We explore the extent to which neocortical circuits generalize, i.e., to what extent can neocortical neurons and the circuits they form be considered as canonical? We find that, as has long been suspected by cortical neuroanatomists, the same basic laminar and tangential organization of the excitatory neurons of the neocortex is evident wherever it has been sought. Similarly, the inhibitory neurons show characteristic morphology and patterns of connections throughout the neocortex. We offer a simple model of cortical processing that is consistent with the major features of cortical circuits: The superficial layer neurons within local patches of cortex, and within areas, cooperate to explore all possible interpretations of different cortical input and cooperatively select an interpretation consistent with their various cortical and subcortical inputs.

We developed a quantitative description of the circuits formed in cat area 17 by estimating the "weight" of the projections between different neuronal types. To achieve this, we made three-dimensional reconstructions of 39 single neurons and thalamic afferents labeled with horseradish peroxidase during intracellular recordings in vivo. These neurons served as representatives of the different types and provided the morphometrical data about the laminar distribution of the dendritic trees and synaptic boutons and the number of synapses formed by a given type of neuron. Extensive searches of the literature provided the estimates of numbers of the different neuronal types and their distribution across the cortical layers. Applying the simplification that synapses between different cell types are made in proportion to the boutons and dendrites that those cell types contribute to the neuropil in a given layer, we were able to estimate the probable source and number of synapses made between neurons in the six layers. The predicted synaptic maps were quantitatively close to the estimates derived from the experimentalelectronmicroscopicstudiesforthecaseofthemainsourcesofexcitatoryandinhibitoryinputtothespinystellatecells,whichform a major target of layer 4 afferents. The map of the whole cortical circuit shows that there are very few "strong" but many "weak" excitatory projections, each of which may involve only a few percentage of the total complement of excitatory synapses of a single neuron.

The majority of synapses in the mammalian cortex originate from cortical neurons. Indeed, the largest input to cortical cells comes from neighboring excitatory cells. However, most models of cortical development and processing do not reflect the anatomy and physiology of feedback excitation and are restricted to serial feedforward excitation. This report describes how populations of neurons in cat visual cortex can use excitatory feedback, characterized as an effective "network conductance", to amplify their feedforward input signals and demonstrates how neuronal discharge can be kept proportional to stimulus strength despite strong, recurrent connections that threaten to cause runaway excitation. These principles are incorporated into models of cortical direction and orientation selectivity that emphasize the basic design principles of cortical architectures.

SUMMARY1. We have studied the neuronal circuitry and structure-function relationships of single neurones in the striate visual cortex of the cat using a combination of electrophysiological and anatomical techniques.2. Glass micropipettes filled with horseradish peroxidase were used to record extracellularly from single neurones. After studying the receptive field properties, the afferent inputs of the neurones were studied by determining their latency of response to electrical stimulation at different positions along the optic pathway. Some cells were thus classified as receiving a mono-or polysynaptic input from afferents of the lateral geniculate nucleus (l.g.n.), via X-or Y-like retinal ganglion cells.3. Two striking correlations were found between dendritic morphology and receptive field type. All spiny stellate cells, and all star pyramidal cells in layer 4A, had receptive fields with spatially separate on and off subfields (S-type receptive fields). All the identified afferent input to these, the major cell types in layer 4, was monosynaptic from X-or Y-like afferents.4. Neurones receiving monosynaptic X-or Y-like input were not strictly segregated in layer 4 and the lower portion of layer 3. Nevertheless the X-and Y-like l.g.n. fibres did not converge on any of the single neurones so far studied.5. Monosynaptic input from the l.g.n. afferents was not restricted to cells lying within layers 4 and 6, the main termination zones of the l.g.n. afferents, but was also received by cells lying in layers 3 and 5.6. The projection pattern of cells receiving monosynaptic input differed widely, depending on the laminar location of the cell soma. This suggests the presence of a number of divergent paths within the striate cortex.7. Cells receiving indirect input from the l.g.n. afferents were located mainly within layers 2, 3 and 5. Most pyramidal cells in layer 3 had axons projecting out ofthe striate cortex, while many axons of the layer 5 pyramids did not.8. The layer 5 cells showed the most morphological variation of any layer, were the most difficult to activate by electrical stimulation, and contained some cells which responded with the longest latencies of any cells in the striate cortex. This suggests that they were several synapses distant from the J.g.n. input.9. The majority of cells in layers 2, 3, 4 and 6 had the same basic S-type receptive field structure.

1. We have studied in vivo the intracellular responses of neurones in cat visual cortex to electrical pulse stimulation of the cortical afferents and have developed a microcircuit that simulates much of the experimental data. 2. Inhibition and excitation are not separable events, because individual neurones are embedded in microcircuits that contribute strong population effects. Synchronous electrical activation of the cortex inevitably set in motion a sequence of excitation and inhibition in every neurone we recorded. The temporal form of this response depends on the cortical layer in which the neurone is located. Superficial layer (layers 2 + 3) pyramidal neurones show a more marked polysynaptic excitatory phase than the pyramids of the deep layers (layers 5 + 6). 3. Excitatory effects on pyramidal neurones, particularly the superficial layer pyramids, are in general not due to monosynaptic input from thalamus, but polysynaptic input from cortical pyramids. Since the thalamic input is transient it does not provide the major, sustained excitation arriving at any cortical neurone. Instead the intracortical excitatory connections provide the major component of the excitation. 4. The polysynaptic excitatory response would be sustained well after the stimulus, were it not for the suppressive effect of intracortical inhibition induced by the pulse stimulation. 5. Intracellular recording combined with ionophoresis of y-aminobutyric acid (GABA) agonists and antagonists showed that intracortical inhibition is mediated by GABAA and GABAB receptors. The GABAA component occurs in the early phase of the impulse response. It is reflected in the strong hyperpolarization that follows the excitatory response and lasts about 50 ms. The GABAB component occurs in the late phase of the response, and is reflected in a sustained hyperpolarization that lasts some 200-300 ms. Both components are seen in all cortical pyramidal neurones. However, the GABAA component appears more powerful in deep layer pyramids than superficial layer pyramids. 6. The microcircuit simulates with good fidelity the above data from experiments in vivo and provides a novel explanation for the apparent lack of significant inhibition during visual stimulation. The basic circuit may be common to all cortical areas studied and thus the microcircuit may be a 'canonical' microcircuit for neocortex.

In layer 4 of cat visual cortex, the monocular, concentric receptive fields of thalamic neurons, which relay retinal input to the cortex, are transformed into 'simple' cortical receptive fields that are binocular and selective for the precise orientation, direction of motion, and size of the visual stimulus. These properties are thought to arise from the pattern of connections from thalamic neurons, although anatomical studies show that most excitatory inputs to layer 4 simple cells are from recurrently connected circuits of cortical neurons. We examined single fibre inputs to spiny stellate neurons. We examined single fibre inputs to spiny stellate neurons in slices of cat visual cortex, and conclude that thalamocortical synapses are powerful and the responses they evoke are unusually invariant for central synapses. However, the responses to intracortical inputs, although less invariant, are strong enough to provide most of the excitation to simple cells in vivo. Our results suggest that the recurrent excitatory circuits of cortex may amplify the initial feedforward thalamic signal, subserving dynamic modifications of the functional properties of cortical neurons.

We have used microanatomy derived from single neurons, and in vivo intracellular recordings to develop a simplified circuit of the visual cortex. The circuit explains the intracellular responses to pulse stimulation in terms of the interactions between three basic populations of neurons, and reveals the following features of cortical processing that are important to computational theories of neocortex. First, inhibition and excitation are not separable events. Activation of the cortex inevitably sets in motion a sequence of excitation and inhibition in every neuron. Second, the thalamic input does not provide the major excitation arriving at any neuron. Instead the intracortical excitatory connections provide most of the excitation. Third, the time evolution of excitation and inhibition is far longer than the synaptic delays of the circuits involved. This means that cortical processing cannot rely on precise timing between individual synaptic inputs.

No abstract