Fourier transformation (FFT) spectra were calculated by the use of the WinTida program and CCh-induced oscillations were analyzed in their maximal frequency and power.

‘Blind’ whole-cell recordings in the supragranular layers of the intact cortex were performed according to the methods described previously (Blanton et al., 1989). In conventional cortical slices, whole-cell recordings were performed from visually identified pyramidal neurons in upper cortical layers using video enhanced infrared Normarski optics. Recording pipettes were pulled from borosilicate glass tubing (CG200F8P, Science Products, Hofheim, Germany) on a vertical puller (PP83, Narishige, Tokyo, Japan) and filled with the following solution (in mM): 117 K-gluconate; 13 KCl; 1 CaCl2; 2 MgCl2; 11 EGTA; 10 K-HEPES; 2 NaATP; and 0.5 NaGTP. This solution was adjusted to pH 7.4 with 1 M KOH and to an osmolarity of 306 mOsm with sucrose. In all experiments 0.5% biocytin (Sigma-Aldrich) was included in the patch electrode solutions for later cell identification. The patch pipettes with resistances of 6–12 MΩ were connected to the headstage of a discontinuous voltage-clamp/current-clamp amplifier (SEC-05L; NPI, Tamm, Germany) and signals were amplified, low-pass filtered at 3 kHz, digitized online with an AD/DA-board (ITC-16; Heka) and analyzed using WinTida software. For experiments involving whole-cell recordings, extracellular MgSO4 was substituted by MgCl2 and all recordings were corrected for liquid junction potentials with –10 mV (Kilb and Luhmann, 2000).

Intra- and Extracellular Biocytin Staining Histology

Intact cortices containing biocytin-filled neurons were fixed for at least 2 days in 4% paraformaldehyde at 4°C, subsequently cryoprotected overnight with 30% sucrose and coronal sections of 100 mm thickness were cut on a cryotome. Conventional coronal slices were treated in the same way, but not resectioned. For extracellular labelling, a small crystal of biocytin was injected into the upper layers of the primary somatosensory cortex (S1) of the intact cerebral cortex in vitro. After 6–8 h at 32–33°C, the tissue was removed from the submerged chamber, carefully flattened between two layers of tissue paper, fixed overnight, cryoprotected and tangential section of 75–100 μm thickness were cut. Sections were stained as described previously (Schröder and Luhmann, 1997). Sections were preincubated for 60 min in phosphate-buffered saline (PBS) containing 0.5% H2O2 to saturate endogenous peroxidase, followed by incubation overnight in avidin-coupled peroxidase (ABC kit; Vectorlabs, Burlingame, CA). After wash in PBS and Tris the slices were incubated for 30 min in 3,3-diaminobenzidine (DAB; Sigma-Aldrich) and for 10 min in DAB containing 0.01% H2O2. The reaction product was intensified by 2–3 min incubation in 0.5% OsO4. Finally, the slices were rinsed in TRIS and distilled water, dehydrated in ethanol and embedded in Durcopan (Fluka, Buchs, Switzerland). Digital photographs of biocytin labelled sections were taken with a Nikon Coolpix 990 camera (Nikon, Düsseldorf, Germany) attached to a Zeiss microscope.

Drugs

The sodium channel blocker tetrodotoxin citrate (TTX; RBI, Natick, MA) was applied in a concentration of 0.1–1 mM and 10 mM atropine sulfate (RBI) was used to block muscarinic acetylcholine receptors. Gamma-amino butyric acid type A (GABA-A) receptors were blocked with 20 mM bicuculline methiodide (BMI; Sigma-Aldrich) or 100 mM gabazine (SR-95531; RBI). The broad spectrum glutamate receptor antagonist kynurenic acid (KA, Sigma-Aldrich) was dissolved freshly at 500 mM in ACSF. The α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic (AMPA)/ kainate antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; RBI) was bath applied at 10 mM and N-methyl-d-aspartate (NMDA) receptors were blocked with (±)-3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (CPP, 20 μM; RBI) or the non-competitive antagonist (+)-MK-801 hydrogen maleate (MK-801, 20 mM; RBI).

Statistics

Statistical analyses were performed with Systat version 9 (SPSS Inc., Chicago, IL) and Auto Signal version 1.5 (SPSS Inc.). Cross-correlograms for 1 ms binwidth were calculated and the strength of the correlation between the neural activity recorded at two different sites was quantified by measuring the peak cross-correlation coefficient (Eggermont, 2000). Values throughout this report are given as mean ± SEM. For statistical comparisons, a paired samples t-test was performed.

Results

Carbachol-induced Network Oscillations in the Intact Cerebral Cortex In Vitro

The intact neocortex in vitro preparation proofed to be most valuable to study large-scale network activities in cortical structures [see also Schwartz et al. (Schwartz et al., 1998) for details on spontaneous neuronal activity in the whole hemisphere preparation]. Intact cortices were analyzed with extra- and intracellular recording techniques over periods of 6–8 h and did not reveal any signs of functional abnormalities. Transient oscillatory network activity occurring 34 ± 2.9 s after washing of CCh could be observed with extracellular recording electrodes in all 70 intact cortices prepared from P0–P7 rats. In normal bathing solution these transient oscillations had a duration of 4.6 ± 0.2 s (n = 68) and a maximal amplitude of 123 ± 7.4 μV (n = 128; Fig. 1A). Color-coded spectrograms (Fig. 1B) and fast Fourier transformations (FFT; Fig. 1C) of the signals indicate that the CCh-induced activity consists of oscillations in the beta (12–30 Hz) and to a lesser extent in the lower gamma (30–80 Hz) frequency range. Whereas the initial oscillation lasting 1–2 s is dominated by higher frequencies in the range of 20–40 Hz, the subsequent activity is characterized by a gradual decrease in oscillatory frequency to 10–15 Hz (Fig. 1B). The FFT spectra revealed the maximal power at an average frequency of 17.7 ± 0.5 Hz (n = 128). In contrast to the persistent CCh-induced network oscillations in adult rat hippocampal slices (Fisahn et al., 1998), oscillatory activity in the intact cerebral cortex of the neonatal rat was primarily restricted to the washing phase of CCh (≥30 μM) and repetitive periods of network oscillations could be observed only rarely at intervals of 30–40 min.

Previous studies in cortical slice preparations (Fisahn et al., 1998; Peinado, 2000) have demonstrated network oscillations at CCh concentrations ranging between 20 and 50 μM. In order to determine the threshold concentration of CCh to elicit the transient oscillatory activity in the intact cerebral cortex, we applied CCh at bath concentrations ranging from 1 to 200 μM (Fig. 2). Network oscillations were absent at CCh concentrations of ≤3 mM (n = 9). The percentage of cortices showing a response to CCh increased from 33% at 10 μM to ≥89% at ≥30 mM (open bars in Fig. 2B). At the threshold concentration of 30 μM, neither the response amplitude (156 ± 19 mV, filled bars in Fig. 2B) nor the frequency (16.7 ± 0.7 Hz, n = 7) of the CCh-induced activity differed from the responses obtained at CCh concentrations of 100–200 μM (Fig. 2A).

The CCh-induced oscillatory activity propagated at 0.5–1.5 mm/s over the cortex, similar to the spontaneous oscillatory calcium waves described previously (Garaschuk et al., 2000) in horizontal brain slices of the newborn rat. The progressive recruitment of oscillatory local circuits during a CCh-induced activity wave is shown in Figure 3. Each FFT spectrogram calculated from the five extracellular recordings, revealed a maximum in the beta frequency range and a minor peak in the gamma range (Fig. 3C). All recordings also showed the typical gradual decrease in the frequency during the oscillatory activity (Fig. 3D).

Pharmacology of CCh-induced Network Oscillations

We first addressed the question of whether the CCh-induced oscillations were mediated by action potential-dependent synaptic interactions. TTX blocked the oscillatory activity in six out of six experiments and this effect was reversible after washing out TTX for >45 min (Fig. 4A). The network activity was also blocked by atropine (n = 6), indicating that muscarinic receptors play an essential role in CCh-induced cortical oscillations (Fig. 4B). The participation of GABA-A receptors in the cholinergically induced activity was studied by bath application of the specific antagonist BMI or gabazine (Table 1). BMI caused a reversible increase in the response amplitude and maximal FFT power of the CCh-induced oscillations (Fig. 4C), but these effects were not significant at the P < 0.05 level. Gabazine had similar, but non-reversible effects (Fig. 4D). Both GABA-A antagonists did neither change the peak frequency nor the propagation pattern of the network activity. These results suggest that GABA-A receptor blockade amplifies, rather than inhibits, the CCh-induced oscillations.

Since the broad spectrum glutamate receptor antagonist kynurenic acid completely and reversibly blocked the network activity, we studied the effects of more specific glutamate receptor antagonists (Table 1, Fig. 5). CNQX did not significantly change the duration, maximal amplitude or frequency of the CCh-induced oscillations, but on average caused an irreversible and significant decrease in the FFT power by ≈27%. The spread of oscillatory activity was unmodified by CNQX. These data indicate that AMPA/kainate receptors contribute to the network oscillations, but are not critically involved in their initiation or propagation. In contrast to CNQX, the NMDA receptor antagonist CPP reversibly blocked all CCh-induced responses. Identical results were obtained with the non-competitive NMDA receptor antagonist MK-801 (n = 16), but the MK–801 effects were irreversible (data not shown). These results demonstrate that the cholinergically induced network oscillations in the intact cerebral cortex of the neonatal rat depend on NMDA receptor activation.

Spatio-temporal Synchronization of CCh-induced Oscillatory Activity

The spatial and temporal coupling of the oscillations was studied in nine intact cortices by performing cross-correlation analyses of the responses recorded with five extracellular electrodes, which were separated by 0.5–2 mm (Fig. 6). CCh-induced beta rhythms support robust synchronization between sites separated horizontally by ∼1 mm and for delays of 2–8 ms (Fig. 6A,B). At a distance of 0.5 mm, the average cross-correlation coefficient estimated 0.558 ± 0.056 (n = 20 pairs). At 1.0 mm, the cross-correlation coefficient decreased to 0.366 ± 0.069 (n = 15 pairs) and at distances of ≥1.5 mm no significant correlation coefficients could be detected (at 1.5 mm, 0.162 ± 0.031, n = 10 and at 2.0 mm 0.075 ± 0.031, n = 5; Fig. 6C,D). These data indicate that local circuits synchronize in the range of a few milliseconds over distances of up to 1 mm already in the very immature cerebral cortex.

In another set of experiments (n = 4) we addressed the question, whether activity-dependent processes may modify the coupling strength between weakly (cross-correlation coefficient of 0.1–0.4) and strongly (>0.7) connected networks. For that purpose CCh was bath applied for 2–4 min at intervals of 20–30 min and the responses recorded simultaneously with up to eight extracellular electrodes were investigated by cross-correlation analyses. In contrast to our hypothesis, that the cross-correlation coefficient between weakly coupled networks would decrease and, vice versa, between strongly connected networks will increase, the coupling strength was remarkably stable for at least 3–4 h (Fig. 7). These data may indicate that additional factors, e.g. patterned synaptic sensory inputs, are required to modify the spatio-temporal properties of the CCh-induced and NMDA-receptor-dependent oscillations in neonatal cerebral cortex.

Cellular Correlate of CCh-induced Network Oscillations

To examine in more detail the cellular responses to CCh application, we performed whole-cell current-clamp recordings from visually identified and biocytin-labelled pyramidal neurons (Fig. 8A) and simultaneous extracellular recordings in 14 coronal slices of P5–7 rat fronto-parietal cortex. The average resting membrane potential was –63.6 ± 1.5 mV and the input resistance 582 ± 52.6 MΩ (n = 17). In all of these cells, bath application of CCh increased the spontaneous synaptic activity and in 13 neurons a prominent membrane depolarization by 24.3 ± 3.8 mV could be observed (Fig. 8B), which in 70% of the cells was sufficient to trigger action potentials. In the remaining four cells, CCh induced a small membrane hyperpolarization by 5.6 ± 1.4 mV. In contrast to our observations in the intact cerebral cortex, CCh-induced network oscillations recorded with extracellular electrodes in slices occurred less frequently and were smaller in amplitude (<120 μV; Fig. 8C). Furthermore, CCh-induced oscillations in the field potential recordings (12.4 ± 1.8 Hz, n = 5) and in single cells (10.8 ± 3.7 Hz, n = 4) were smaller in their maximal frequency (Fig. 8D) in slices as compared to the intact cortex. However, a smaller peak in the lower beta frequency range could be also observed (arrow in Fig. 8D). These data indicate that a sufficiently large network of intra-cortical excitatory connections is essential for the generation of robust CCh-induced beta oscillations and that this network is only partially preserved in the cortical slice preparation.

Therefore, we also performed whole-cell current-clamp recordings from biocytin-stained pyramidal neurons (Fig. 9A) and simultaneous extracellular recordings in the primary somatosensory cortex of the intact cortex preparation. The average resting membrane potential (–64.1 ± 2.8 mV) and the input resistance (567 ± 59.8 MΩ, n = 13) were very similar as compared to the values obtained from pyramidal neurons recorded in conventional cortex slices (see above), indicating that neurons in the in toto cortex show comparable passive membrane properties. Bath application of CCh elicited in supra-granular pyramidal neurons a sudden membrane depolarization lasting 6–10 s associated with a barrage of action potentials (Fig. 9B,C). The average discharge frequency at the beginning of the response was 14.3 ± 0.8 Hz (n = 13), close to the maximal discharge frequencies elicited by injection of strong depolarizing current pulses under control conditions (15.6 ± 0.8 Hz, n = 13). The intracellular response was transiently synchronized with the local field potential over periods of 0.5–2 s (vertical arrows in Fig. 9D; see also Fig. 10C). However, due to action potential failures not every cycle in the field potential recording correlated with a spike (asterisks in Fig. 9D), as previously also reported for gamma oscillations in the hippocampus of the behaving rat (Bragin et al., 1995) and for CCh-induced oscillations in hippocampal slices (Fisahn et al., 1998). In order to avoid action potential generation, neurons were hyperpolarized to membrane potentials negative to –80 mV by constant negative current injection before and during CCh application. Under this condition, CCh elicited membrane potential oscillations in the beta frequency range (16.2 ± 1 Hz; n = 11), which were transiently synchronized with the local field potential (Fig. 10A,B). Figure 10C illustrates the changes in coupling strength between the intracellularly recorded membrane potential of a pyramidal neuron and the local field potential during CCh-induced oscillatory activity of about 10 s duration. Cross-correlation coefficients show a remarkable variation from one second to the next, because membrane potential oscillations and field potentials were transiently running out of phase. These data demonstrate that single neurons participate to the local network response in a dynamic manner.

The effect of NMDA receptor blockade on CCh-induced responses was studied in biocytin-labelled pyramidal neurons recorded in coronal slices and in intact cortices (Fig. 11). Addition of 20 mM CPP, reduced the CCh-induced depolarization in pyramidal neurons recorded in neocortical slices from 25 ± 5 mV to 6.3 ± 1.3 mV (n = 7). In pyramidal neurons analyzed in the intact cortex preparation, the response was reduced by CPP from 10.5 ± 2.4 mV to 1.9 ± 1.3 mV (n = 3). This result indicates that NMDA receptor blockade in neonatal cortex reduces excitatory synaptic interactions and prevents the synchronization of action potential-dependent coupling in local networks.

Anatomical Substrate of Intracortical Synchronization in Neonatal Cortex

Extracellular biocytin injections into the upper layers of S1 in the intact cerebral cortex of 10 P4–7 rats revealed a dense network of anterogradely labelled horizontal fibers and retrogradely stained neurons organized in discrete clusters up to 2.7 mm from the center of the injection site (Fig. 12A). The majority of these cells resembled in their morphology pyramidal neurons (Fig. 12B,C), although an unequivocal identification of the cells in tangential slices is difficult. These data indicate that intracortical horizontal connections of considerable extent are capable to synchronize the CCh-induced oscillations over spatial dimensions of at least 1 mm.

Discussion

In this study we used a novel intact in vitro preparation and conventional neocortical slices to study the properties of cholinergically induced oscillations in the cerebral cortex of the neonatal rat. The main results of these experiments are as follows. (i) Bath application of the cholinergic agonist CCh induces transient network oscillations in the beta and to a lesser extent also in the gamma frequency range, spreading with 0.5–1.5 mm/s over the whole cortical hemisphere. (ii) CCh-induced oscillations are blocked by TTX, atropine and CPP or MK-801, but not by CNQX, bicuculline methiodide or gabazine, suggesting that they require action-potential-dependent synaptic interactions and are sustained by an excitatory circuit involving muscarinic and NMDA receptors. (iii) CCh-induced oscillations are synchronized between sites separated by ∼1 mm and for delays of 2–8 ms, indicating cooperative activity between neighboring cortical modules already at earliest stages of development. The coupling strength between these oscillatory modules is stable over periods of at least 3–4 h. (iv) CCh-induced network oscillations are transiently synchronized over periods of 0.5–2 s with membrane potential oscillations recorded in pyramidal neurons. (v) Long-range horizontal axonal connections form a suitable anatomical framework to synchronize CCh-induced activity between neighboring modules.

Anatomical Substrate of Spatio-temporal Correlation

Previous studies in developing cat visual cortex have already demonstrated horizontal intrinsic connections ranging over several millimeters (Luhmann et al., 1986, 1990; Assal and Innocenti, 1993; Galuske and Singer, 1996) and our previous studies in neonatal rat cortical slices have shown lateral axonal projections up to 2 mm [see Fig. 4A in Luhmann et al. (Luhmann et al., 1999)]. Since horizontal intrinsic connections in rat cerebral cortex conduct at 0.15–0.55 m/s (Murakoshi et al., 1993), monosynaptic interactions over 1 mm should reveal a latency of 1.8–6.7 ms. This value is exactly in the range of our results in the cross-correlograms (2–8 ms). Therefore, lateral intracortical connections probably form the anatomical substrate of synchronized CCh-induced rhythmic activity in newborn rat cortex. Recently Chiu and Weliky (Chiu and Weliky, 2001) demonstrated in developing ferret visual cortex in vivo clusters of correlated spontaneous activity that were separated by ∼1 mm and proposed that this pattern is generated by horizontal patchy connections.

But why does the neonatal cortical network oscillate in the beta frequency range? Kopell et al. (Kopell et al., 2000) suggested on the basis of their computational network model that synchronization over long distances and long conduction delays is mainly supported by beta rhythms. This hypothesis is supported by experimental data demonstrating synchronized activity in the beta frequency range between widely spaced cortical areas of awake cats during a visuomotor task (Roelfsema et al., 1997). In the newborn rat cortex, synchronization processes are restricted to ∼1 mm, but due to the slow kinetics of passive and active membrane properties at this age (Luhmann et al., 2000b), long conduction delays occur even over these relatively small distances. The beta rhythm seems to be most suited to mediate synchronization of CCh-induced oscillatory activity in the neonatal cerebral cortex. During further development and maturation of faster membrane kinetics and synaptic properties, the system may use faster rhythms (i.e. in the gamma range) to synchronize intracortical activity at nearby sites (Buhl et al., 1998; Dickson et al., 2000). Steriade et al. (Steriade et al., 1996) demonstrated in adult cat cortex, that during the transition from sleep to wakefulness fast oscillations (30–40 Hz) are not only synchronized between neighboring (1–2 mm) sites, but also between different cortical areas separated by >5 mm.

Cellular Basis of CCh-induced Network Oscillations

Beierlein et al. (Beierlein et al., 2000) have recently demonstrated in P14–21 rats that a network of low-threshold-spiking inhibitory interneurons, when activated by muscarinic or metabotropic glutamate receptor agonists, can coordinate the firing pattern of neocortical neurons over a distance of ∼400 μm. This widespread synchronous inhibition is critically depending on electrical synapses containing the gap junction protein Cx36 (Deans et al., 2001). Gap junctions in combination with GABAergic synaptic connections are also capable of synchronizing action potential generation in regular spiking nonpyramidal neurons at beta and gamma frequencies (Szabadics et al., 2001). These data indicate that cortical networks of inhibitory interneurons connected by electrical (and chemical) synapses are well suited to synchronize neuronal activity and to generate different cortical rhythms (Galarreta and Hestrin, 2001). In contrast, the CCh-induced network oscillations observed in the newborn rat cerebral cortex are mediated by chemical synapses and are not dependent on intact GABAergic synaptic transmission. In agreement with previous observations in neonatal rat hippocampus (Reece and Schwartzkroin, 1991; Avignone and Cherubini, 1999), application of CCh to the intact cerebral cortex of the newborn rat produced an increase in synaptic network activity and most often a prominent membrane depolarization. The ascending cholinergic projection from the basal forebrain (Dinopoulos et al., 1989; Calarco and Robertson, 1995; Mechawar and Descarries, 2001) and muscarinic receptors on cortical neurons (Buwalda et al., 1995; Hohmann et al., 1995) are clearly present in the perinatal rat and an important function of the cholinergic system and of muscarinic receptors in cortical morphogenesis has been previously demonstrated in various studies (Hohmann and Berger-Sweeney, 1998). Our data imply that a network with sufficient recurrent excitatory connections can generate oscillatory activity by itself as long as the neuronal activity exceeds a critical threshold, which is surpassed by the combined activation of muscarinic and NMDA receptors (Fig. 13). Activation of muscarinic or NMDA receptors induces in newborn rat cortical neurons an intracellular calcium rise (Yuste and Katz, 1991) and Peinado (Peinado, 2000) recently demonstrated that cholinergically induced calcium waves propagate in newborn cortex over many millimeters. In contrast to the spontaneous neuronal domains in developing neocortex, which are insensitive to TTX, propagate locally at ∼100 μm/s and are abolished by gap junction blockers (Yuste et al., 1992, 1995), the CCh-induced propagating calcium waves are blocked by TTX (Peinado, 2000). In the newborn rat cortex, large-scale oscillatory calcium waves propagating at 2.1 mm/s require action potentials and activation of AMPA and NMDA (Garaschuk et al., 2000). Although AMPA receptor blockade caused a decrease in the average amplitude and FFT power of the CCh-induced oscillatory activity (Table 1), our observations in the intact cerebral cortex suggest that NMDA receptors play a primary role in cholinergically induced oscillations. Application of NMDA antagonists only reduced the CCh-induced membrane depolarization, but completely blocked the oscillatory network activity. These data indicate that NMDA receptors are required to reach the critical threshold for eliciting the network oscillations and that the CCh-induced increase in synaptic activity is insufficient to trigger the oscillations.

Functional Relevance of Cholinergically Induced Synchronized Oscillations

The cholinergic system plays an important role in the maturation and in developmental plasticity of the cerebral cortex (Hohmann and Berger-Sweeney, 1998). Numerous studies in rodent somatosensory cortex (Hohmann et al., 1988) and cat visual cortex (Bear and Singer, 1986) have demonstrated that an impairment in cholinergic function results in a deficit of cortical morphogenesis or early synaptic plasticity. Our data indicate that cholinergically induced oscillatory activity may influence these processes during the early stages of corticogenesis. The synchronization of CCh-induced oscillatory activity between neighboring neocortical modules is sustained by an excitatory circuit involving AMPA/kainate receptors and critically depending on NMDA receptors (Fig. 13). The crucial role of NMDA receptors during early neocortical development, i.e. in pattern formation, has been well documented in the cat visual (Kleinschmidt et al., 1987) and mouse somatosensory cortex (Iwasato et al., 1997, 2000). During the first postnatal week, a large number of glutamatergic synapses in rat cortical structures are exclusively mediated by NMDA receptors (Isaac et al., 1997; Rumpel et al., 1998) and activity-dependent modifications in synaptic strength require NMDA receptor activation (Crair and Malenka, 1995; Durand et al., 1996). Furthermore, during early postnatal development NMDA currents in rat cerebral cortex show an activity-dependent decrease in duration (Carmignoto and Vicini, 1992) by upregulation of the NR2A subunit (Flint et al., 1997), further supporting the hypothesis that NMDA receptors mediate activity-dependent plasticity during neonatal stages (Crair, 1999). Since in the rat hippocampal CA1 region, NMDA receptors are also critically involved in synaptic plasticity during CCh-induced theta oscillation (Huerta and Lisman, 1993, 1995), we expected activity-dependent modifications in the synchronization process between neighboring networks over repeated CCh applications. Whereas in the olfactory system of insects, odor-evoked synchronized oscillations increase in spike time precision and coherence over repeated odor stimulations (Stopfer and Laurent, 1999), CCh-induced beta oscillations in the neonatal rat cortex are characterized by their stability of synchronization. Probably additional synaptic activity arising from the sensory periphery is required for modifying these synaptic connections according to Hebbian (Hebb, 1949) rules (Singer, 1995; Katz and Shatz, 1996; Sur et al., 1999).

We suggest that in the neonatal cerebral cortex in vivo the three following prerequisites have to be fulfilled in order to induce physiologically relevant synaptic modifications between spatially separated cortical modules (Fig. 13). (i) The neuronal network has to be connected over considerable distances via horizontal axonal fibers, which mainly use NMDA receptor synapses. (ii) Ascending modulatory systems, as the cholinergic system, have to convert the neocortical network into an oscillatory mode with a preference in the beta frequency range, which tolerates long conduction delays. (iii) Patterned synaptic inputs provide the decisive signal to modify horizontal connections between functionally related sites in an experience-dependent manner and according to Hebbian rules. However, whether the early formation of cortical circuits and patterns involving lateral interactions critically depends on these three conditions can be tested only under in vivo conditions or in intact in vitro preparations.

Table 1

Pharmacology of CCh-induced oscillations in neonatal rat cerebral cortex

 n Amplitude (mV) Frequency (Hz) FFT power (nV2
Effect of GABA-A or ionotropic glutamate receptor blockade on maximal amplitude, frequency and FFT power of carbachol-induced network oscillations recorded with extracellular electrodes in the in vitro intact cerebral cortex of the newborn rat. Data are expressed as mean ± SEM and statistically significant differences versus control are indicated at *P < 0.05, **P < 0.01 and ***P < 0.001 (paired samples t-test). 
GABA-A receptor antagonists     
    Control 15 167.0 ± 29.6 17.3 ± 1.2 3.88 ± 0.47 
    20 mM BMI 15 187.6 ± 38.4 16.0 ± 1.1 5.85 ± 1.35 
    Recovery 15 148.9 ± 15.9 16.2 ± 1.2 3.99 ± 0.58 
    Control 19 132.9 ± 14 15.4 ± 0.8 3.76 ± 0.49 
    100 mM gabazine 19 209.4 ± 45.8* 16.4 ± 1.2 4.87 ± 1.05 
    Recovery 19 204.4 ± 35.5* 15.0 ± 4.6 5.55 ± 1.04* 
Glutamate receptor antagonists     
    Control 10 142.4 ± 19.8 21.7 ± 2 6.11 ± 0.77 
    500 mM KA 10  0 ± 0*** n.a. n.a. 
    Recovery 10 141.5 ± 18.9 20.3 ± 1.6 6.09 ± 0.99 
    Control 41 136.5 ± 10 17.4 ± 0.9 5.65 ± 0.66 
    10 mM CNQX 41 123.4 ± 9.9 16.1 ± 0.6 4.16 ± 0.48** 
    Recovery 41 129.2 ± 9.4 16.8 ± 0.7 4.08 ± 0.51** 
    Control 10 162.0 ± 20.2 19.2 ± 1.2 8.39 ± 1.23 
    20 mM CPP 10  0 ± 0*** n.a. n.a. 
    Recovery 10 171.1 ± 19.6 18.6 ± 2.1 8.64 ± 1.11 
 n Amplitude (mV) Frequency (Hz) FFT power (nV2
Effect of GABA-A or ionotropic glutamate receptor blockade on maximal amplitude, frequency and FFT power of carbachol-induced network oscillations recorded with extracellular electrodes in the in vitro intact cerebral cortex of the newborn rat. Data are expressed as mean ± SEM and statistically significant differences versus control are indicated at *P < 0.05, **P < 0.01 and ***P < 0.001 (paired samples t-test). 
GABA-A receptor antagonists     
    Control 15 167.0 ± 29.6 17.3 ± 1.2 3.88 ± 0.47 
    20 mM BMI 15 187.6 ± 38.4 16.0 ± 1.1 5.85 ± 1.35 
    Recovery 15 148.9 ± 15.9 16.2 ± 1.2 3.99 ± 0.58 
    Control 19 132.9 ± 14 15.4 ± 0.8 3.76 ± 0.49 
    100 mM gabazine 19 209.4 ± 45.8* 16.4 ± 1.2 4.87 ± 1.05 
    Recovery 19 204.4 ± 35.5* 15.0 ± 4.6 5.55 ± 1.04* 
Glutamate receptor antagonists     
    Control 10 142.4 ± 19.8 21.7 ± 2 6.11 ± 0.77 
    500 mM KA 10  0 ± 0*** n.a. n.a. 
    Recovery 10 141.5 ± 18.9 20.3 ± 1.6 6.09 ± 0.99 
    Control 41 136.5 ± 10 17.4 ± 0.9 5.65 ± 0.66 
    10 mM CNQX 41 123.4 ± 9.9 16.1 ± 0.6 4.16 ± 0.48** 
    Recovery 41 129.2 ± 9.4 16.8 ± 0.7 4.08 ± 0.51** 
    Control 10 162.0 ± 20.2 19.2 ± 1.2 8.39 ± 1.23 
    20 mM CPP 10  0 ± 0*** n.a. n.a. 
    Recovery 10 171.1 ± 19.6 18.6 ± 2.1 8.64 ± 1.11 
Figure 1.

Bath application of carbachol to the in vitro intact cerebral cortex of a P3 rat elicits transient network oscillations of more than 150 μV amplitude and 3 s duration. (A) Extracellular AC recording of the CCh-induced beta and gamma activity in parietal cortex. (B) Color-coded spectrogram of the recording shown in (A) at identical time scale. Note progressive decrease in the frequency from ∼20–40 Hz at the onset to ∼10 Hz during the final oscillation. (C) Fast Fourier transformation of the recording in (A). Note maximal power at 13 and 18.5 Hz as also illustrated in (B).

Figure 1.

Bath application of carbachol to the in vitro intact cerebral cortex of a P3 rat elicits transient network oscillations of more than 150 μV amplitude and 3 s duration. (A) Extracellular AC recording of the CCh-induced beta and gamma activity in parietal cortex. (B) Color-coded spectrogram of the recording shown in (A) at identical time scale. Note progressive decrease in the frequency from ∼20–40 Hz at the onset to ∼10 Hz during the final oscillation. (C) Fast Fourier transformation of the recording in (A). Note maximal power at 13 and 18.5 Hz as also illustrated in (B).

Figure 2.

Threshold concentration of CCh to elicit all-or-none network oscillations. (A) extracellular recordings in the intact cortex of a P5 rat during washing of 10, 30 and 100 μM CCh. (B) Percentage of cortices with response to CCh (open bars) and response amplitude (filled bars) at CCh concentrations ranging from 1 to 200 μM (n = 9 preparations). Bars represent mean ± SEM.

Figure 2.

Threshold concentration of CCh to elicit all-or-none network oscillations. (A) extracellular recordings in the intact cortex of a P5 rat during washing of 10, 30 and 100 μM CCh. (B) Percentage of cortices with response to CCh (open bars) and response amplitude (filled bars) at CCh concentrations ranging from 1 to 200 μM (n = 9 preparations). Bars represent mean ± SEM.

Figure 3.

Widespread propagation of CCh-induced oscillations in the cerebral cortex of a P5 rat. (A) Schematic illustration of the electrode positions in a lateral view of the intact cortex with locations of the primary motor (M1), somatosensory (S1), auditory (Au1), visual (V1) and secondary somatosensory (S2) cortex according to Zilles and Wree (Zilles and Wree, 1985). (B) Simultaneous extracellular recordings in parietal cortex with five tungsten electrodes positioned in rostro-caudal direction with a tip separation of ∼0.5 mm as illustrated in (A). (C) Fast Fourier spectra of the CCh-induced oscillations shown in (B). (D) Color-coded spectrogram of the recordings shown in (B) at identical time scale. Note propagation of oscillatory activity from the most rostral electrode (1) towards the caudal pole at a velocity of ≈0.7 mm/s.

Figure 3.

Widespread propagation of CCh-induced oscillations in the cerebral cortex of a P5 rat. (A) Schematic illustration of the electrode positions in a lateral view of the intact cortex with locations of the primary motor (M1), somatosensory (S1), auditory (Au1), visual (V1) and secondary somatosensory (S2) cortex according to Zilles and Wree (Zilles and Wree, 1985). (B) Simultaneous extracellular recordings in parietal cortex with five tungsten electrodes positioned in rostro-caudal direction with a tip separation of ∼0.5 mm as illustrated in (A). (C) Fast Fourier spectra of the CCh-induced oscillations shown in (B). (D) Color-coded spectrogram of the recordings shown in (B) at identical time scale. Note propagation of oscillatory activity from the most rostral electrode (1) towards the caudal pole at a velocity of ≈0.7 mm/s.

Figure 4.

CCh-induced network oscillations in the intact cortex are reversibly blocked by 0.1 μM TTX (A) and 10 μM atropine (B). Bath application of 20 μM bicuculline methiodide (C) or 100 μM gabazine (D) causes an increase in the amplitude of the cholinergically induced oscillations.

Figure 4.

CCh-induced network oscillations in the intact cortex are reversibly blocked by 0.1 μM TTX (A) and 10 μM atropine (B). Bath application of 20 μM bicuculline methiodide (C) or 100 μM gabazine (D) causes an increase in the amplitude of the cholinergically induced oscillations.

Figure 5.

Effects of glutamate receptor antagonists on CCh-induced network oscillations. Blockade of glutamate receptors with 500 μM kynurenic acid reversibly abolishes beta and gamma activity. Whereas the AMPA/kainate receptor antagonist CNQX (10 μM) does not block CCh-induced oscillations, bath application of the NMDA antagonist CPP causes reversible blockade of network activity. All recordings were obtained from the same intact neocortex of a P5 rat. Fourier spectra were calculated from the extracellular recordings illustrated to the left.

Figure 5.

Effects of glutamate receptor antagonists on CCh-induced network oscillations. Blockade of glutamate receptors with 500 μM kynurenic acid reversibly abolishes beta and gamma activity. Whereas the AMPA/kainate receptor antagonist CNQX (10 μM) does not block CCh-induced oscillations, bath application of the NMDA antagonist CPP causes reversible blockade of network activity. All recordings were obtained from the same intact neocortex of a P5 rat. Fourier spectra were calculated from the extracellular recordings illustrated to the left.

Figure 6.

CCh-induced oscillations in newborn rat cortex are synchronized over distances of ∼1 mm. Cross-correlograms of the recordings shown in Figure 3. Downward arrows in columns A and B mark peak of correlation on the time axis. (E) Schematic illustration of the five recording sites (1–5) in parietal cortex of a P5 rat.

CCh-induced oscillations in newborn rat cortex are synchronized over distances of ∼1 mm. Cross-correlograms of the recordings shown in Figure 3. Downward arrows in columns A and B mark peak of correlation on the time axis. (E) Schematic illustration of the five recording sites (1–5) in parietal cortex of a P5 rat.

Figure 7.

CCh-induced interactions between spatially separated networks are stable for hours. CCh was bath applied for 2–3 min to a P2 rat intact cortex at intervals of 20–30 min and the cross-correlation coefficients between weakly (<0.4) and strongly (>0.7) coupled networks were calculated for each oscillatory activity recorded with extracellular electrodes. After each oscillation, CCh was washed out until the next application.

Figure 7.

CCh-induced interactions between spatially separated networks are stable for hours. CCh was bath applied for 2–3 min to a P2 rat intact cortex at intervals of 20–30 min and the cross-correlation coefficients between weakly (<0.4) and strongly (>0.7) coupled networks were calculated for each oscillatory activity recorded with extracellular electrodes. After each oscillation, CCh was washed out until the next application.

Figure 8.

CCh application to the neonatal rat parietal cortical slice induces in a pyramidal cell an increase in spontaneous synaptic activity and a membrane depolarization. (A) Morphology of a biocytin-stained pyramidal cell in P7 rat cerebral cortex. Dotted line marks cortical surface. Note presence of dye-coupled cells and horizontal axon collateral (downward arrows). (B) CCh causes a gradual depolarization and a marked increase in depolarizing potentials. Resting membrane potential was –59 mV. (C) Simultaneous whole-cell recording in current-clamp mode (upper trace) and extracellular field potential (lower trace) recorded close to the cell. (D) Fast Fourier transformation of the extracellular recording in (C) (lower trace). Note maximal power in alpha and beta frequency range.

Figure 8.

CCh application to the neonatal rat parietal cortical slice induces in a pyramidal cell an increase in spontaneous synaptic activity and a membrane depolarization. (A) Morphology of a biocytin-stained pyramidal cell in P7 rat cerebral cortex. Dotted line marks cortical surface. Note presence of dye-coupled cells and horizontal axon collateral (downward arrows). (B) CCh causes a gradual depolarization and a marked increase in depolarizing potentials. Resting membrane potential was –59 mV. (C) Simultaneous whole-cell recording in current-clamp mode (upper trace) and extracellular field potential (lower trace) recorded close to the cell. (D) Fast Fourier transformation of the extracellular recording in (C) (lower trace). Note maximal power in alpha and beta frequency range.

Figure 9.

Whole-cell current-clamp recording with a biocytin-containing patch electrode (upper trace in BD) and simultaneous field potential recordings with two extracellular electrodes (FP1 and FP2) in the primary somatosensory cortex of the intact cortex preparation. The patch-clamp electrode was positioned between the two extracellular electrodes separated in rostro-caudal direction by ∼1 mm. (A) Photograph of biocytin-labelled pyramidal neuron in 100 μm coronal section of the intact cortex. (B) CCh elicits a prominent membrane depolarization associated with repetitive action potentials and in the extracellular recordings typical network oscillations. (C, D) Whole-cell and extracellular recordings shown in (B) at higher time resolution. Note good temporal correlation between membrane potential deflections and extracellular oscillation in FP1 (vertical arrows in D). Asterisks in (D) mark action potential failures.

Figure 9.

Whole-cell current-clamp recording with a biocytin-containing patch electrode (upper trace in BD) and simultaneous field potential recordings with two extracellular electrodes (FP1 and FP2) in the primary somatosensory cortex of the intact cortex preparation. The patch-clamp electrode was positioned between the two extracellular electrodes separated in rostro-caudal direction by ∼1 mm. (A) Photograph of biocytin-labelled pyramidal neuron in 100 μm coronal section of the intact cortex. (B) CCh elicits a prominent membrane depolarization associated with repetitive action potentials and in the extracellular recordings typical network oscillations. (C, D) Whole-cell and extracellular recordings shown in (B) at higher time resolution. Note good temporal correlation between membrane potential deflections and extracellular oscillation in FP1 (vertical arrows in D). Asterisks in (D) mark action potential failures.

Figure 10.

Temporal relationship between CCh-induced membrane potential oscillation of a supragranular pyramidal neuron (upper trace in A) and local field potential oscillation recorded at a distance of ∼0.3 mm (lower trace). Corresponding cross-correlogram is shown in (B). (C) Changes in cross-correlation coefficient between intracellular membrane potential and local field potential during a CCh-induced response (same experiment as in A).

Figure 10.

Temporal relationship between CCh-induced membrane potential oscillation of a supragranular pyramidal neuron (upper trace in A) and local field potential oscillation recorded at a distance of ∼0.3 mm (lower trace). Corresponding cross-correlogram is shown in (B). (C) Changes in cross-correlation coefficient between intracellular membrane potential and local field potential during a CCh-induced response (same experiment as in A).

Figure 11.

Reduction of the CCh-induced depolarization by NMDA receptor blockade. Whole-cell current-clamp recording of a layer II/III pyramidal cell in a P7 rat cortical slice at –60 mV under control conditions (A) and in a bathing solution containing 20 μM CPP (B).

Figure 11.

Reduction of the CCh-induced depolarization by NMDA receptor blockade. Whole-cell current-clamp recording of a layer II/III pyramidal cell in a P7 rat cortical slice at –60 mV under control conditions (A) and in a bathing solution containing 20 μM CPP (B).

Figure 12.

Organization of horizontal intrinsic connections in a 75 μm tangential section of a P6 rat intact cortex. (A) An extracellular injection of a small biocytin crystal into S1 (open star) reveals clusters (filled stars) of anterogradely labelled fibers and retrogradely stained neurons up to 2.7 mm from center of the injection site. (B, C) Magnification of the cells marked in (A).

Figure 12.

Organization of horizontal intrinsic connections in a 75 μm tangential section of a P6 rat intact cortex. (A) An extracellular injection of a small biocytin crystal into S1 (open star) reveals clusters (filled stars) of anterogradely labelled fibers and retrogradely stained neurons up to 2.7 mm from center of the injection site. (B, C) Magnification of the cells marked in (A).

Figure 13.

Schematic model illustrating the generation of CCh-induced oscillations in neonatal cerebral cortex. The cholinergic input (green) arising from the basal forebrain excites neocortical pyramidal neurons via activation of muscarinic receptors (light green) and hyperpolarizes a minor proportion of neurons (blue) via activation of an inhibitory cholinergic receptor (dark green). The cholinergically activated neurons generate a local oscillatory network activity via glutamatergic synapses using predominantly NMDA and to a lesser extent AMPA/kainate receptors (yellow). Neighboring cortical networks, separated by up to 1 mm, become synchronized via activation of intrinsic excitatory connections. This wave of oscillatory activity propagates with 0.5–1.5 mm/s over the cortex.

Figure 13.

Schematic model illustrating the generation of CCh-induced oscillations in neonatal cerebral cortex. The cholinergic input (green) arising from the basal forebrain excites neocortical pyramidal neurons via activation of muscarinic receptors (light green) and hyperpolarizes a minor proportion of neurons (blue) via activation of an inhibitory cholinergic receptor (dark green). The cholinergically activated neurons generate a local oscillatory network activity via glutamatergic synapses using predominantly NMDA and to a lesser extent AMPA/kainate receptors (yellow). Neighboring cortical networks, separated by up to 1 mm, become synchronized via activation of intrinsic excitatory connections. This wave of oscillatory activity propagates with 0.5–1.5 mm/s over the cortex.

References

Assal F, Innocenti GM (
1993
) Transient intra-areal axons in developing cat visual cortex.
Cereb Cortex
 
3
:
290
–303.
Avignone E, Cherubini E (
1999
) Muscarinic receptor modulation of GABA-mediated giant depolarizing potentials in the neonatal rat hippocampus.
J Physiol (Lond)
 
518
:
97
–107.
Bear MF, Singer W (
1986
) Modulation of visual cortical plasticity by acetylcholine and norepinephrine.
Nature
 
320
:
172
–176.
Beierlein M, Gibson JR, Connors BW (
2000
) A network of electrically coupled interneurons drives synchronized inhibition in neocortex.
Nat Neurosci
 
3
:
904
–910.
Blanton MG, LoTurco JJ, Kriegstein AR (
1989
) Whole cell recording from neurons in slices of reptilian and mammalian cerebral cortex.
J Neurosci Methods
 
30
:
203
–210.
Bonner TI (
1989
) The molecular basis of muscarinic receptor diversity.
Trends Neurosci
 
12
:
148
–151.
Bragin A, Jandó G, Nádasdy Z, Hetke J, Wise K, Buzsáki G (
1995
) Gamma (40–100 Hz) oscillation in the hippocampus of the behaving rat.
J Neurosci
 
15
:
47
–60.
Buhl EH, Tamás G, Fisahn A (
1998
) Cholinergic activation and tonic excitation induce persistent gamma oscillations in mouse somatosensory cortex in vitro.
J Physiol (Lond)
 
513
:
117
–126.
Buwalda B, de Groote L, Van der Zee EA, Matsuyama T, Luiten PG (
1995
) Immunocytochemical demonstration of developmental distribution of muscarinic acetylcholine receptors in rat parietal cortex.
Dev Brain Res
 
84
:
185
–191.
Calarco CA, Robertson RT (
1995
) Development of basal forebrain projections to visual cortex: DiI studies in rat.
J Comp Neurol
 
354
:
608
–626.
Carmignoto G, Vicini S (
1992
) Activity-dependent decrease in NMDA receptor responses during development of the visual cortex.
Science
 
258
:
1007
–1011.
Chiu CY, Weliky M (
2001
) Spontaneous activity in developing ferret visual cortex in vivo.
J Neurosci
 
21
:
8906
–8914.
Crair MC (
1999
) Neuronal activity during development: permissive or instructive?
Curr Opin Neurobiol
 
9
:
88
–93.
Crair MC, Malenka RC (
1995
) A critical period for long-term potentiation at thalamocortical synapses.
Nature
 
375
:
325
–328.
Deans MR, Gibson JR, Sellitto C, Connors BW, Paul DL (
2001
) Synchronous activity of inhibitory networks in neocortex requires electrical synapses containing Connexin36.
Neuron
 
31
:
477
–485.
Dickson CT, Biella G, De Curtis M (
2000
) Evidence for spatial modules mediated by temporal synchronization of carbachol-induced gamma rhythm in medial entorhinal cortex.
J Neurosci
 
20
:
7846
–7854.
Dinopoulos A, Eadie LA, Dori I, Parnavelas JG (
1989
) The development of basal forebrain projections to the rat visual cortex.
Exp Brain Res
 
76
:
563
–571.
Durand GM, Kovalchuk Y, Konnerth A (
1996
) Long-term potentiation and functional synapse induction in developing hippocampus.
Nature
 
381
:
71
–75.
Eggermont JJ (
2000
) Sound-induced synchronization of neural activity between and within three auditory cortical areas.
J Neurophysiol
 
83
:
2708
–2722.
Feller MB, Wellis DP, Stellwagen D, Werblin FS, Shatz CJ (
1996
) Requirement for cholinergic synaptic transmission in the propagation of spontaneous retinal waves.
Science
 
272
:
1182
–1187.
Fisahn A, Pike FG, Buhl EH, Paulsen O (
1998
) Cholinergic induction of network oscillations at 40 Hz in the hippocampus in vitro.
Nature
 
394
:
186
–189.
Flint AC, Maisch US, Weishaupt JH, Kriegstein AR, Monyer H (
1997
) NR2A subunit expression shortens NMDA receptor synaptic currents in developing neocortex.
J Neurosci
 
17
:
2469
–2476.
Galarreta M, Hestrin S (
2001
) Electrical synapses between GABA-releasing interneurons.
Nat Rev Neurosci
 
2
:
425
–433.
Galuske RAW, Singer W (
1996
) The origin and topography of long-range intrinsic projections in cat visual cortex: a developmental study.
Cereb Cortex
 
6
:
417
–430.
Garaschuk O, Linn J, Eilers J, Konnerth A (
2000
) Large-scale oscillatory calcium waves in the immature cortex.
Nat Neurosci
 
3
:
452
–459.
Goodman CS, Shatz CJ (
1993
) Developmental mechanisms that generate precise patterns of neuronal connectivity.
Neuron
 
10
(Suppl.):
77
–98.
Hebb DO (1949) The organization of behaviour. New York: John Wiley.
Hohmann CF, Berger-Sweeney J (
1998
) Cholinergic regulation of cortical development and plasticity. New twists to an old story.
Perspect Dev Neurobiol
 
5
:
401
–425.
Hohmann CF, Brooks AR, Coyle JT (
1988
) Neonatal lesions of the basal forebrain cholinergic neurons result in abnormal cortical development.
Dev Brain Res
 
42
:
253
–264.
Hohmann CF, Potter ED, Levey AI (
1995
) Development of muscarinic receptor subtypes in the forebrain of the mouse.
J Comp Neurol
 
358
:
88
–101.
Huerta PT, Lisman JE (
1993
) Heightened synaptic plasticity of hippocampal CA1 neurons during a cholinergically induced rhythmic state.
Nature
 
364
:
723
–725.
Huerta PT, Lisman JE (
1995
) Bidirectional synaptic plasticity induced by a single burst during cholinergic theta oscillation in CA1 in vitro.
Neuron
 
15
:
1053
–1063.
Isaac JTR, Crair MC, Nicoll RA, Malenka RC (
1997
) Silent synapses during development of thalamocortical inputs.
Neuron
 
18
:
269
–280.
Iwasato T, Erzurumlu RS, Huerta PT, Chen DF, Sasaoka T, Ulupinar E, Tonegawa S (
1997
) NMDA receptor-dependent refinement of somatotopic maps.
Neuron
 
19
:
1201
–1210.
Iwasato T, Datwani A, Wolf AM, Nishiyama H, Taguchi Y, Tonegawa S, Knöpfel T, Erzurumlu RS, Itohara S (
2000
) Cortex-restricted disruption of NMDAR1 impairs neuronal patterns in the barrel cortex.
Nature
 
406
:
726
–731.
Katz LC, Shatz CJ (
1996
) Synaptic activity and the construction of cortical circuits.
Science
 
274
:
1133
–1138.
Khalilov I, Esclapez M, Medina I, Aggoun D, Lamsa K, Leinekugle X, Khazipov R, Ben-Ari Y (
1997
) A novel in vitro preparation: the intact hippocampal formation.
Neuron
 
19
:
743
–749.
Kilb W, Luhmann HJ (
2000
) Characterization of a hyperpolarization activated inward current in Cajal–Retzius cells in rat neonatal neocortex.
J Neurophysiol
 
84
:
1681
–1691.
Kleinschmidt A, Bear MF, Singer W (
1987
) Blockade of ‘NMDA’ receptors disrupts experience-dependent plasticity of kitten striate cortex.
Science
 
238
:
355
–358.
Kopell N, Ermentrout GB, Whittington MA, Traub RD (
2000
) Gamma rhythms and beta rhythms have different synchronization properties.
Proc Natl Acad Sci USA
 
97
:
1867
–1872.
Luhmann HJ, Martínez-Millán L, Singer W (
1986
) Development of horizontal intrinsic connections in cat striate cortex.
Exp Brain Res
 
63
:
443
–448.
Luhmann HJ, Singer W, Martínez-Millán L (
1990
) Horizontal interactions in cat striate cortex: I. Anatomical substrate and postnatal development.
Eur J Neurosci
 
2
:
344
–367.
Luhmann HJ, Schubert D, Kötter R, Staiger JF (
1999
) Cellular morphology and physiology of the perinatal rat cerebral cortex.
Dev Neurosci
 
21
:
298
–309.
Luhmann HJ, Dzhala VI, Ben-Ari Y (
2000
) Generation and propagation of 4-AP-induced epileptiform activity in neonatal intact limbic structures in vitro.
Eur J Neurosci
 
12
:
2757
–2768.
Luhmann HJ, Reiprich RA, Hanganu I, Kilb W (
2000
) Cellular physiology of the neonatal rat cerebral cortex: intrinsic membrane properties, sodium and calcium currents.
J Neurosci Res
 
62
:
574
–584.
McKinney M, Coyle JT, Hedreen JC (
1983
) Topographic analysis of the innervation of the rat neocortex and hippocampus by the basal forebrain cholinergic system.
J Comp Neurol
 
217
:
103
–121.
Mechawar N, Descarries L (
2001
) The cholinergic innervation develops early and rapidly in the rat cerebral cortex: a quantitative immuno-cytochemical study.
Neuroscience
 
108
:
555
–567.
Murakoshi T, Guo J-Z, Ichinose T (
1993
) Electrophysiological identification of horizontal synaptic connections in rat visual cortex in vitro.
Neurosci Lett
 
163
:
211
–214.
Peinado A (
2000
) Traveling slow waves of neural activity: a novel form of network activity in developing neocortex.
J Neurosci
 
20
:
NIL1
–NIL6.
Reece LJ, Schwartzkroin PA (
1991
) Effects of cholinergic agonists on immature rat hippocampal neurons.
Dev Brain Res
 
60
:
29
–42.
Roelfsema PR, Engel AK, König P, Singer W (
1997
) Visuomotor integration is associated with zero time-lag synchronization among cortical areas.
Nature
 
385
:
157
–161.
Rumpel S, Hatt H, Gottmann K (
1998
) Silent synapses in the developing rat visual cortex: evidence for postsynaptic expression of synaptic plasticity.
J Neurosci
 
18
:
8863
–8874.
Schröder R, Luhmann HJ (
1997
) Morphology, electrophysiology and pathophysiology of supragranular neurons in rat primary somato-sensory cortex.
Eur J Neurosci
 
9
:
163
–176.
Schwartz TH, Rabinowitz D, Unni V, Kumar VS, Smetters DK, Tsiola A, Yuste R (
1998
) Networks of coactive neurons in developing layer 1.
Neuron
 
20
:
541
–552.
Singer W (
1995
) Development and plasticity of cortical processing architectures.
Science
 
270
:
758
–764.
Steriade M, Dossi RC, Paré D, Oakson G (
1991
) Fast oscillations (20–40 Hz) in thalamocortical systems and their potentiation by mesopontine cholinergic nuclei in the cat.
Proc Natl Acad Sci USA
 
88
:
4396
–4400.
Steriade M, McCormick DA, Sejnowski TJ (
1993
) Thalamocortical oscillations in the sleeping and aroused brain.
Science
 
262
:
679
–685.
Steriade M, Amzica F, Nuñez A (
1993
) Cholinergic and noradrenergic modulation of the slow (∼0.3 Hz) oscillation in neocortical cells.
J Neurophysiol
 
70
:
1385
–1400.
Steriade M, Amzica F, Contreras D (
1996
) Synchronization of fast (30–40 Hz) spontaneous cortical rhythms during brain activation.
J Neurosci
 
16
:
392
–417.
Stopfer M, Laurent G (
1999
) Short-term memory in olfactory network dynamics.
Nature
 
402
:
664
–668.
Sur M, Angelucci A, Sharma J (
1999
) Rewiring cortex: the role of patterned activity in development and plasticity of neocortical circuits.
J Neurobiol
 
41
:
33
–43.
Szabadics J, Lorincz A, Tamás G (
2001
) β and γ frequency synchronization by dendritic GABAergic synapses and gap junctions in a network of cortical interneurons.
J Neurosci
 
21
:
5824
–5831.
Wong ROL (
1993
) The role of spatio-temporal firing patterns in neuronal development of sensory systems.
Curr Opin Neurobiol
 
3
:
595
–601.
Yuste R, Katz LC (
1991
) Control of postsynaptic Ca2+ influx in developing neocortex by excitatory and inhibitory neurotransmitters.
Neuron
 
6
:
333
–344.
Yuste R, Peinado A, Katz LC (
1992
) Neuronal domains in developing neocortex.
Science
 
257
:
665
–669.
Yuste R, Nelson DA, Rubin WW, Katz LC (
1995
) Neuronal domains in developing neocortex: mechanisms of coactivation.
Neuron
 
14
:
7
–17.
Zilles K, Wree A (1985) Cortex: areal and laminar structure. In: The rat nervous system (Paxinos G, ed.), vol. 1, pp. 375–415. Sydney: Academic Press.