Showing papers by "Kevan A. C. Martin published in 1994"

Polyneuronal innervation of spiny stellate neurons in cat visual cortex

Abstract: Our hypothesis was that spiny stellate neurons in layer 4 of cat visual cortex receive polyneuronal innervation. We characterised the synapses of four likely sources of innervation by three simple criteria: the type of synapse, the target (spine, dendritic shaft), and the area of the presynaptic bouton. The layer 6 pyramids had the smallest boutons and formed asymmetric synapses mainly with the dendritic shaft. The thalamic afferents had the largest boutons and formed asymmetric synapses mainly with spines. The spiny stellates had medium-sized boutons and formed asymmetric synapses mainly with spines. We used these to make a "template" to match against the boutons forming synapses with the spiny stellate dendrite. Of the asymmetric synapses, 45% could have come from layer 6 pyramidal neurons, 28% from spiny stellate neurons, and 6% from thalamic afferents. The remaining 21% of asymmetric synapses could not be accounted for without assuming some additional selectivity of the presynaptic axons. Additional asymmetric synapses may come from a variety of sources, including other cortical neurons and subcortical nuclei such as the claustrum. Of the symmetric synapses, 84% could have been provided by clutch cells, which form large boutons. The remainder, formed by small boutons, probably come from other smooth neurons in layer 4, e.g., neurogliaform and bitufted neurons. Our analysis supports the hypothesis that the spiny stellate receives polyneuronal innervation, perhaps from all the sources of boutons in layer 4. Although layer 4 is the major recipient of thalamic afferents, our results show that they form only a few percent of the synapses of layer 4 spiny stellate neurons.

Map of the synapses formed with the dendrites of spiny stellate neurons of cat visual cortex

Abstract: The synaptic input of six spiny stellate neurons in sublamina 4A of cat area 17 was assessed by electron microscopy. The neurons were physiologically characterized and filled with horseradish peroxidase in vivo. After processing the neurons were reconstructed at the light microscopic level using computer-assisted methods and analyzed quantitatively. The extensive branching of the dendritic tree about 50 microns from the soma meant that the distal branches constituted five times the length of proximal dendrite. Proximal and distal portions of a single dendrite from each neuron were examined in series of ultrathin sections (1,456 sections) in the electron microscope. The majority (79%) of the 263 synapses examined were asymmetric; the remainder (21%) were symmetric. Symmetric synapses formed 35% of synapses sampled on proximal dendrites and were usually located on the shaft. They formed only 4% of synapses sampled on distal dendrites. Spines accounted for less than half of the total asymmetric synapses (45%); the remainder were on shafts. Symmetric synapses formed with four of 92 spines. Nine spines formed no synapses. Spiny stellate neurons in cat visual cortex appear to differ considerably from pyramidal neurons in having a significant asymmetric (excitatory) synaptic input to the dendritic shaft.

Synaptic output of physiologically identified spiny stellate neurons in cat visual cortex

Abstract: Spiny stellate neurons of area 17 of the cat's visual cortex were physiologically characterised and injected intracellularly with horseradish peroxidase. Six neurons from sublamina 4A were selected. Five had the S-type of simple receptive fields; one had a complex receptive field. Their axons formed boutons mainly in layers 3 and 4. An electron microscopic examination of 45 boutons showed that each bouton formed one asymmetric synapse on average. Spines were the most frequent synaptic target (74%); dendritic shafts formed the remainder (26%). On the basis of ultrastructural characteristics, 8% of the target dendrites were characterised as originating from smooth gamma-aminobutyrate-ergic (GABAergic) neurons. Thus the major output of spiny stellate neurons is to other spiny neurons, probably pyramidal neurons in layer 3 and spiny stellates in layer 4.

Hybrid analog-digital architectures for neuromorphic systems

Abstract: Signal restoration is necessary to perform computations of significant complexity. In digital computers each state variable is restored to a binary value, but this strategy is incompatible with analog computation. Nevertheless, cortical neurons, whose major mode of operation is analog, are able to perform prodigious feats of computation. The authors' research on visual cortex suggests that cortical neurons are able to compute reliably because they are organized into populations in which the signal at each neuron is restored to an appropriate analog value by a collective strategy. The strategy depends on feedback amplification that restores an input signal towards a stored analog memory. This principle is similar to recall by autoassociative neural networks. Networks of cortical amplifiers can solve simple visual processing tasks. They are well-suited to sensory processing because the same principle that restores their analog signals can also extract meaningful features from ambiguous sensory input. The authors describe a hybrid analog-digital CMOS architecture for constructing networks of cortical amplifiers. This neuromorphic architecture is a step towards exploring analog computers whose distributed signal restoration permits them to perform reliably sequential computations of great depth. >