Letters to Nature

Nature 433, 868-873 (24 February 2005) | doi:10.1038/nature03252; Received 9 September 2004; Accepted 6 December 2004

Excitatory cortical neurons form fine-scale functional networks

Yumiko Yoshimura1,2, Jami L. M. Dantzker1,2 and Edward M. Callaway1

  1. Systems Neurobiology Laboratories, The Salk Institute for Biological Studies, 10010 North Torrey Pines Road, La Jolla, California 92037, USA
  2. Present addresses: Department of Visual Neuroscience, Research Institute of Environmental Medicine, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan (Y.Y.); Department of Neurology and Neurological Sciences, Stanford University, 300 Pasteur Drive, Room M016, Stanford, California 94305-5122, USA (J.L.M.D.)

Correspondence to: Edward M. Callaway1 Correspondence and requests for materials should be addressed to E.M.C. (Email: callaway@salk.edu).

Top

The specificity of cortical neuron connections creates columns of functionally similar neurons spanning from the pia to the white matter1, 2, 3, 4, 5, 6. Here we investigate whether there is an additional, finer level of specificity that creates subnetworks of excitatory neurons within functional columns. We tested for fine-scale specificity of connections to cortical layer 2/3 pyramidal neurons in rat visual cortex by using cross-correlation analyses of synaptic currents evoked by photostimulation. Recording simultaneously from adjacent layer 2/3 pyramidal cells, we find that when they are connected to each other (20% of all recorded pairs) they share common input from layer 4 and within layer 2/3. When adjacent layer 2/3 neurons are not connected to each other, they share very little (if any) common excitatory input from layers 4 and 2/3. In contrast, all layer 2/3 neurons share common excitatory input from layer 5 and inhibitory input from layers 2/3 and 4, regardless of whether they are connected to each other. Thus, excitatory connections from layer 4 to layer 2/3 and within layer 2/3 form fine-scale assemblies of selectively interconnected neurons; inhibitory connections and excitatory connections from layer 5 link neurons across these fine-scale subnetworks. Relatively independent subnetworks of excitatory neurons are therefore embedded within the larger-scale functional architecture; this allows neighbouring neurons to convey information more independently than suggested by previous descriptions of cortical circuitry.

The cerebral cortex consists of a complex network of neuronal connections, the organization of which is believed to contribute critically to perception and behaviour. Over the last several decades, the idea of the 'functional column' has provided a dominant influence on studies of the organization and function of cortical circuits1. The specificities of connections that both create and maintain functional architecture are well established1, 2, 3, 4, 5, 6, 7 and provide a substrate for interactions between neurons with similar functional attributes. However, each excitatory neuron connects to only a minority of others in the same column8, 9. This sparse connectivity is consistent with two different scenarios, each of which has implications for the way that cortical circuits process information. In the first scenario, connections would be dependent on the spatial overlap of dendrites and axons, but would otherwise be determined probabilistically, independent of other connections in the network10, 11, 12, 13. Neighbouring neurons with extensively overlapping dendrites might share common input by chance, but information would be averaged across neurons to create a reliable but relatively uniform output. In the alternative scenario, there might be a fine-scale organization of connections between excitatory neurons within functional columns. The probability that two neurons are connected might be dependent on whether they share common input from other sources. This would reflect rules of connectivity that are not random but instead create substructure within each functional column. Fine-scale selectivity embedded within the columnar functional architecture might give rise to relatively independent neuronal networks that process information uniquely from their immediate neighbours.

To address these issues we took advantage of the ability of focal uncaging of glutamate ('photostimulation') to generate action potentials asynchronously in a small, spatially restricted population of neurons in rat visual cortex brain slices (see Supplementary Figs S1 and S2). By combining this type of stimulation with intracellular recordings of excitatory and inhibitory synaptic currents in pairs of adjacent layer 2/3 pyramidal neurons, we were able to use the timing of evoked synaptic currents to infer whether individual stimulated neurons provided common input to both recorded cells or if instead the recorded cells received input from separate neuronal populations. If a single presynaptic neuron is stimulated and it provides input to both recorded cells, it will generate synchronous synaptic currents; inputs from different presynaptic neurons that fire action potentials asynchronously will generate asynchronous synaptic currents.

We used established cross-correlation analysis methods14 to normalize for synchrony resulting from time-locking of action potential generation to the stimulus (see Methods). To obtain correlation probabilities, the numbers of synchronous synaptic currents attributable to shared input were expressed as a proportion of the total numbers of evoked synaptic currents from each cell (see Methods for details). The correlation probability closely estimates the probability that when a photostimulated presynaptic neuron fires an action potential and evokes a synaptic current in one of the two recorded layer 2/3 pyramidal neurons, the same presynaptic neuron will also evoke a synaptic current in the second recorded neuron. To determine the extent of shared input from the different laminar sources to each pair of recorded layer 2/3 pyramidal cells, separate calculations of correlation probability were made based on stimulation sites in each cortical layer. For example, for the pair of layer 2/3 pyramidal neurons illustrated in Fig. 1a, the correlation probability for stimulation sites in layer 4 was 0.22, meaning that for 22% of the cases in which a layer 4 neuron was stimulated and evoked an excitatory postsynaptic current (EPSC) in one layer 2/3 cell, that same layer 4 neuron also evoked a synchronous EPSC that was detected in the other recorded layer 2/3 cell. Similarly, the correlation probability of 0.30 obtained for this same pyramidal neuron pair for stimulation sites in layer 2/3 shows that these cells share about 30% of their excitatory layer 2/3 inputs.

Figure 1: Cross-correlation analyses of photostimulation-evoked excitatory postsynaptic currents (EPSCs) simultaneously recorded in adjacent pairs of layer 2/3 pyramidal neurons.
Figure 1 : Cross-correlation analyses of photostimulation-evoked excitatory postsynaptic currents (EPSCs) simultaneously recorded in adjacent pairs of layer 2/3 pyramidal neurons. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact npg@nature.com

a, b, Results are shown for a pair of pyramidal neurons that was synaptically connected (a) and for a pair that was not connected (b). For each of the two cells, reconstructions of the locations of photostimulation sites (coloured squares) relative to the locations of laminar borders and cell bodies of recorded pyramidal neurons (triangles) are shown. The colour of each square indicates the sum of amplitudes of EPSCs that were observed in response to photostimulation at that site. To the right of these plots, example voltage clamp recordings are shown for stimulation sites indicated by the large numbered squares. Simultaneous recordings from representative sites in each cortical layer are shown for 'cell a' (red) and 'cell b' (black). The short horizontal lines above each trace indicate the onset of photostimulation. The histograms to the far right of each panel show matched (black) and shifted (red) correlograms computed from data collected upon stimulation in each layer. The corresponding correlation probabilities (CPs) computed from these analyses are also indicated.

High resolution image and legend (170K)

To test whether connected pyramidal neurons belong to functional subnetworks, we compared correlation probabilities between connected and unconnected layer 2/3 pyramidal neuron pairs. We found that the correlation probabilities for excitatory connections from layer 4 and from layer 2/3 to layer 2/3 pyramidal neuron pairs depended on whether the recorded pyramidal neurons were connected to each other. When the recorded layer 2/3 pyramids were not connected, the correlation probabilities for EPSCs were very low (Figs 1b and 2a). For unconnected cell pairs, correlation probabilities averaged 3.8 plusminus 1.1% (mean plusminus s.e.m.) for layer 2/3 stimulation sites (range -5.3 to 10.3, 17 cell pairs) and 3.6 plusminus 0.9% for layer 4 stimulation sites (range -2.2 to 9.1, 17 cell pairs). In sharp contrast to the low correlation probability values for cell pairs that were not connected, correlation probabilities were high when layer 2/3 pyramids were connected (Figs 1a and 2a). For connected cell pairs, correlation probabilities averaged 20.1 plusminus 2.7% for layer 2/3 stimulation sites (range 4.5 to 37.6, 16 cell pairs) and 16.8 plusminus 2.1% for layer 4 stimulation sites (range 2.9 to 30.0, 16 cell pairs). The differences in correlation probabilities for connected versus unconnected cell pairs were highly significant (P < 0.0001 for both layer 2/3 and layer 4 stimulation; Mann-Whitney U-test).

Figure 2: Correlation probabilities for EPSCs and IPSCs in layer 2/3 pyramidal cell pairs.
Figure 2 : Correlation probabilities for EPSCs and IPSCs in layer 2/3 pyramidal cell pairs. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact npg@nature.com

a, b, Correlation probabilities for EPSCs (a) and IPSCs (b) are shown separately for photostimulation sites in each cortical layer and for connected (circles) versus unconnected (squares) pairs of cells. Open circles correspond to pyramidal cells that were reciprocally connected and filled circles indicate one-way connections. Mean values for each group are indicated by horizontal lines. c, d, An experiment testing connections between pyramidal cell pairs (c) and summary of results (d). Action potentials were triggered by injecting positive current into one current-clamped cell (upper traces) while recording EPSCs in the other voltage-clamped cell (lower traces, 3 examples shown for each). The cell pair illustrated in c had a one-way connection from 'cell b' to 'cell a'. Connections, when present, were highly reliable for both layer 4 to layer 2/3 (4% failure) and layer 2/3 to layer 2/3 (8% failure) pairs. Connections were found in 20% of cell pairs for both layer 4 to layer 2/3 and layer 2/3 to layer 2/3 pairs.

High resolution image and legend (100K)

These data indicate that excitatory connections from layer 4 are highly selective at a fine scale, connecting preferentially to layer 2/3 pyramidal neurons that are in turn connected to each other. Furthermore, the fine-scale groupings of neurons that are defined by the selectivity of input from layer 4 are further reinforced by the fine-scale selectivity of excitatory connections within layer 2/3.

In contrast to excitatory connections from layer 4 and within layer 2/3, the specificity of excitatory connections from layer 5 was not dependent on whether the simultaneously recorded layer 2/3 pyramids were connected to each other. For stimulation sites in layer 5, correlation probabilities based on EPSCs averaged 9.8 plusminus 1.7% for unconnected cell pairs and 10.7 plusminus 1.7% for connected cell pairs (no significant difference, P > 0.78; Figs 1 and 2a). The finding that correlation probabilities do not differ between connected and unconnected cell pairs indicates that these connections link neurons across the fine-scale subnetworks established by the specificities of layer 4 and layer 2/3 connections.

The organization of inhibitory connections to adjacent layer 2/3 pyramidal cells was also independent from whether the layer 2/3 cells were connected. Therefore, inhibitory connections also do not respect the fine-scale groupings defined by the specificities of layer 4 and layer 2/3 excitatory connections. To estimate the extent of common inhibitory input onto adjacent pyramidal neurons from interneurons in each layer, the same photostimulation methods were employed except that the layer 2/3 pyramidal cells were voltage-clamped at 0 mV (see Methods) to allow recordings of inhibitory postsynaptic currents (IPSCs) as outward currents (Fig. 3). The correlation probabilities for IPSCs were typically high, regardless of whether the layer 2/3 pyramidal cells were connected and regardless of the layer that was stimulated (Figs 2b and 3). For connected cell pairs, correlation probabilities averaged 23.8 plusminus 2.7% (range 11.5 to 40.0) for stimulation sites in layer 2/3 and 18.0 plusminus 6.8% (range 9.8 to 28.4) for layer 4. For unconnected cell pairs the correlation probabilities averaged 23.8 plusminus 1.7% (range 16.2 to 33.1) for layer 2/3, and 19.1 plusminus 2.5% (range 6.3 to 31.6) for layer 4. There were no significant differences between connected cell pairs or between layers. Layer 5 is not included in this analysis because, as expected from previous studies15, photostimulation in layer 5 resulted in only modest or no significant increase in IPSCs above spontaneous IPSC levels (see Methods).


It is important to note that cortical inhibition comes from diverse cell types that make synapses to different parts of the dendritic arbours of pyramidal cells and for which connections might differ in their reliability and/or detectability16. It is therefore possible that the IPSCs we detect during photostimulation might preferentially represent inputs from certain inhibitory cell types. For example, IPSCs detected during photostimulation could be biased towards those from basket cells (which make strong synapses close to or at the cell body)16, relative to those from inhibitory cell types that make synapses at electrotonically distant sites16. Therefore, our results might not generalize to all sources of inhibition of layer 2/3 pyramidal neurons.

In summary, the results presented here indicate that excitatory connections from layer 4 to layer 2/3 pyramidal cells and within layer 2/3 are highly specific on a fine scale, creating groups of selectively interconnected neurons (Fig. 4). Such groupings are predicted from hebbian learning rules17, which are likely to regulate the development and maintenance of excitatory cortical connections18, 19, 20.

Figure 4: Schematic diagram illustrating the organization of cortical connections proposed in this study.
Figure 4 : Schematic diagram illustrating the organization of cortical connections proposed in this study. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact npg@nature.com

Excitatory connections from layer 4 to layer 2/3 and within layer 2/3 define groups of selectively interconnected neurons (red or blue). The organization of excitation from layer 5 (grey triangles) and inhibition from layers 2/3 and 4 interneurons (IN, grey ovals) does not respect the fine-scale interconnected cell groups defined by excitatory connections from layer 4 and within layer 2/3.

High resolution image and legend (59K)

Neither inhibitory connections nor excitatory connections from layer 5 provide shared input preferentially to connected layer 2/3 pyramidal cells (Fig. 4). Instead, these connections can serve to link neurons across the fine-scale subnetworks defined by excitatory connections from layers 2–4, and might act to modulate activity between the subnetworks. These observations do not, however, rule out the possibility that these connections might be specific for modes of organization not examined in these studies. For example there could be specificity related to an organization orthogonal to the fine-scale excitatory networks in layers 2–4, or on a different spatial scale.

Previous studies, using either photostimulation or more conventional methods, have demonstrated selectivity of connections to cortical neurons of distinctly different types15, 21, 22, 23, 24, 25, 26, 27, 28. Here we have extended the photostimulation method to explicitly test for the possibility of fine-scale selectivity of cortical connections to neurons of the same type. We observe that the specificity of excitatory connections from layer 4 and within layer 2/3 depends on whether adjacent layer 2/3 pyramidal neurons are interconnected, independent of their locations (Fig. 4). These results therefore demonstrate selectivity of connections at a finer scale than cortical columnar architecture.

Like other species, the rat visual cortex has a columnar, retinotopic organization. But orientation-selective neurons in rat visual cortex are not organized into orientation columns; instead, adjacent neurons can be selective for disparate orientations29. We therefore suggest that the fine-scale specificity of excitatory connections in rat visual cortex may be related to the emergence of orientation selectivity, similar to the columnar specificity found in cats and tree shrews3, 4, 5, 6. However, even in cat visual cortex where orientation columns are present, the excitatory connections from layer 4 to layer 3 are sparse; paired recordings show that layer 4 excitatory neurons connect to layer 3 pyramidal cells only 10% of the time9. And despite the presence of orientation columns, neighbouring neurons in cat visual cortex can differ in their selectivities for other features and sometimes demonstrate functional 'micro-organization'30. We therefore suggest that fine-scale selectivity of excitatory connections, embedded within the coarser columnar organization, is a common feature of cortical circuits.

Top

Methods

Slice preparation, photostimulation and recordings

The methods used for these experiments are similar to those reported previously15, with some differences as detailed here. A vibratome was used to cut 300 microm thick coronal brain slices from the primary visual cortex of P21–26 Long-Evans rats. Slices were cut in artificial cerebral spinal fluid (ACSF; 124 mM NaCl, 5 mM KCl, 1.25 mM KH2PO4, 1.3 mM MgSO4, 3.2 mM CaCl2, 26 mM NaHCO3 and 10 mM glucose) and stored in an interface chamber at approx34 °C for at least one hour until they were transferred to a recording chamber containing ACSF with 60–80 microM 'caged' glutamate (gamma-(alpha-carboxy-2-nitrobenzyl) ester, trifluoroacetate, l-glutamic acid) at room temperature; this is only half the concentration of caged glutamate used in previous studies15, and was used in order to reduce the numbers of neurons that fire action potentials synchronously (see Supplementary Fig. S2) and to reduce the size of direct responses so that postsynaptic currents (PSCs) can be detected upon stimulation at sites close to the recorded neurons. An infrared Olympus DIC microscope with a times40, 0.8 NA water immersion lens was used to visualize and target recording electrodes to pairs of layer 2/3 pyramidal neurons with somata separated by less than 50 microm for whole-cell recordings. The mean plusminus SEM distances between recorded cells were 34.6 plusminus 3.9 microm for non-connected pairs and 35.1 plusminus 4.1 microm for connected pairs of neurons. Cell bodies of recorded neurons were at least 50 microm from the surface of the slice. Glass recording electrodes (4–6 MOmega resistance) were filled with an intracellular solution consisting of 130 mM K-gluconate, 6 mM KCl, 2 mM MgCl2, 0.2 mM EGTA, 10 mM HEPES, 2.5 mM Na2ATP, 0.5 mM Na2GTP, 10 mM K-phosphocreatine and 0.3% biocytin, adjusted with KOH to pH 7.25. For some experiments in which IPSCs were recorded, potassium was replaced with cesium. All intracellular recordings had access resistances less than 20 MOmega. In all paired recordings, connections between neuron pairs were assessed by injecting current to evoke action potentials in one of the cells recorded in current-clamp while testing for PSCs during voltage-clamp recording in the other cell. For each pair, connections were tested in both directions for at least 50 trials, generating single action potentials in each presynaptic neuron. When connections were not detected with this procedure, they were also tested by stimulation in trains of 4–5 action potentials at 50 Hz to induce possible potentiation of weak connections. Control experiments measuring spatial and temporal properties of photostimulation-evoked action potentials (see Supplementary Information) used extracellular loose-patch recordings made with the same recording electrodes, except that they were filled with ACSF.

Photostimulation was achieved by uncaging glutamate with 10 ms flashes of ultraviolet light from an argon-ion laser focused through the times40 microscope objective. This results in the generation of action potentials only in neurons with cell bodies within 100 microm (and usually less than 50 microm) of the site of uncaging (see Supplementary Fig. S1). Photostimulation-evoked synaptic currents were measured from voltage-clamped neurons, with the holding potentials at -65 mV to measure EPSCs and at 0 mV to measure IPSCs. Spontaneous synaptic currents were also recorded in interleaved trials with no stimulation.

Data analysis

Maps of photostimulation sites were aligned to laminar borders in fixed and stained tissue15 (for example, Figs 1 and 3) and each site was assigned a laminar identity. Sites within 50 microm of laminar borders were discarded from further analyses in order to limit the number of evoked synaptic currents arising from neurons with cell bodies potentially outside the stimulated layer. The electrical recordings from photostimulation and no-stimulation (control) trials were analysed using peak analysis software from Synaptosoft and other custom software. The times of onset and amplitudes of all EPSCs or IPSCs occurring within 150 ms of stimulation were marked. Rise times of PSCs were measured as the time taken for the amplitude to increase from 10% to 90% of its peak value. Results from analyses of the laminar sources and strengths of excitatory and inhibitory input to layer 2/3 pyramidal neurons (not shown) were indistinguishable from those described previously15, and there were no systematic differences that correlated with results from cross-correlation analyses.

Cross-correlograms of EPSCs and/or IPSCs were computed for each pair of simultaneously recorded layer 2/3 pyramidal neurons; separate correlograms were computed for stimulation sites from each cortical layer (layers 2/3, 4 and 5 for EPSCs and layers 2/3 and 4 for IPSCs). Other layers provided weak or variable input to recorded neurons, preventing evoked PSCs from being clearly distinguished from spontaneous PSCs. Correlograms were also computed for spontaneous PSCs. Cross-correlation data were binned into histograms using 4 ms bins; the central bin included values of 0 plusminus 2 ms. Data from the stimulation trials (from the same layer) were also used to create shifted correlograms for each layer and cell pair14. To calculate the correlation probability, the shifted correlogram was subtracted from the unshifted correlogram for the corresponding layer, and then the value in the central bin was divided by the average estimated total number of evoked PSCs (for the two cells) observed for all trials in the relevant layer. The average number of evoked PSCs was calculated as the total number of measured PSCs for 'cell a' minus the expected number of spontaneous PSCs for that cell, plus the same value calculated for 'cell b', divided by 2.

Top

References

  1. Mountcastle, V. B. Introduction. Computation in cortical columns. Cereb. Cortex 13, 2–4 (2003) | Article | PubMed | ISI |
  2. Hubel, D. H. & Wiesel, T. N. Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. J. Physiol. (Lond.) 160, 106–154 (1962) | PubMed | ISI | ChemPort |
  3. Chapman, B. , Zahs, K. R. & Stryker, M. P. Relation of cortical cell orientation selectivity to alignment of receptive fields of the geniculocortical afferents that arborize within a single orientation column in ferret visual cortex. J. Neurosci. 11, 1347–1358 (1991) | PubMed | ISI | ChemPort |
  4. Ferster, D. , Chung, S. & Wheat, H. Orientation selectivity of thalamic input to simple cells of cat visual cortex. Nature 380, 249–252 (1996) | Article | PubMed | ISI | ChemPort |
  5. Alonso, J. M. , Usrey, W. M. & Reid, R. C. Rules of connectivity between geniculate cells and simple cells in cat primary visual cortex. J. Neurosci. 21, 4002–4015 (2001) | PubMed | ISI | ChemPort |
  6. Mooser, F. , Bosking, W. H. & Fitzpatrick, D. A morphological basis for orientation tuning in primary visual cortex. Nature Neurosci. 7, 872–879 (2004) | Article |
  7. Callaway, E. M. Local circuits in primary visual cortex of the macaque monkey. Annu. Rev. Neurosci. 21, 47–74 (1998) | Article | PubMed | ISI | ChemPort |
  8. Thomson, A. M. & Morris, O. T. Selectivity in the inter-laminar connections made by neocortical neurones. J. Neurocytol. 31, 239–246 (2002) | Article | PubMed | ISI |
  9. Thomson, A. M. , West, D. C. , Wang, Y. & Bannister, A. P. Synaptic connections and small circuits involving excitatory and inhibitory neurons in layers 2–5 of adult rat and cat neocortex: triple intracellular recordings and biocytin labelling in vitro. Cereb. Cortex 12, 936–953 (2002) | Article | PubMed | ISI |
  10. Hellwig, B. A quantitative analysis of the local connectivity between pyramidal neurons in layers 2/3 of the rat visual cortex. Biol. Cybern. 82, 111–121 (2000) | PubMed | ISI | ChemPort |
  11. Hellwig, B. , Schuz, A. & Aertsen, A. Synapses on axon collaterals of pyramidal cells are spaced at random intervals: a Golgi study in the mouse cerebral cortex. Biol. Cybern. 71, 1–12 (1994) | Article | PubMed | ChemPort |
  12. Braitenberg, V. & Schuz, A. Anatomy of the Cortex (Springer-Verlag, Berlin, 1991)
  13. Binzegger, T. , Douglas, R. J. & Martin, K. A. A quantitative map of the circuit of cat primary visual cortex. J. Neurosci. 24, 8441–8453 (2004) | Article | PubMed | ISI | ChemPort |
  14. Aertsen, A. M. , Gerstein, G. L. , Habib, M. K. & Palm, G. Dynamics of neuronal firing correlation: modulation of "effective connectivity". J. Neurophysiol. 61, 900–917 (1989) | PubMed | ISI | ChemPort |
  15. Dantzker, J. L. & Callaway, E. M. Laminar sources of synaptic input to cortical inhibitory interneurons and pyramidal neurons. Nature Neurosci. 3, 701–707 (2000) | Article |
  16. Kawaguchi, Y. & Kondo, S. Parvalbumin, somatostatin and cholecystokinin as chemical markers for specific GABAergic interneuron types in the rat frontal cortex. J. Neurocytol. 31, 277–287 (2002) | Article | PubMed | ISI |
  17. Hebb, D. O. The Organization of Behavior (Wiley, New York, 1949)
  18. Yoshimura, Y. , Ohmura, T. & Komatsu, Y. Two forms of synaptic plasticity with distinct dependence on age, experience, and NMDA receptor subtype in rat visual cortex. J. Neurosci. 23, 6557–6566 (2003) | PubMed | ISI | ChemPort |
  19. Sur, M. , Schummers, J. & Dragoi, V. Cortical plasticity: time for a change. Curr. Biol. 12, R168–R170 (2002) | Article | PubMed | ISI | ChemPort |
  20. Katz, L. C. & Shatz, C. J. Synaptic activity and the construction of cortical circuits. Science 274, 1133–1138 (1996) | Article | PubMed | ISI | ChemPort |
  21. Sawatari, A. & Callaway, E. M. Diversity and cell type specificity of local excitatory connections to neurons in layer 3B of monkey primary visual cortex. Neuron 25, 459–471 (2000) | Article | PubMed | ISI | ChemPort |
  22. Schubert, D. , Kotter, R. , Zilles, K. , Luhmann, H. J. & Staiger, J. F. Cell type-specific circuits of cortical layer IV spiny neurons. J. Neurosci. 23, 2961–2970 (2003) | PubMed | ISI | ChemPort |
  23. Agmon, A. & Connors, B. W. Correlation between intrinsic firing patterns and thalamocortical synaptic responses of neurons in mouse barrel cortex. J. Neurosci. 12, 319–329 (1992) | PubMed | ISI | ChemPort |
  24. Gibson, J. R. , Beierlein, M. & Connors, B. W. Two networks of electrically coupled inhibitory neurons in neocortex. Nature 402, 75–79 (1999) | Article | PubMed | ISI | ChemPort |
  25. Gonchar, Y. & Burkhalter, A. Connectivity of GABAergic calretinin-immunoreactive neurons in rat primary visual cortex. Cereb. Cortex 9, 683–696 (1999) | Article | PubMed | ISI | ChemPort |
  26. Gonchar, Y. & Burkhalter, A. Distinct GABAergic targets of feedforward and feedback connections between lower and higher areas of rat visual cortex. J. Neurosci. 23, 10904–10912 (2003) | PubMed | ISI | ChemPort |
  27. Meskenaite, V. Calretinin-immunoreactive local circuit neurons in area 17 of the cynomolgus monkey, Macaca fascicularis. J. Comp. Neurol. 379, 113–132 (1997) | Article | PubMed | ISI | ChemPort |
  28. Staiger, J. F. et al. Innervation of interneurons immunoreactive for VIP by intrinsically bursting pyramidal cells and fast-spiking interneurons in infragranular layers of juvenile rat neocortex. Eur. J. Neurosci. 16, 11–20 (2002) | Article | PubMed | ISI |
  29. Girman, S. V. , Sauve, Y. & Lund, R. D. Receptive field properties of single neurons in rat primary visual cortex. J. Neurophysiol. 82, 301–311 (1999) | PubMed | ISI | ChemPort |
  30. DeAngelis, G. C. , Ghose, G. M. , Ohzawa, I. & Freeman, R. D. Functional micro-organization of primary visual cortex: receptive field analysis of nearby neurons. J. Neurosci. 19, 4046–4064 (1999) | PubMed | ISI | ChemPort |
Top

Supplementary Information

Supplementary information accompanies this paper.

Top

Acknowledgements

We are grateful for support from the National Institutes of Health. We thank Y. Komatsu and F. Briggs and members of the Callaway laboratory for discussions.

Top

Competing interests statement

The authors declare no competing financial interests.

MORE ARTICLES LIKE THIS

These links to content published by NPG are automatically generated.

NEWS AND VIEWS

A tale of two spikes

Nature Neuroscience News and Views (01 Apr 1999)

Extra navigation

.

naturejobs

natureproducts


ADVERTISEMENT