No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Cellular organization of reciprocal patchy networks in layer III of cat visual cortex (area 17)
Abstract
There is no direct information available concerning the exact spatial characteristics of long-range axons and their relationship with the patchy phenomena observed after extracellular injection of retrograde tracers. In the present study, using the recently introduced neuronal tracer biocytin, we demonstrate by detailed three-dimensional reconstruction of 10 pyramidal cells in layer III, that their clustered axonal terminals form a specific patchy network in layers II and III. The reconstructed network occupied an area of 6.5 x 3.5 mm parallel to the cortical surface elongated in an anteroposterior direction. The average centre-to-centre distance between patches within the network was 1.1 mm. On average, the axonal field of each of the 10 pyramidal cells contained a total of 417 boutons at four to eight distinct sites (patches), and in each patch, an average of 79 boutons was provided by the same cell. The identified connections between the patches were predominantly reciprocal. Detailed analyses have shown that many pyramidal cells of the network are directly interconnected so that each of them can receive one to four, chiefly axospinous, contacts onto the distal segment of its apical and basal dendrites from the axon of another pyramidal cell belonging to a different patch labelled from the same injection site. We hypothesize that the possible functional role of the network is to link remote sites with similar physiological characteristics, such as orientation preference, supporting the model of Mitchison and Crick [(1982) Proc. natn. Acad. Sci. U.S.A. 79, 3661-3665].
... For supragranular pyramidal neurons, multiple studies report a mixed axonal architecture of a bouton-dense local or home cluster, spatially coincident with the soma of the individual neuron, and a variable number of distal terminal clusters (cat V1: 4-8, Kisvarday and Eysel, 1992;up to seven;Binzegger et al., 2007;macaque TE, up to 43, Tanigawa et al., 2005). These are spatially extended from the soma for ∼1.0 mm in cat V1, and up to 8.0 mm in macaque area TE (Tanigawa et al., 2005;Wang et al., 2017). ...
... Distal clusters commonly receive uo to 10 converging branches from the same neuron, which contribute differing proportions of boutons to a single cluster (cat visual cortex: n = 10 neurons, Martin et al., 2014;n = 33 neurons;Martin et al., 2017, n = 50 neurons). Intracellular fills reveal that a single branch of the same neuron can diverge to more than one cluster, thereby contacting spatially separated postsynaptic neurons (cat visual cortex: Kisvarday and Eysel, 1992). The distal patches of one neuron will have membership in other local networks. ...
... The architectural significance of this mixed linear-and-clustered geometry is unclear; but, intriguingly in the context of cortical organization, something similar has been observed for some feedback axons (e.g., V4 to V1, figure 12 in Rockland et al., 1994) and for some pulvinocortical axons in V2 (figures 8, 17 in Rockland et al., 1999). In cat V1, the intrinsic collaterals of a single neuron can be myelinated or unmyelinated in their trajectory to distal clusters (Kisvarday and Eysel, 1992;Koestinger et al., 2017). Myelination has commonly been investigated in the context of conduction velocity, but is increasingly considered with other processes such as adaptive remodeling in response to different conditions of sensory and non-sensory processing (Fields, 2015;Bonetto et al., 2020;Yang et al., 2020;de Faria et al., 2021). ...
... There are several studies where extracellular tracer injections were used to visualize the projection pattern of localized groups of thousands of neurons. There are only very few neuroanatomical studies dealing with the distant projection pattern of single cells (Ojima et al., 1991;Kisvarday & Eysel, 1992). It is still largely unknown how many of the synapses of a typical pyramidal cell are established locally, or to which degree the remaining long-range projections are spatially clustered. ...
... Generalized over many studies, we propose to set the local connectivity range to r loc = 500 µm, which is the upper boundary we found in the literature. For example in (Levitt, Lewis, Yoshioka and Lund 1993;Lewis, Melchitzky and Burgos 2002;Melchitzky, Gonzale-Burgos, Barrionuevo and Lewis 2001;Ojima, Honda and Jones 1991) it is r loc ∼ 300 µm, and in (Kisvarday, Martin, Freund, Magloczky, Whitteridge and Somogyi 1986;Stettler, Das, Bennet and Gilbert 2002;Burkhalter and Charles 1990;Kisvarday and Eysel 1992) it is r loc ∼ 400 − 500 µm. The main criterion is the range of the dendritic and local axonal arborizations (Ghosh, Fyffe and Porter 1988;Lohmann and Rörig 1994;Ghosh and Porter 1988). ...
... Ojima, Honda and Jones (1991) mention that there are more local than long axon collaterals (approximately 9:3). Kisvarday and Eysel (1992) count the boutons made by axon collaterals and find twice as many boutons in the local compared to the distant range. The relation between local and distant synapses may differ, depending on the specific location or function of the neuron, the cortical area, the species, and many more factors. ...
... Inclusion of cell-type and layer-specific connections in the present hybrid scheme has some implications for how we proceed with setting up equivalent populations consisting of morphologically detailed model neurons. Different cell types belonging to a particular population may have different spatial distributions of synapses, or the populations may consist of different morphological classes of neurons (Kisvárday and Eysel 1992;Nowak 2003;Stepanyants et al. 2008). An example is layer 4 in which spiny stellate cells lack apical dendrites, while pyramidal cells have apical dendrites extending into layer 1. ...
... For each of the 16 cell types, we acquired representative morphological reconstructions of predominantly cat visual cortex neurons from several sources (Contreras et al. 1997;Kisvárday and Eysel 1992;Mainen and Sejnowski 1996;Stepanyants et al. 2008) (cf. Fig. 2, Table 7). ...
... p4, ss4 L23 , ss4 L4 , cf., Fig. 2). Cell types y: (Kisvárday and Eysel 1992;Mainen and Sejnowski 1996;Contreras et al. 1997;Stepanyants et al. 2008) • Cat somatosensory cortex from NeuroMorpho.org (Contreras et al. 1997;Ascoli et al. 2007). ...
The measurement of local field potentials (LFP), the low-frequency part of extracellularly recorded potentials, is one of the most commonly used methods for probing hippocampal and cortical activity in vivo. It offers the possibility to monitor the activity of many neurons close to the recording electrode simultaneously but has the limitation that it may be difficult to interpret and relate to the underlying neuronal activity. The recording electrode picks up activity from proximal neurons, but what about more distant neurons? An important piece of information for a correct interpretation of the LFP is to decide the size of the tissue that substantially contributes to the LFP, i.e., the reach of the LFP signal. In this chapter we present a simple model that describes how population geometry, spatial decay of single-cell LFP contributions, and correlation between LFP sources determine the relation between LFP amplitude and population size and use it to study the spatial reach of the LFP. The model can also be used to study different frequency bands of the LFP separately as well as the spatial decay outside the active neuronal population.
... Inclusion of cell-type and layer-specific connections in the present hybrid scheme has some implications for how we proceed with setting up equivalent populations consisting of morphologically detailed model neurons. Different cell types belonging to a particular population may have different spatial distributions of synapses, or the populations may consist of different morphological classes of neurons (Kisvárday and Eysel 1992;Nowak 2003;Stepanyants et al. 2008). An example is layer 4 in which spiny stellate cells lack apical dendrites, while pyramidal cells have apical dendrites extending into layer 1. ...
... For each of the 16 cell types, we acquired representative morphological reconstructions of predominantly cat visual cortex neurons from several sources (Contreras et al. 1997;Kisvárday and Eysel 1992;Mainen and Sejnowski 1996;Stepanyants et al. 2008) (cf. Fig. 2, Table 7). ...
... p4, ss4 L23 , ss4 L4 , cf., Fig. 2). Cell types y: (Kisvárday and Eysel 1992;Mainen and Sejnowski 1996;Contreras et al. 1997;Stepanyants et al. 2008) • Cat somatosensory cortex from NeuroMorpho.org (Contreras et al. 1997;Ascoli et al. 2007). ...
With rapidly advancing multi-electrode recording technology, the local field potential (LFP) has again become a popular measure of neuronal activity in both research and clinical applications. Proper understanding of the LFP requires detailed mathematical modeling incorporating the anatomical and electrophysiological features of neurons near the recording electrode, as well as synaptic inputs from the entire network. Here we propose a hybrid modeling scheme combining efficient point-neuron network models with biophysical principles underlying LFP generation by real neurons. The LFP predictions rely on populations of network-equivalent multicompartment neuron models with layer-specific synaptic connectivity, can be used with an arbitrary number of point-neuron network populations, and allows for a full separation of simulated network dynamics and LFPs. We apply the scheme to a full-scale cortical network model for a ∼1 mm² patch of primary visual cortex, predict laminar LFPs for different network states, assess the relative LFP contribution from different laminar populations, and investigate effects of input correlations and neuron density on the LFP. The generic nature of the hybrid scheme and its public implementation in
HybridLFPy
form the basis for LFP predictions from other and larger point-neuron network models, as well as extensions of the current application with additional biological detail.
... The most intensively investigated brain areas belong to visual cortices. In reference to this, there are several studies dealing with visual cortices in different species such as treeshrew ( Rockland and Lund, 1982; Bosking et al., 1997 ), ferret ( Rockland, 1985; Wallace and Bajwa, 1991 ), cat ( Gilbert and Wiesel, 1983; Martin and Whitteridge, 1984; Matsubara et al., 1985; Luhmann et al., 1990; Kisvárday and Eysel, 1992; Löwel and Singer, 1992; Kisvárday et al., 1997; Buzás et al., 2006 ), monkey ( Rockland, 1985; Lund et al., 1993; Yoshioka et al., 1996; Malach et al., 1993; Sincich and Blasdel, 2001; Angelucci et al., 2002; Tanigawa et al., 2005 ), and human ( Burkhalter and Bernardo, 1989 ). There are two important constraints that determine the method of choice while investigating long-range lateral connections. ...
... Contemporary anatomical data obtained for the cat visual cortex concluded on the basis of spatial parameters of cortico-cortical connections that these implied a relationship with the functional architecture of the cortex and in particular linked columns of similar orientations. Single axon reconstructions along with population labeling showed that the average patch size has a diameter of approximately 200–400 μm ( Gilbert and Wiesel, 1983; Luhmann et al., 1986; Kisvárday and Eysel, 1992 ), a dimension that matches with the width of an iso-orientation column. Furthermore, the average inter-patch distance is 600–1100 μm ( Luhmann et al., 1986; Löwel et al., 1987; Kisvárday and Eysel, 1992 ) corresponding to the average hypercolumn distance wherein the orientation preference completes a full cycle ( Hubel and Wiesel, 1974 ). ...
... Single axon reconstructions along with population labeling showed that the average patch size has a diameter of approximately 200–400 μm ( Gilbert and Wiesel, 1983; Luhmann et al., 1986; Kisvárday and Eysel, 1992 ), a dimension that matches with the width of an iso-orientation column. Furthermore, the average inter-patch distance is 600–1100 μm ( Luhmann et al., 1986; Löwel et al., 1987; Kisvárday and Eysel, 1992 ) corresponding to the average hypercolumn distance wherein the orientation preference completes a full cycle ( Hubel and Wiesel, 1974 ). Importantly, the above functional and anatomical data provide strong support to the early recognition that the spatial organization of lateral connections is ideally suited to promote spatially extended excitatory convergence between cells of similar orientations for constructing elongated receptive fi elds (RFs) ( Bolz and Gilbert, 1989; Gilbert, 1977; Gray et al., 1989 ). Inevitably, these data provide some experimental support to the concept of " like connects to like " and through this idea, further explains perceptual phenomena and a number of psychophysical observations where long-range connections are expected to play a key role (for review see Gilbert, 1992 ; Kapadia et al., 1995 ). ...
In the cerebral cortex, the clustered or patchy nature of the neural network can be found in many mammalian species. Therefore, this can be taken as representing a basic architectural feature of the brain organization. Our aim is to provide an overview of the patchy connectivity system considering topography in relation to the functional connectivity, ultimately delineating its putative role through examples excerpted from the literature and also to introduce some theoretical considerations. For consistency, we focus on the primary visual cortex since most functional anatomical data are available only for this brain area. We will deal only with the patchy lateral network because, despite the wealth of information, it is still not clear what general principles of cortical processing this network subserves.
... To combine encoded features, it has been proposed that neurons use intricate networks of longrange horizontal connections which often preferentially connect clusters of neurons that have similar orientation preference (Gilbert and Wiesel, 1989;Kisvarday et al., 1989;Kisvárday and Eysel, 1992;Bosking et al., 1997). Such functional specificity may allow single neurons to combine information from a larger area in the visual field (i.e. ...
... Our results show that synchronization between V1 neuron's spikes and the LFP from a neural cluster up to 2mm apart happens when both neural clusters have the same stimulus orientation preference. These results can be explained by the physiology of horizontal connections which preferentially link remote clusters of neurons with similar orientation preference (Burkhalter, 1989;Gilbert and Wiesel, 1989;Kisvarday et al., 1989;Schwarz andBolz, 1991, 1991;Kisvárday and Eysel, 1992;Malach et al., 1993Malach et al., , 1993Bosking et al., 1997). The synchronization between columns as separate as 2mm apart may help the flow of information to the downstream as several studies have shown that temporarily correlated activity of large pools of neurons can help the control of the flow of information across cortical areas (Salinas and Sejnowski, 2001;Rubino et al., 2006). ...
In the primary visual cortex, neurons with similar receptive field properties are bound together through widespread networks of horizontal connections that span orientation columns. How connectivity across the cortical surface relates to stimulus information is not fully understood. We recorded spiking activity and the local field potential (LFP) from the primary visual cortex of marmoset monkeys and examined how connectivity between distant orientation columns affect the encoding of visual orientation.
Regardless of their spatial separation, recording sites with similar orientation preferences have higher coherence between spiking activity and the local field potential than sites with different preferred orientation. Using information theoretic methods, we measured the amount of stimulus information that is shared between pairs of sites. More stimulus information can be decoded from pairs with the same preferred stimulus orientation than the pairs with a different preferred orientation, and the amount of information is significantly correlated with the magnitude of beta-band spike-LFP coherence. These effects remained after controlling for firing rate differences.
Our results thus show that spike-LFP synchronization in the beta-band is associated with the encoding of stimulus information within the primary visual cortex of marmoset monkeys.
Significance Statement
A fundamental step in processing images in the visual cortex is coordinating the neural activity across distributed populations of neurons. Here, we demonstrate that populations of neurons in the primary visual cortex of marmoset monkeys with the same stimulus orientation preference temporally coordinate their activity patterns when presented with a visual stimulus. We find maximum synchronization in the beta range depends on the similarity of orientation preference at each pair of the neural population.
... Here again there is relatively little data on corticocortical connections. There are not even estimates of convergence, as are available for several types of interneurons (Kisvardy and Eysel, 1992), because of the greater difficulty in identifying extrinsic connections in EM material. In principle, some indirect gauge of convergence can be derived by comparing the size of an injection site (of retrograde tracer) with the resultant size of projection focus, although this approach has its own limitations. ...
... Whether all of these axons are myelinated (another possible variable in the determination of conduction velocity) is not known. Other studies report that geniculocortical (Freund et al., 1989) and callosal axons are myelinated at their point of entry into and within the gray matter; large basket cell axons (jones, 1984;Kisvardy and Eysel, 1992) are also myelinated. ...
Over the last twenty-five years, there has been an extensive effort, still growing for that matter, to explore and understand the organization of extrastriate cor tex in primates. We now recognize that most of caudal neocortex is visual in some sense and that this large visual region includes many distinct areas. Some of these areas have been well defined, and connections, neural properties, and the functional consequences of deactivations have been studied. More recently, non invasive imaging of cortical activity patterns during visual tasks has led to an expanding stream of papers on extrastriate visual cortex of humans, and results have been related to theories of visual cortex organization that have emerged from research on monkeys. Against this backdrop, the time seems ripe for a review of progress and a glance at the future. One caveat important to emphasize at the very onset is that the reader may be puzzled or confused by the use of different terminologies. Individual investi gators commonly tend to favor different terminologies, but in general some prove more advantageous than others. As discussed by Rowe and Stone (1977) as well as by others, there is an unfortunate tendency for role-indicating names to lead to fixed ideas about function, in contrast to those that are more neutral and adaptable to new findings.
... These considerations on the role of extrastriate feedback lead to the prediction that the emergence of V1 gammasynchronization depends strongly on the lateral, recurrent connectivity within area V1. In area V1, there exists an extensive lattice of recurrent, patchy connections ( Figure 5B; Gilbert and Wiesel, 1983;Rockland and Lund, 1983;Lund et al., 2003) with axons that are thick (∼1-3 micron) and myelinated, presumably having high conduction velocities corresponding to axonal delays on the order of a few milliseconds (Kisvarday and Eysel, 1992). This network extends, both in cats and monkeys, to about 7 mm on average (up to ∼10 mm) and carries information over several degrees of visual field, covering an area that is several times the CRF size (Gilbert and Wiesel, 1983;Rockland and Lund, 1983;Kisvarday and Eysel, 1992;Angelucci et al., 2002a,b;Lund et al., 2003; Figure 5B). ...
... In area V1, there exists an extensive lattice of recurrent, patchy connections ( Figure 5B; Gilbert and Wiesel, 1983;Rockland and Lund, 1983;Lund et al., 2003) with axons that are thick (∼1-3 micron) and myelinated, presumably having high conduction velocities corresponding to axonal delays on the order of a few milliseconds (Kisvarday and Eysel, 1992). This network extends, both in cats and monkeys, to about 7 mm on average (up to ∼10 mm) and carries information over several degrees of visual field, covering an area that is several times the CRF size (Gilbert and Wiesel, 1983;Rockland and Lund, 1983;Kisvarday and Eysel, 1992;Angelucci et al., 2002a,b;Lund et al., 2003; Figure 5B). These long-range excitatory FIGURE 5 | Mechanisms of gamma-synchronization: connections and lamina. ...
During visual stimulation, neurons in visual cortex often exhibit rhythmic and synchronous firing in the gamma-frequency (30-90 Hz) band. Whether this phenomenon plays a functional role during visual processing is not fully clear and remains heavily debated. In this paper, we explore the function of gamma-synchronization in the context of predictive and efficient coding theories. These theories hold that sensory neurons utilize the statistical regularities in the natural world in order to improve the efficiency of the neural code, and to optimize the inference of the stimulus causes of the sensory data. In visual cortex, this relies on the integration of classical receptive field (CRF) data with predictions from the surround. Here, we outline two main hypotheses about gamma-synchronization in visual cortex. First, we hypothesize that the precision of gamma-synchronization reflects the extent to which CRF data can be accurately predicted by the surround. Second, we hypothesize that different cortical columns synchronize to the extent that they accurately predict each other’s CRF visual input. We argue that these two hypotheses can account for a large number of empirical observations made on the stimulus dependencies of gamma-synchronization. Furthermore, we show that they are consistent with the known laminar dependencies of gamma-synchronization and the spatial profile of intercolumnar gamma-synchronization, as well as the dependence of gamma-synchronization on experience and development. Based on our two main hypotheses, we outline two additional hypotheses. First, we hypothesize that the precision of gamma-synchronization shows, in general, a negative dependence on RF size. In support, we review evidence showing that gamma-synchronization decreases in strength along the visual hierarchy, and tends to be more prominent in species with small V1 RFs. Second, we hypothesize that gamma-synchronized network dynamics facilitate the emergence of spiking output that is particularly information-rich and sparse.
... These considerations on the role of extrastriate feedback lead to the prediction that the emergence of V1 gammasynchronization depends strongly on the lateral, recurrent connectivity within area V1. In area V1, there exists an extensive lattice of recurrent, patchy connections ( Figure 5B; Gilbert and Wiesel, 1983;Rockland and Lund, 1983;Lund et al., 2003) with axons that are thick (∼1-3 micron) and myelinated, presumably having high conduction velocities corresponding to axonal delays on the order of a few milliseconds (Kisvarday and Eysel, 1992). This network extends, both in cats and monkeys, to about 7 mm on average (up to ∼10 mm) and carries information over several degrees of visual field, covering an area that is several times the CRF size (Gilbert and Wiesel, 1983;Rockland and Lund, 1983;Kisvarday and Eysel, 1992;Angelucci et al., 2002a,b;Lund et al., 2003; Figure 5B). ...
... In area V1, there exists an extensive lattice of recurrent, patchy connections ( Figure 5B; Gilbert and Wiesel, 1983;Rockland and Lund, 1983;Lund et al., 2003) with axons that are thick (∼1-3 micron) and myelinated, presumably having high conduction velocities corresponding to axonal delays on the order of a few milliseconds (Kisvarday and Eysel, 1992). This network extends, both in cats and monkeys, to about 7 mm on average (up to ∼10 mm) and carries information over several degrees of visual field, covering an area that is several times the CRF size (Gilbert and Wiesel, 1983;Rockland and Lund, 1983;Kisvarday and Eysel, 1992;Angelucci et al., 2002a,b;Lund et al., 2003; Figure 5B). These long-range excitatory FIGURE 5 | Mechanisms of gamma-synchronization: connections and lamina. ...
During visual stimulation, neurons in visual cortex often exhibit rhythmic and synchronous firing in the gamma-frequency (30-90 Hz) band. Whether this phenomenon plays a functional role during visual processing is not fully clear and remains heavily debated. In this paper, we explore the function of gamma-synchronization in the context of predictive and efficient coding theories. These theories hold that sensory neurons utilize the statistical regularities in the natural world in order to improve the efficiency of the neural code, and to optimize the inference of the stimulus causes of the sensory data. In visual cortex, this relies on the integration of classical receptive field (CRF) data with predictions from the surround. Here, we outline two main hypotheses about gamma-synchronization in visual cortex. First, we hypothesize that the precision of gamma-synchronization reflects the extent to which CRF data can be accurately predicted by the surround. Second, we hypothesize that different cortical columns synchronize to the extent that they accurately predict each other’s CRF visual input. We argue that these two hypotheses can account for a large number of empirical observations made on the stimulus dependencies of gamma-synchronization. Furthermore, we show that they are consistent with the known laminar dependencies of gamma-synchronization and the spatial profile of intercolumnar gamma-synchronization, as well as the dependence of gamma-synchronization on experience and development. Based on our two main hypotheses, we outline two additional hypotheses. First, we hypothesize that the precision of gamma-synchronization shows, in general, a negative dependence on RF size. In support, we review evidence showing that gamma-synchronization decreases in strength along the visual hierarchy, and tends to be more prominent in species with small V1 RFs. Second, we hypothesize that gamma-synchronized network dynamics facilitate the emergence of spiking output that is particularly information-rich and sparse.
... 1.9; Rockland und Lund, 1982). Diese Gruppen von Neuronen liegen jeweils etwa 1 mm voneinander entfernt und sind reziprok miteinander verknüpft (Kisvárday und Eysel, 1992). Eine derartige Anordnung, in der langreichweitige Horizontalverbindungen selektiv verteilte Gruppen von Neuronen verbinden, wurde mittlerweile für zahlreiche Säugetierspezies und nicht nur im visuellen Kortex beschrieben (Katze: Gilbert und Wiesel, 1983;Kisvárday und Eysel, 1992;Frettchen: Rockland, 1985;Spitzhörnchen: Rockland und Lund, 1982;Makaken und Totenkopfaffen: Rockland und Lund, 1983;Livingstone und Hubel, 1984;Lund et al., 1993;Mensch: Burkhalter et al., 1993). ...
... Diese Gruppen von Neuronen liegen jeweils etwa 1 mm voneinander entfernt und sind reziprok miteinander verknüpft (Kisvárday und Eysel, 1992). Eine derartige Anordnung, in der langreichweitige Horizontalverbindungen selektiv verteilte Gruppen von Neuronen verbinden, wurde mittlerweile für zahlreiche Säugetierspezies und nicht nur im visuellen Kortex beschrieben (Katze: Gilbert und Wiesel, 1983;Kisvárday und Eysel, 1992;Frettchen: Rockland, 1985;Spitzhörnchen: Rockland und Lund, 1982;Makaken und Totenkopfaffen: Rockland und Lund, 1983;Livingstone und Hubel, 1984;Lund et al., 1993;Mensch: Burkhalter et al., 1993). ...
... Introduction Long-range horizontal connections (LRCs, Fig 1A) are characteristic anatomical structures observed in the primary visual cortex (V1) of various mammalian species, such as monkeys [1,2], cats [3,4], tree shrews [5,6], gray squirrels [7], ferrets [8], and rats [9,10]. Given their extraordinarily long wiring (up to 2-3 mm), LRCs are distinguished from local connections of a short lateral spread (up to 1 mm) [11][12][13][14][15]. Given the argument that the brain has evolved to develop its structure by balancing wiring costs and functional efficiency [16][17][18][19][20], it may be disadvantageous to develop such connections with high structural costs [21][22][23] unless they play a crucial role in information processing. ...
Long-range horizontal connections (LRCs) are conspicuous anatomical structures in the primary visual cortex (V1) of mammals, yet their detailed functions in relation to visual processing are not fully understood. Here, we show that LRCs are key components to organize a "small-world network" optimized for each size of the visual cortex, enabling the cost-efficient integration of visual information. Using computational simulations of a biologically inspired model neural network, we found that sparse LRCs added to networks, combined with dense local connections, compose a small-world network and significantly enhance image classification performance. We confirmed that the performance of the network appeared to be strongly correlated with the small-world coefficient of the model network under various conditions. Our theoretical model demonstrates that the amount of LRCs to build a small-world network depends on each size of cortex and that LRCs are beneficial only when the size of the network exceeds a certain threshold. Our model simulation of various sizes of cortices validates this prediction and provides an explanation of the species-specific existence of LRCs in animal data. Our results provide insight into a biological strategy of the brain to balance functional performance and resource cost.
... The third example is a network embedded into two-dimensional space introduced in a paper by Voges & Perrinet [97] to model the dynamics of neocortical networks with realistic horizontal connectivity. The "PB model", as it is called by the authors, incorporates both local and non-local connections between cells as observed for instance in the laminar structure of the visual cortex of cats [97,98]. Local connectivity (footprint ≲ 150-300 μm) is observed to be approximately isotropic, with nearby cells being more likely to be connected than cells farther apart. ...
Sustainable research on computational models of neuronal networks requires published models to be understandable, reproducible, and extendable. Missing details or ambiguities about mathematical concepts and assumptions, algorithmic implementations, or parameterizations hinder progress. Such flaws are unfortunately frequent and one reason is a lack of readily applicable standards and tools for model description. Our work aims to advance complete and concise descriptions of network connectivity but also to guide the implementation of connection routines in simulation software and neuromorphic hardware systems. We first review models made available by the computational neuroscience community in the repositories ModelDB and Open Source Brain, and investigate the corresponding connectivity structures and their descriptions in both manuscript and code. The review comprises the connectivity of networks with diverse levels of neuroanatomical detail and exposes how connectivity is abstracted in existing description languages and simulator interfaces. We find that a substantial proportion of the published descriptions of connectivity is ambiguous. Based on this review, we derive a set of connectivity concepts for deterministically and probabilistically connected networks and also address networks embedded in metric space. Beside these mathematical and textual guidelines, we propose a unified graphical notation for network diagrams to facilitate an intuitive understanding of network properties. Examples of representative network models demonstrate the practical use of the ideas. We hope that the proposed standardizations will contribute to unambiguous descriptions and reproducible implementations of neuronal network connectivity in computational neuroscience.
... The third example is a network embedded into two-dimensional space introduced in a paper by Voges & Perrinet [96] to model the dynamics of neocortical networks with realistic horizontal connectivity. The "PB model", as it is called by the authors, incorporates both local and non-local connections between cells as observed for instance in the laminar structure of the visual cortex of cats [96,97]. Local connectivity (footprint 150 − 300 µm) is observed to be approximately isotropic, with nearby cells being more likely to be connected than cells farther apart. ...
Sustainable research on computational models of neuronal networks requires published models to be understandable, reproducible, and extendable. Missing details or ambiguities about mathematical concepts and assumptions, algorithmic implementations, or parameterizations hinder progress. Such flaws are unfortunately frequent and one reason is a lack of readily applicable standards and tools for model description. Our work aims to advance complete and concise descriptions of network connectivity but also to guide the implementation of connection routines in simulation software and neuromorphic hardware systems. We first review models made available by the computational neuroscience community in the repositories ModelDB and Open Source Brain, and investigate the corresponding connectivity structures and their descriptions in both manuscript and code. The review comprises the connectivity of networks with diverse levels of neuroanatomical detail and exposes how connectivity is abstracted in existing description languages and simulator interfaces. We find that a substantial proportion of the published descriptions of connectivity is ambiguous. Based on this review, we derive a set of connectivity concepts for deterministically and probabilistically connected networks and also address networks embedded in metric space. Beside these mathematical and textual guidelines, we propose a unified graphical notation for network diagrams to facilitate an intuitive understanding of network properties. Examples of representative network models demonstrate the practical use of the ideas. We hope that the proposed standardizations will contribute to unambiguous descriptions and reproducible implementations of neuronal network connectivity in computational neuroscience.
... Long-range horizontal connections (LRCs) are characteristic circuit structures observed in the visual cortex of various mammalian species including monkeys (Stettler, Das, Bennett, & Gilbert, 2002), cats (Das & Gilbert, 1995;Gibson & Wiesel, 1979), tree shrews (Bosking et al., 1997;Chisum, Mooser, & Fitzpatrick, 2003), ferrets (Durack & Katz, 1996;Van Hooser, Heimel, Chung, & Nelson, 2006), and rats (Rumberger, Tyler, & Lund, 2001). From their long range of connection (up to 2-3 mm), LRCs are distinguished from local connections of a short lateral spread (up to 0.5 mm) (Kisvárday & Eysel, 1992;Kisvárday et al., 1986;Malach, Amir, Harel, & Grinvald, 1993;Stepanyants, Martinez, & Ferecsko, 2009;Voges, https://doi.org/10.1016/j.neunet.2020.11.013 0893-6080/© 2020 The Authors. Published by Elsevier Ltd. ...
The brain successfully performs visual object recognition with a limited number of hierarchical networks that are much shallower than artificial deep neural networks (DNNs) that perform similar tasks. Here, we show that long-range horizontal connections (LRCs), often observed in the visual cortex of mammalian species, enable such a cost-efficient visual object recognition in shallow neural networks. Using simulations of a model hierarchical network with convergent feedforward connections and LRCs, we found that the addition of LRCs to the shallow feedforward network significantly enhances the performance of networks for image classification, to a degree that is comparable to much deeper networks. We found that a combination of sparse LRCs and dense local connections dramatically increases performance per wiring cost. From network pruning with gradient-based optimization, we also confirmed that LRCs could emerge spontaneously by minimizing the total connection length while maintaining performance. Ablation of emerged LRCs led to a significant reduction of classification performance, which implies these LRCs are crucial for performing image classification. Taken together, our findings suggest a brain-inspired strategy for constructing a cost-efficient network architecture to implement parsimonious object recognition under physical constraints such as shallow hierarchical depth.
... It is wellknown that cortical neurons with a preference for a particular direction are grouped into directional columns [6,65]. Many studies documented patchy patterns of cortical horizontal interconnections that have approximately the same center-to-center distance, such as between functional columns [66][67][68][69]. In accordance with the wiring economy thesis [70], this connectional pattern, such as cortical columns per se, can be a result of an attempt to minimize the wiring cost for the intracortical connectivity. ...
A biophysically detailed description of the mechanisms of the primary vision is still being developed. We have incorporated a simplified, filter-based description of retino-thalamic visual signal processing into the detailed, conductance-based refractory density description of the neuronal population activity of the primary visual cortex. We compared four mechanisms of the direction selectivity (DS), three of them being based on asymmetrical projections of different types of thalamic neurons to the cortex, distinguishing between (i) lagged and nonlagged, (ii) transient and sustained, and (iii) On and Off neurons. The fourth mechanism implies a lack of subcortical bias and is an epiphenomenon of intracortical interactions between orientation columns. The simulations of the cortical response to moving gratings have verified that first three mechanisms provide DS to an extent compared with experimental data and that the biophysical model realistically reproduces characteristics of the visual cortex activity, such as membrane potential, firing rate, and synaptic conductances. The proposed model reveals the difference between the mechanisms of both the intact and the silenced cortex, favoring the second mechanism. In the fourth case, DS is weaker but significant; it completely vanishes in the silenced cortex.DS in the On-Off mechanism derives from the nonlinear interactions within the orientation map. Results of simulations can help to identify a prevailing mechanism of DS in V1. This is a step towards a comprehensive biophysical modeling of the primary visual system in the frameworks of the population rate coding concept.
... Due to the evidence that horizontal connections are largely reciprocal [25], it is reasonable to model this phenomenon through a symmetric distance. Since the construction of the patches P(·) was meant to restrict which cells can interact with one another, it is natural to define them so that p is connected to q if and only if q is connected to p. ...
... The complex circuitry shown in Fig. 3 is still only a simplified representation of the interactions within the primary visual cortex and between other cortical and subcortical areas. There are also recurrent horizontal connections within layers, especially II/III (Gilbert and Wiesel, 1979;Kisvarday and Eysel, 1992), feedback from extrastriate areas to layers II/III (Gilbert and Li, 2013;Salin and Bullier, 1995), and inputs and outputs from and to other subcortical areas (Coppola and Disney, 2018;Purushothaman et al., 2012). ...
Synopsis:
The primary visual cortex (V1) is the first cortical structure in the visual stream and lies in the most posterior region of the occipital lobe. It primarily receives input from the lateral geniculate nucleus and individual cells respond selectively to the orientation of edges and contours based on luminance or color contrast. V1 is the first cortical area with cells that integrate information between the two eyes, respond selectively to stereoscopic depth, and respond selectively to the direction of motion. V1 sends output projections to extrastriate visual areas along either the dorsal or ventral stream.
... The neurons for each layer use the same characteristics and morphologies as the model introduced in [19]. Morphologies are reduced from those presented in [36] and compared in [31] against cat visual cortex neurons from [37,38]. To produce the slow background oscillation patterns the noise input to different neuron populations was adjusted, as were the synaptic delay parameters. ...
Neurostimulation using weak electric fields has generated excitement in recent years due to its potential as a medical intervention. However, study of this stimulation modality has been hampered by inconsistent results and large variability within and between studies. In order to begin addressing this variability we need to properly characterise the impact of the current on the underlying neuron populations. To develop and test a computational model capable of capturing the impact of electric field stimulation on networks of neurons. We construct a cortical tissue model with distinct layers and explicit neuron morphologies. We then apply a model of electrical stimulation and carry out multiple test case simulations. The cortical slice model is compared to experimental literature and shown to capture the main features of the electrophysiological response to stimulation. Namely, the model showed 1) a similar level of depolarisation in individual pyramidal neurons, 2) acceleration of intrinsic oscillations, and 3) retention of the spatial profile of oscillations in different layers. We then apply alternative electric fields to demonstrate how the model can capture differences in neuronal responses to the electric field. We demonstrate that the tissue response is dependent on layer depth, the angle of the apical dendrite relative to the field, and stimulation strength. We present publicly available computational modelling software that predicts the neuron network population response to electric field stimulation.
... Most of them are GABAergic and connect superficial and deep cortical layers within a narrow vertical column, with no profuse horizontal connections (DeFelipe and Jones, 1985). Others in M1, such as GABAergic "basket" cells, give rise to axon collaterals that travel for Ͼ1 mm horizontally (Jones, 1975;DeFelipe et al., 1986), as in the visual cortex (Kisvárday and Eysel, 1992;Fitzpatrick, 1996). Thus, axon collaterals display varying spatial configurations and functions. ...
Cortical networks are characterized by the origin, destination, and reciprocity of their connections, as well as by the diameter, conduction velocity, and synaptic efficacy of their axons. The network formed by parietal and frontal areas lies at the core of cognitive-motor control because the outflow of parietofrontal signaling is conveyed to the subcortical centers and spinal cord through different parallel pathways, whose orchestration determines, not only when and how movements will be generated, but also the nature of forthcoming actions. Despite intensive studies over the last 50 years, the role of corticocortical connections in motor control and the principles whereby selected cortical networks are recruited by different task demands remain elusive. Furthermore, the synaptic integration of different cortical signals, their modulation by transthalamic loops, and the effects of conduction delays remain challenging questions that must be tackled to understand the dynamical aspects of parietofrontal operations. In this article, we evaluate results from nonhuman primate and selected rodent experiments to offer a viewpoint on how corticocortical systems contribute to learning and producing skilled actions. Addressing this subject is not only of scientific interest but also essential for interpreting the devastating consequences for motor control of lesions at different nodes of this integrated circuit. In humans, the study of corticocortical motor networks is currently based on MRI-related methods, such as resting-state connectivity and diffusion tract-tracing, which both need to be contrasted with histological studies in nonhuman primates.
... Interactions between neocortical columns were also considered. Studies show that local excitation in a particular column (which engages pyramidal and interneuron populations within that column) induces activation in adjacent columns due to axonal connections of pyramidal neurons across different local circuits, also known as "patchy connections" [45][46][47][48] . Interneurons have very localized axons and do not contribute to the patchy connections which are made by pyramidal cells 400-1000 μm 49-51 from a given neocortical column. ...
Transcranial magnetic stimulation (TMS) is a powerful technique to noninvasively activate neurons in the brain. However, the relationship between TMS-generated electric fields (E-fields) and specific cortical responses is not well understood. The goal of this study was to investigate the relationship between induced E-fields and neocortical activation measured by metabolic responses. Human subject-specific detailed finite element models (FEM) of the head were constructed to calculate the distribution of induced cortical E-field vectors. Positron emission tomography (PET) recordings were made during concurrent TMS application as a measure of cortical activation. A functional model of local circuit connections was developed to study the relationship between applied magnetic fields and neocortical activation and was fitted to experimental data. Sensitivity of interneurons to induced tangential E-fields was over twice as strong as pyramidal neuron sensitivity to induced normal E-fields which may help explain why cortical electrophysiological responses to TMS have specific sensitivities to coil orientation. Furthermore, this study produced an algorithm for predicting electrophysiological responses in human neocortex with high accuracy (>95%) that could provide an invaluable tool for planning of specific regional cortical activation critical in both research and clinical applications.
... Neuron reconstructions are obtained from cat visual cortex and cat somatosensory cortex (source: NeuroMorpho.org by Kisvárday and Eysel (1992); Mainen and Sejnowski (1996) Table 7). Each morphology M y is here shown in relation to the layer boundaries (horizontal lines). ...
Chronic and acute implants of multi-electrode arrays that cover several mm$^2$ of neural tissue provide simultaneous access to population signals like extracellular potentials and the spiking activity of 100 or more individual neurons. While the recorded data may uncover principles of brain function, its interpretation calls for multiscale computational models with corresponding spatial dimensions and signal predictions. Such models can facilitate the search of mechanisms underlying observed spatiotemporal activity patterns in cortex. Multi-layer spiking neuron network models of local cortical circuits covering ~1 mm$^2$ have been developed, integrating experimentally obtained neuron-type specific connectivity data and reproducing features of in-vivo spiking statistics. With forward models, local field potentials (LFPs) can be computed from the simulated spiking activity. To account for the spatial scale of common neural recordings, we extend a local network and LFP model to 4x4 mm$^2$. The upscaling preserves the neuron densities, and introduces distance-dependent connection probabilities and delays. As detailed experimental connectivity data is partially lacking, we address this uncertainty in model parameters by testing parameter combinations within biologically plausible bounds. Based on model predictions of spiking activity and LFPs, we find that the upscaling procedure preserves the overall spiking statistics of the original model and reproduces asynchronous irregular spiking across populations and weak pairwise spike-train correlations observed in sensory cortex. In contrast with the weak spike-train correlations, the correlation of LFP signals is strong and distance-dependent, compatible with experimental observations. Enhanced spatial coherence in the low-gamma band may explain the recent experimental report of an apparent band-pass filter effect in the spatial reach of the LFP.
... Le fait que le rayon d'action des neurones inhibiteurs soit plus large que celui des neurones excitateurs n'a cependant pas été mis en évidence in vivo. En fait, les deux seules études qui à notre connaissance examinent la question prouvent que ce rayon est sensiblement le même voir plus petit pour les neurones inhibiteurs (Kisvárday et al., 1997;Mariño et al., 2005 1. les neurones excitateurs forment effectivement de nombreuses connexions excitatrices avec des neurones excitateurs proches (Kisvárday & Eysel, 1992;Kisvárday et al., 1997;Yousef et al., 1999;Mariño et al., 2005), 2. les neurones excitateurs activent également des neurones inhibiteurs (Liu et al., 2013), 3. les neurones inhibiteurs inhibent les neurones excitateurs (Kisvárday et al., 1993), 4. il a aussi été montré que certains neurones inhibiteurs communiquent entre eux par des jonctions gap, ce type de connexions n'étant pas présente entre neurones excitateurs (Galarreta & Hestrin, 1999;Gibson et al., 1999;Fukuda et al., 2006). Les jonctions gap permettent de connecter électriquement un grand nombre de cellules, ce qui accélère la transmission de l'information par rapport à une connexion par synapses chimiques. ...
... The local cluster around the dendritic tree is typically the largest in extent and contains the most boutons ( Fig. 1h reiterates the concept of local versus distal). The distal clusters vary in number from cell to cell and for an individual cell, no two clusters are the same size or contain the same number of boutons (Kisvarday and Eysel 1992;Binzegger et al. 2004Binzegger et al. , 2005Binzegger et al. , 2007. When ranked according to the number of boutons in each cluster, the distribution of boutons across the clusters is seen to follow an exponential distribution (Binzegger et al. 2007). ...
Pyramidal cells in the superficial layers of neocortex of higher mammals form a lateral network of axon clusters known as the 'daisy' network. The role of these axon clusters remains speculative and we still lack a comprehensive quantitative description of the single neurons forming the daisy or their heterogeneity. We filled intracellularly 50 superficial layer pyramidal neurons in the cat primary visual cortex and reconstructed the axonal tree and their synaptic boutons in 3D. Individual bouton clusters were identified using an objective mean-shift algorithm. By parameterizing the morphology of these 50 axonal trees and the 217 bouton clusters they formed, we were able to extract one set of relatively constant parameters and another set of variable parameters. Both sets combined allowed us to outline a comprehensive biological blueprint of superficial layer pyramidal neurons. Overall, our detailed analysis supports the hypothesis that pyramidal neurons use their lateral clusters to combine differential contextual cues, required for context-dependent processing of natural scenes.
... In the mammalian visual cortex, boutons of many superficial layer (L2/3) pyramidal cells are known to terminate in distinct clusters forming patchy projection of the axons (Rockland and Lund 1982;Gilbert and Wiesel 1983;Rockland and Lund 1983;Martin and Whitteridge 1984;Kisvárday and Eysel 1992;Binzegger et al. 2007). L4 neurons also provide clustered projections although with different spatial constraints compared to their superficial layer counterparts (Karube and Kisvárday 2011). ...
To uncover the functional topography of layer 6 neurons, optical imaging was combined with three-dimensional neuronal reconstruction. Apical dendrite morphology of 23 neurons revealed three distinct types. Type Aa possessed a short apical dendrite with many oblique branches, Type Ab was characterized by a short and less branched apical dendrite, whereas Type B had a long apical dendrite with tufts in layer 2. Each type had a similar number of boutons, yet their spatial distribution differed from each other in both radial and horizontal extent. Boutons of Type Aa and Ab were almost restricted to the column of the parent soma with a laminar preference to layer 4 and 5/6, respectively. Only Type B contributed to long horizontal connections (up to 1.5 mm) mostly in deep layers. For all types, bouton distribution on orientation map showed an almost equal occurrence at iso- (52.6 ± 18.8 %) and non-iso-orientation (oblique, 27.7 ± 14.9 % and cross-orientation 19.7 ± 10.9 %) sites. Spatial convergence of axons of nearby layer 6 spiny neurons depended on soma separation of the parent cells, but only weakly on orientation preference, contrary to orientation dependence of converging axons of layer 4 spiny cells. The results show that layer 6 connections have only a weak dependence on orientation preference compared with those of layers 2/3 (Buzás et al., J Comp Neurol 499:861–881, 2006) and 4 (Karube and Kisvárday, Cereb Cortex 21:1443–1458, 2011).
Electronic supplementary material
The online version of this article (doi:10.1007/s00429-016-1284-z) contains supplementary material, which is available to authorized users.
... Of particular interest to us are horizontally projecting intrinsic axonal patches that arise from these cells and project to the basal dendritic trees of neighboring pyramidal cells Wiesel, 1979, 1983;Rockland andLund, 1982, 1983;Livingstone and Hubel, 1984;Martin and Whitteridge, 1984;Rockland, 1985;Kisvárday et al., 1986;McGuire et al., 1991;Fujita and Fujita, 1996). As many as 80-95% of the horizontal projection synapses of individual supragranular pyramidal cells are formed with other nearby supragranular pyramidal cells (Kisvárday et al., 1986;McGuire et al., 1991) in reciprocally connected patches (Kisvárday and Eysel, 1992). These reciprocal patches have been reported in many different cortical areas in the primate brain, and are believed to provide an anatomical basis for functional domains among the patches (Mitchison and Crick, 1982;Matsubara et al., 1985;T'so et al., 1986;T'so and Gilbert, 1988;Gilbert and Wiesel, 1989;Malach et al., 1993;Malonek et al., 1994). ...
... In the visual system, such reciprocal pathways exist between areas at the same or different hierarchical levels [21]. In addition neurons within each area are often densely and reciprocally interconnected [34]. What is the function of reciprocal or return projections in the cortex? ...
Recent advances in our understanding of the functional organization of the cerebral cortex pose the problem of how cortical neural activity is integrated. Even within one sensory domain, such as vision, there is a multitude of functionally segregated areas. However, visual perception always appears unitary and coherent. In order to achieve perceptual coherence as well as an integrated behavioral response to stimuli composed of many elementary features and attributes, these distributed neural signals must be put together. In this chapter, we present a set of computer simulations based on the anatomy and physiology of the cortex that address the problem of cortical integration. We propose that integration takes place at many different levels of organization: locally within populations of neurons, within a given cortical area (linking) and across several areas (binding). We discuss some stringent temporal constraints for neural integration and the possible effects of patterns of correlations on behavior.
... In the cat, the striate cortex exhibits a basic periodicity of structure and function at the scale of about 1 mm. This organization is manifested both in terms of the center-to-center distance of isoorientation bands (Kisvarday and Eysel 1993;Kisvarday et al. 1997;Lowel et al. 1987) and in terms of the average separation of neighboring axonal termination patches in supragranular layers (Kisvarday and Eysel 1992). To constrain our model, we assume that an area of roughly 1.0 mm 2 forms a basic macrounit in the striate cortical mosaic array. ...
... But the most pronounced connection pattern is formed by the horizontal projections which run within the layers 2/3 and mostly connect to other distant pyramidal neurons. The projections can extend up to several millimeters before they exhibit small axonal arbors visible as patches in stained slices (Kisvarday and Eysel, 1992). The major fraction of the axons seems to connect pyramidal neurons with similar physiological properties like orientation preference or ocular dominance (Gilbert and Wiesel, 1989). ...
... Key words: long-range intracortical connections; callosal connections; experience-dependent development; optical imaging; area 17; strabismus Long-range tangential axon collaterals are a prominent feature of cortical circuitry (Fisken et al., 1975). In the mammalian visual cortex, they interconnect regularly spaced clusters of cells [tree shrew (Rockland and L und, 1982), squirrel and macaque monkey (Rockland and L und, 1983;Livingstone and Hubel, 1984), cat (Gilbert and Wiesel, 1983;K isvárday and Eysel, 1992), and ferret (Rockland, 1985)], which share preferences for similar orientations or colors (Ts'o et al., 1986(Ts'o et al., , 1988Gilbert and Wiesel, 1989;Gray et al., 1989;Hata et al., 1991;Malach et al., 1993Malach et al., , 1994. It has been proposed that these connections: (1) contribute to the generation of large composite receptive fields (Singer and Tretter, 1976;Gilbert and Wiesel, 1985;Gilbert, 1990, Schwarz andBolz, 1991), (2) mediate inhibitory and subthreshold excitatory effects from beyond the classical receptive field (Blakemore and Tobin, 1972;Nelson and Frost, 1978;Morrone et al., 1982;Allman et al., 1985), (3) contribute to orientation and direction tuning (Eysel et al., 1987(Eysel et al., , 1990, (4) are responsible for adaptive changes of cortical maps after deafferentiation (Kaas et al., 1990;Heinen and Skavenski, 1991;Gilbert and Wiesel, 1992;Darian-Smith and Gilbert, 1994), and (5) synchronize the responses of spatially distributed neurons as a function of stimulus coherence (Gray et al., 1989;König et al., 1993;Singer, 1993). ...
The development of both long-range intracortical and interhemispheric connections depends on visual experience. Previous experiments showed that in strabismic but not in normal cats, clustered horizontal axon projections preferentially connect cell groups activated by the same eye. This indicates that there is selective stabilization of fibers between neurons exhibiting correlated activity. Extending these experiments, we investigated in strabismic cats: (1) whether tangential connections remain confined to columns of similar orientation preference within the subsystems of left and right eye domains; and (2) whether callosal connections also extend predominantly between neurons activated by the same eye and preferring similar orientations. To this end, we analyzed in strabismic cats the topographic relationships between orientation preference domains and both intrinsic and callosal connections of area 17. Red and green latex microspheres were injected into monocular iso-orientation domains identified by optical imaging of intrinsic signals. Additionally, domains sharing the ocular dominance and orientation preference of the neurons at the injection sites were visualized by 2-deoxyglucose (2-DG) autoradiography. Quantitative analysis revealed that 56% of the retrogradely labeled cells within the injected area 17 and 60% of the transcallosally labeled neurons were located in the 2-DG-labeled iso-orientation domains. This indicates: (1) that strabismus does not interfere with the tendency of long-range horizontal fibers to link predominantly neurons of similar orientation preference; and (2) that the selection mechanisms for the stabilization of callosal connections are similar to those that are responsible for the specification of the tangential intrinsic connections.
... Interestingly, the only layer 3 interneuron (of nine tested) in the cat to receive excitatory input from layer 4 had similar long, myelinated, horizontal axon collaterals. These long collaterals could therefore provide patchy inhibition (Kisvárday and Eysel, 1992) that was controlled by correlated activity in layers 3 and 4 and direct thalamo-cortical input. No equivalently long, patchy horizontal interneuronal projections were found in the rat in this study, but one such layer 4 cell was reported previously (Thomson et al., 1996). ...
Dual and triple intracellular recordings with biocytin labelling in slices of adult neocortex explored small circuits of synaptically connected neurons. 679 paired recordings in rat and 319 in cat yielded 135 and 42 excitatory postsynaptic potentials (EPSPs) and 37 and 26 inhibitory postsynaptic potentials (IPSPs), respectively. Patterns of connectivity and synaptic properties were similar in the two species, although differences of scale and in the range of morphologies were observed. Excitatory ‘forward’ projections from layer 4 to 3, like those from layer 3 to 5, targeted pyramidal cells and a small proportion of interneurons, while excitatory ‘back’ projections from layer 3 to 4 selected interneurons, including parvalbumin immuno-positive basket cells. Layer 4 interneurons that inhibited layer 3 pyramidal cells included both basket cells and dendrite-targeting cells. Large interneurons, resembling cells previously described as large basket cells, in layers 4 and 3 (cat), with long myelinated horizontal axon collaterals received frequent excitatory inputs from both layers. A very high rate of connectivity was observed between pairs of interneurons, often with quite different morphologies, and the resultant IPSPs, like the EPSPs recorded in interneurons, were brief compared with those recorded in pyramidal and spiny stellate cells.
Stimulus size modulation of neuronal firing activity is a fundamental property of the primary visual cortex. Numerous biological experiments have shown that stimulus size modulation is affected by multiple factors at different spatiotemporal scales, but the exact pathways and mechanisms remain incompletely understood. In this paper, we establish a large‐scale neuronal network model of primary visual cortex with layer 2/3 to study how gamma oscillation properties are modulated by stimulus size and especially how long‐range connections affect the modulation as realistic neuronal properties and spatial distributions of synaptic connections are considered. It is shown that long‐range horizontal synaptic connections are sufficient to produce dimensional modulation of firing rates and gamma oscillations. In particular, with increasing grating stimulus size, the firing rate increases and then decreases, the peak frequency of gamma oscillations decreases and the spectral power increases. These are consistent with biological experimental observations. Furthermore, we explain in detail how the number and spatial distribution of long‐range connections affect the size modulation of gamma oscillations by using the analysis of neuronal firing activity and synaptic current fluctuations. Our results provide a mechanism explanation for size modulation of gamma oscillations in the primary visual cortex and reveal the important and unique role played by long‐range connections, which contributes to a deeper understanding of the cognitive function of gamma oscillations in visual cortex.
In the mammalian visual system, stimulus context was investigated with respect to the ways it influenced neuronal mean response magnitude (the average number of spikes fired per second), response temporal structure (the timing of spikes with respect to one another), and the extent to which distributed neurones fired spikes synchronous due to synaptic interaction between them. Neurones were presented with bipartite grating stimuli, in which the spatio-temporal relationship between the grating activating the excitatory receptive field and that presented to the surrounding visual space could be varied systematically. Simultaneous extracellular recordings were made of the responses of up to four single neurones separated by 750-1000µm, in the lateral geniculate nucleus (LGN) of the thalamus in the cat, or the primary visual cortex (V1) of non-human primates or cats. Changing context systematically influenced the activity of groups of cells. The responses of 83% of primate V1 cells to discontinuous stimuli, in which the centre/surround orientation difference was greater than 45°, contained stronger oscillations at frequencies below 80Hz, than responses to continuous stimuli. Many cat and primate V1 neurones exhibited elevated response magnitudes to such stimuli. In primate V1, the strength of a cell's oscillatory discharge was dependent on stimulus configuration rather than response magnitude. In the LGN and V1, cell pairs with different orientation preferences fired synchronised responses when stimulated by specific discontinuous grating configurations. Stimulus specific synchronised LGN input, and reciprocal excitatory and inhibitory cortico-cortical connections could generate these properties of cells, and the network in which they exist. A model is proposed to account for the function significance of contour discontinuities in generating coherent neural representations of objects in the visual world. It involves response synchronisation in horizontal, feedforward and feedback interactions, within and between the LGN, V1, V2 and V4.
The brain performs visual object recognition using much shallower hierarchical stages than artificial deep neural networks employ. However, the mechanism underlying this cost-efficient function is elusive. Here, we show that cortical long-range connectivity(LRC) may enable this parsimonious organization of circuits for balancing cost and performance. Using model network simulations based on data in tree shrews, we found that sparse LRCs, when added to local connections, organize a small-world network that dramatically enhances object recognition of shallow feedforward networks. We found that optimization of the ratio between LRCs and local connections maximizes the small-worldness and task performance of the network, by minimizing the total length of wiring needed for integration of the global information. We also found that the effect of LRCs varies by network size, which explains the existence of species-specific LRCs in mammalian visual cortex of various sizes. Our results demonstrate a biological strategy to achieve cost-efficient brain circuits.
Highlights
Long-range connections (LRCs) enhance the object recognition of shallow networks
Sparse LRCs added to dense local connections organize a small-world type network
Small-worldness of networks modulates the balance between performance and wiring cost
Distinct LRCs in various species are due to the size-dependent effect of LRCs
Significance statement
The hierarchical depth of the visual pathway in the brain is constrained by biological factors, whereas artificial deep neural networks consist of super-deep structures (i.e., as deep as computational power allows). Here, we show that long-range horizontal connections (LRCs) observed in mammalian visual cortex may enable shallow biological networks to perform cognitive tasks that require deeper artificial structures, by implementing cost-efficient organization of circuitry. Using model simulations based on anatomical data, we found that sparse LRCs, when added to dense local circuits, organize “small-world” type networks and that this dramatically enhances image classification performance by integrating both local and global components of visual stimulus. Our findings show a biological strategy of brain circuitry to balance sensory performance and wiring cost in the networks.
One sentence summary
Cortical long-range connections organize a small-world type network to achieve cost-efficient functional circuits under biological constraints
Anatomy and function of long-range intrinsic and callosal axons in primary visual cortex are reviewed. In cats, both arborize in a patchy manner, in an orderly relationship to the visuotopic map and visual stimulus features. Patches tend to link neurons with similar contour and direction preference aligned along a collinear visual field axis. Direct investigation of callosal action on visual responses reveals a multiplicative shift without changing neuronal selectivity. Both gain and bias toward excitation or inhibition depend on global stimulus attributes. Interactions are more pronounced for neurons processing similar, in particular cardinal, visual features. As feature selectivity emerges already in ongoing neuronal activity, it is hypothesized that perceptual grouping is anticipated via the feature bias in patchy connections. By comparing data from lateral and feedback circuits, we conclude that visual callosal connections are more similar to intrinsic connections and can be interpreted as extending this circuit across the hemispheres.
Since Santiago Ramon y Cajal, neuroscientists have been fascinated by the shapes of dendritic arbors for more than 100 years. However, the principle underlying these complex and diverse structures remains elusive. Here we propose that evolution has tinkered with brain design to maximize its functionality while minimizing the cost associated with building and maintaining it. We hypothesize that the functionality of a neuron benefits from a larger repertoire of connectivity patterns between dendrites and surrounding axons, and the cost of a dendritic arbor increases with its total length and path length from synapses to soma. We solved this optimization problem by drawing an analogy with maximization of the entropy for a given energy in statistical physics. The solution predicts several scaling relationships between arbor dimensions and closely fits with experimental data. Moreover, our theory may explain why basal dendrites of pyramidal cells and Purkinje cells, the two major cell types in the mammalian brains, exhibit distinct morphologies.
Our understanding of the organization of the basal ganglia has advanced markedly over the last 10 years, mainly due to increased knowledge of their anatomical, neurochemical and physiological organization. These developments have led to a unifying model of the functional organization of the basal ganglia in both health and disease. The hypothesis is based on the so-called "direct" and "indirect" pathways of the flow of cortical information through the basal ganglia and has profoundly influenced the field of basal ganglia research, providing a framework ibr anatomical, physiological and clinical studies. The recent introduction of powerful techniques for the analysis of neuronal networks has led to further developments in our understanding of the basal ganglia. The objective of this commentary is to build upon the established model of the basal ganglia connectivity and review new anatomical findings that lead to the refinement of some aspects of the model. Four issues will be discussed. (1) The existence of several routes for the flow of cortical information along "indirect" pathways. (2) The synaptic convergence of information flowing through the "direct" and "indirect" pathways at the single-cell level in the basal ganglia output structures. (3) The convergence of functionally diverse information from the globus pallidus and the ventral pallidum at different levels of the basal ganglia. (4) The interconnections between the two divisions of the pallidal complex and the subthalamic nucleus and the characterization of the neuronal network underlying the indirect pathways. The findings summarized in this commentary confirm and elaborate the models of the direct and indirect pathways of information flow through the basal ganglia and provide a morphological framework for future studies. (C) 1998 IBRO. Published by Elsevier Science Ltd.
In monkey prefrontal cortex, the intrinsic axon collaterals of supragranular pyramidal neurons extend horizontally for considerable distances through the gray matter and give rise to stripe-like clusters of axon terminals (Levitt et al. [1993] J. Comp. Neurol. 338:360–376). Because understanding the functional role of these connections requires knowledge of their synaptic targets, we made injections of biotinylated dextran amine (BDA) into layer 3 of macaque prefrontal area 9 and examined the labeled intrinsic axon collaterals by electron microscopy. Labeled axon terminals formed exclusively asymmetric synapses, and 95.6% of the postsynaptic structures were dendritic spines, presumably belonging to other pyramidal neurons. The remaining postsynaptic structures were dendritic shafts, many of which had the morphological characteristics of local circuit neurons. The prefrontal injections also labeled associational projections that traveled through the white matter to terminate in other areas of prefrontal cortex. All of the synapses formed by these associational axons were asymmetric, and 91.9% were onto dendritic spines. The similarities in synaptic targets of the prefrontal intrinsic and associational axon terminals suggested that these projections might arise from the same neurons, an interpretation confirmed in dual label, retrograde tracing studies.
To determine the specificity of the synaptic targets of these prefrontal connections, two additional comparisons were made. In the posterior parietal cortex (area 7a), 94.2% of the synapses furnished by BDA-labeled intrinsic collaterals of supragranular pyramidal neurons were also with dendritic spines. In contrast, only 75.6% of unlabeled asymmetric synapses in the prefrontal cortex were onto dendritic spines. These comparisons suggest that the axons of supragranular pyramidal neurons in primate association cortices are preferentially directed to specific targets.
Finally, after injections of BDA, a small number of retrogradely labeled pyramidal neurons were observed within the anterogradely labeled clusters of intrinsic axon terminals. At the ultrastructural level, synapses between anterogradely labeled axon terminals and retrogradely labeled dendritic spines were identified. These findings suggest that reciprocal, monosynaptic connections may exist between pyramidal neurons located in different stripe-like clusters, providing a potential anatomical substrate for reverberating excitatory circuits within the primate association cortices. J. Comp. Neurol. 390:211–224, 1998.
“Intelligent activity may reasonably be regarded as the key note of mammalian progress” (Romer 1962). This progress became possible with the acquisition of a neocortex, with its great analytic, associative and synthetic potential. Other mammalian characteristics, such as improvements in the circulation and in temperature regulation and, in most mammals, the long gestation period, giving birth to live young, and the development of nursing, with concomitant care and training of the young (Romer 1962), are conditional for the development, imprinting and functioning of a complicated brain. The main characteristics of the mammalian brain all are dependent on the presence of a neocortex.
Intracortical lateral connections have initially been shown in anatomical studies by degeneration methods in monkey (Fisken et al. 1975) and cat (Creutzfeldt et al. 1977). However, only the use of modern tracer techniques has disclosed details like the periodic spatial pattern of such intracortical connections (Rockland and Lund 1982, 1983; Gilbert and Wiesel 1983; Gilbert 1985) or the axonal and dendritic distribution of single excitatory or inhibitory cells (Gilbert and Wiesel 1979; Somogyi et al. 1983; Kisvárday et al. 1985). Intracortical excitation can be laterally transmitted by a long-range axonal network of excitatory connections between pyramidal cells spanning up to about 5 mm in the adult cat visual cortex (Kisvárday and Eysel 1991; see also the contribution of Kisvárday in this volume). Intracortical inhibition can be mediated by GABAergic cells which comprise on the average 20% of the cortical neurons in rat, cat and monkey (Ribak 1978; Hendrickson et al. 1981; Gabbott and Somogyi 1986; Fitz Patrick et al. 1987; Hendry et al. 1987). These nonpyramidal cells are a heterogenous group which show a variety of morphological specializations (Somogyi 1986). Some of the GABAergic cells give rise to locally restricted axonal systems (Kisvárday et al. 1985), others send axons over distances from 0.5 to 2 mm (Somogyi et al. 1983; Matsubara et al. 1987b).
Extrastriate visual cortex consists of multiple areas. As reviewed elsewhere (Kaas, 1989; Colby and Duhamel, 1991; and several chapters in this volume), there are still many questions concerning specific boundaries and subdivisions, and the criteria for area identification themselves remain under discussion. How areas interact is even less well known and is very much a topic of active research. Is there an overall architecture? Are there patterns of sequential or synchronous coactivation?
Stimulus-dependent synchronous oscillatory firing of populations of cortical neurons at frequencies around 40 Hz has been observed within and between many different cortical areas in primates and cats both awake and anaesthetized (for review see Singer, 1993). It has been proposed that this synchronous activity is used to group separated parts of single objects (Engel et al., 1991a; Sporns et al., 1991) and even signal visual awareness (Crick and Koch, 1990). Although there is still much debate as to the role and significance of synchrony, it is generally agreed that many cortical neurons can fire synchronously under some conditions. Therefore, it is worthwhile to discover exactly how cortical tissue generates and sustains synchronized oscillatory firing of its component neurons, both within a single column and between columns that may be located in different hemispheres (Engel et al., 1991b).
The intrinsic circuitry of the motor cortex comprises a complex network of connections whose synaptic relationships are poorly understood. This study was designed to determine the characteristics of subsets of GABAergic neurons containing the calcium-binding proteins parvalbumin (PV) and calbindin (CB), and their relationships with intrinsic axons in motor cortex. Immunohistochemically identified PV-containing neuronal profiles were more evenly distributed across cortical laminae (38% in II-III, 32% in V, 30% in VI) and more numerous (2.1/1) than CB-containing neuronal profiles (71% in II-III, 17% in V, 12% in VI). Relationships between neurons and axons intrinsic to motor cortex were visualized with fluorescent markers using the laser scanning confocal microscope. Similar percentages of PV (43%) and CBimmunoreactive (IR) (40%) neurons formed sparsely distributed appositions (1-5/neuron) with anterogradely labeled axons. The mean distances of such appositions from the somata were significantly different f...
This chapter focuses on the basic mechanisms that contribute to the structure of receptive fields (RFs) and response properties of cells in the striate cortex. Excitatory amino acids (EAA) are the main excitatory transmitters in the striate cortex. Application of their agonists increases the activity of the neurons, and application of the antagonists leads to a decrease in activity. When the glutamatergic receptors of striate cortex cells are pharmacologically stimulated by glutamate microiontophoresis, the vast majority of cells show excitatory responses accompanied by enhanced visual responses. Neurons in the striate cortex have RFs of different subfield structure and size. Simple cells have smaller RFs with adjacent bands of ON and OFF subfields. Complex cells are characterized by larger RFs at a given eccentricity in the visual field and overlapping, spatially nonseparable regions with ON and OFF responses, and hypercomplex cells (simple or complex-like) are defined by the presence of additional end-inhibitory zones. These spatial characteristics of the striate cortex RFs appear in part generated, or at least significantly sculptured, by the balance between excitatory and inhibitory synaptic inputs.
This chapter summarizes the structural and functional organization of the lateral geniculate nucleus (LGN) and the relationships of its subcomponent structures to the signal streams arriving from retina and transmitted to primary visual cortex. In gross structure, the LGN has a modified sigmoid shape with an upturned tail posteriorly, an almost horizontal body in the mid-section, and a down-turned head, anteriorly. The three dorsal laminae contain neurons ranging in size from small to large. Because of the presence of large neurons, these laminae are termed the magnocellular layers. Medial from the laminated portion of LGN is a highly compact structure, the medial interlaminar nucleus (MIN). It too is considered part of LGN. The MIN occupies about half of the medial surface of the LGN and is concentrated anteriorly in the nucleus. Retinal fibers approach from the ventral and lateral aspect in the optic tract (OT) and penetrate through the ventral surface of the nucleus to terminate in the appropriate layer according to their eye of origin and position of parent cell body in the retina. Each of the components of LGN receives a unique and signature set of fiber types from retina.
Publisher Summary
This chapter discusses functional topography of horizontal neuronal networks in cat visual cortex. The topographical relationship between orientation selectivity and lateral connectivity is examined. At each penetration, orientation selectivity of a small cluster of cells was determined for the dominant eye using computer controlled stimuli. The location of penetrations was marked on an enlarged photograph of the exposed area for which surface blood vessels were used as landmarks. Findings provide direct evidence that in the visual cortex of the cat individual large basket cells could mediate lateral inhibition to virtually all orientations. In addition to iso-orientation inhibition, there is functional topographical evidence for non-iso-orientation inhibition in the strictest sense. Excitatory boutons were often connected through a short stalk to the axon stem, an uncommon feature of basket boutons. On the other hand, basket axons frequently established pericellular nests around the target neurons, a feature that has never been seen for pyramidal and spiny stellate cells. These features were taken into account for all boutons together with the global pattern of their parent axons in deciding which of the two groups, the putative excitatory or the putative inhibitory, they belong to. Lateral connections in primate visual areas seem to allow for considerable amount of crosstalk between specific compartments of orientation selectivity.
A prominent and stereotypical feature of cortical circuitry in the striate cortex is a plexus of long-range horizontal connections, running for 6-8 mm parallel to the cortical surface, which has a clustered distribution. This is seen for both intrinsic cortical connections within a particular cortical area and the convergent and divergent connections running between area 17 and other cortical areas. To determine if these connections are related to the columnar functional architecture of cortex, we combined labeling of the horizontal connections by retrograde transport of rhodamine-filled latex microspheres (beads) and labeling of the orientation columns by 2-deoxyglucose autoradiography. We first mapped the distribution of orientation columns in a small region of area 17 or 18, then made a small injection of beads into the center of an orientation column of defined specificity, and after allowing for retrograde transport, labeled vertical orientation columns with the 2-deoxyglucose technique. The retrogradely labeled cells were confined to regions of orientation specificity similar to that of the injection site, indicating that the horizontal connections run between columns of similar orientation specificity. This relationship was demonstrated for both the intrinsic horizontal and corticocortical connections. The extent of the horizontal connections, which allows single cells to integrate information over larger parts of the visual field than that covered by their receptive fields, and the functional specificity of the connections, suggests possible roles for these connections in visual processing.
The intrinsic connections of the cortex have long been known to run vertically, across the cortical layers. In the present study we have found that individual neurons in the cat primary visual cortex can communicate over suprisingly long distances horizontally (up to 4 mm), in directions parallel to the cortical surface. For all of the cells having widespread projections, the collaterals within their axonal fields were distributed in repeating clusters, with an average periodicity of 1 mm. This pattern of extensive clustered projections has been revealed by combining the techniques of intracellular recording and injection of horseradish peroxidase with three-dimensional computer graphic reconstructions. The clustering pattern was most apparent when the cells were rotated to present a view parallel to the cortical surface. The pattern was observed in more than half of the pyramidal and spiny stellate cells in the cortex and was seen in all cortical layers. In our sample, cells made distant connections within their own layer and/or within another layer. The axon of one cell had clusters covering the same area in two layers, and the clusters in the deeper layer were located under those in the upper layer, suggesting a relationship between the clustering phenomenon and columnar cortical architecture. Some pyramidal cells did not project into the white matter, forming intrinsic connections exclusively. Finally, the axonal fields of all our injected cells were asymmetric, extending for greater distances along one cortical axis than along the orthogonal axis. The axons appeared to cover areas of cortex representing a larger part of the visual field than that covered by the excitatory portion of the cell's own receptive field. These connections may be used to generate larger receptive fields or to produce the inhibitory flanks in other cells' receptive fields.
An experimental neurohistological study has been made of the intrinsic connections of the cortex of area 17 of the monkey, of the commissural connections of the visual cortex of the cat and monkey and of the association fibres passing into area 17 of the cat. In light microscopic studies the axonal degeneration method of Nauta has been used, and the site and mode of termination of the degenerating fibres has also been determined with the electron microscope...
The neuronal structure and connectivity underlying receptive field organisation of cells in the cat visual cortex have been investigated. Intracellular recordings were made using a micropipette filled with a histochemical marker, which was injected into the cells after their receptive fields had been characterised. This allowed visualisation of the dendritic and axonal arborisations of functionally identified neurones.
The extraction of stereo-disparity information from two images depends upon establishing a correspondence between them. In
this article we analyze the nature of the correspondence computation and derive a cooperative algorithm that implements it.
We show that this algorithm successfully extracts information from random-dot stereograms, and its implications for the psychophysics
and neurophysiology of the visual system are briefly discussed.
After making a map of response properties of neurons in a roughly 3 X 4 mm region of area 18 in the cat, we injected wheat-germ agglutinin horseradish peroxidase (WGA-HRP) and succinylated concanavalin A (Con A) into physiologically identified regions of the map. We observed up to 10 patches of retrogradely labeled cells surrounding each injection site. The majority of the patches occurred within 1.4 mm of the center of the injection site, but rare patches were found as far as 3.4 mm from the injection site. The mean center-to-center spacing of the intrinsic patches was about 1 mm, while the mean distance between the center of the injection site and the nearest patches was less than 1 mm. The labeled cells included both nonpyramidal and pyramidal types and were found in all layers, although they were usually most dense in layers II-IV. Between 2% and 9% of the cells within a cortical column were labeled after a single injection of WGA-HRP or Con A into area 18. Injections of different tracers into 2 neighboring areas resulted in a uniform and less patchy distribution of labeled cells, which suggests that the patches observed after a single injection were only a portion of a continuous horizontal system of interconnections. The patterns and positions of the intrinsic patches were compared to the distribution of the following receptive-field properties: preferred orientation, receptive-field location, and eye preference. The preferred orientations of the recording sites within the injected and labeled areas were different and, most frequently, orthogonal to each other. This is a highly specific projection, since regions with orientation values like those of the injection site were "within range," yet not labeled. We were unable to detect any relationship between the ocular preferences of the injected and labeled cell regions. Injections into areas predominantly driven by the contralateral eye resulted in labeled regions exhibiting varied eye preference distributions. In some animals they were like the injection site and in others there were equal numbers of contra- and ipsilateral eye-dominated regions. The overall distribution of the patches around the injection site was elongated along the anterior-posterior cortical axis of the brain. The patches extended further in the posterior than the anterior direction. These observations appear to be related to the finding that the cortical magnification factor is greater along the anterior-posterior than the medial-lateral axis of area 18.(ABSTRACT TRUNCATED AT 400 WORDS)
A fundamental step in visual pattern recognition is the establishment of relations between spatially separate features. Recently, we have shown that neurons in the cat visual cortex have oscillatory responses in the range 40-60 Hz (refs 1, 2) which occur in synchrony for cells in a functional column and are tightly correlated with a local oscillatory field potential. This led us to hypothesize that the synchronization of oscillatory responses of spatially distributed, feature selective cells might be a way to establish relations between features in different parts of the visual field. In support of this hypothesis, we demonstrate here that neurons in spatially separate columns can synchronize their oscillatory responses. The synchronization has, on average, no phase difference, depends on the spatial separation and the orientation preference of the cells and is influenced by global stimulus properties.
Cortical neurons using the inhibitory neurotransmitter gamma-aminobutyric acid (GABA) are known to contribute to the formation of neuronal receptive field properties in the primary visual cortex (area 17) of the cat. In order to determine the cortical location of GABA containing neurons and what proportion of cortical neurons might use GABA as their transmitter, we analysed their distribution quantitatively using a post-embedding GABA immunohistochemical method on semithin sections in conjunction with stereological procedures. The mean total numerical density of neurons in the medial bank of the lateral gyrus (area 17) of five adult cats was 54,210 +/- 634 per mm3 (mean +/- SD). An average of 20.60 +/- 0.48% (mean +/- SEM) of the neurons were immunoreactive for GABA. The density of GABA-immunoreactive neurons was somewhat higher in layers II, III and upper VI, compared with layers I, IV, V and lower VI, with the lowest density being in layer V. The proportion of GABA-immunopositive cells relative to immunonegative neurons gradually decreased from the pia to the white matter. Layer I was different from other layers in that approximately 95% of its neurons were GABA-immunoreactive. The results allowed the calculation of the absolute numbers of GABAergic neurons in each layer under a given cortical surface area and could provide the basis for the quantitative treatment of cortical circuits.
Primary visual coding can be characterized by the receptive field (RF) properties of single neurons. Subject of this paper is our search for a global, second coding step beyond the RF-concept that links related features in a visual scene. In recent models of visual coding, oscillatory activities have been proposed to constitute such linking signals. We tested the neurophysiological relevance of this hypothesis for the visual system. Single and multiple spikes as well as local field potentials were recorded simultaneously from several locations in the primary visual cortex (A17 and A18) using 7 or 19 individually advanceable fiber-microelectrodes (250 or 330 microns apart). Stimulus-evoked (SE)-resonances of 35-85 Hz were found in these three types of signals throughout the visual cortex when the primary coding channels were activated by their specific stimuli. Stimulus position, orientation, movement direction and velocity, ocularity and stationary flicker caused specific SE-resonances. Coherent SE-resonances were found at distant cortical positions when at least one of the primary coding properties was similar. Coherence was found 1) within a vertical cortex column, 2) between neighbouring hypercolumns, and 3) between two different cortical areas. We assume that the coherence of SE-resonances is mediated by recurrent excitatory intra- and inter-areal connections via phase locking between assemblies that represent the linking features of the actual visual scene. Visually related activities are, thus, transiently labelled by a temporal code that signalizes their momentary association.
Anatomical studies in the visual cortex have shown the presence of long-range horizontal connections with clustered axonal collaterals, suggesting interactions over distances of several millimeters. We used cross-correlation analysis in cat striate cortex to detect interactions between cells over comparable distances. Using one cell as a reference, we recorded from other cells with a second electrode at varying distances and looked for correlated firing between the two recording sites. This technique allowed us to combine a physiological measure of the strength and type of connection between cells with a characterization of their receptive field properties. The observed interactions were excitatory, and extended over horizontal distances of several millimeters. Furthermore, the interactions were between orientation columns of like specificity, resulting in a waxing and waning in the strength of interaction as the electrodes passed through different orientation columns. We studied relationships between strength of correlation and other receptive field properties and found a tendency for facilitatory interactions between cells sharing the same eye preference. A large proportion of our correlations was due to common input. This feature, and the similarity of interactions between cells in the same column with the reference cell, suggest a high degree of interconnectivity between and within the columns. As the distance between the two electrodes increased, the overlap of the receptive fields of the cells participating in the interactions gradually diminished. At the furthest distances recorded, the cell pairs had nonoverlapping receptive fields separated by several degrees. The distribution and range of these interactions corresponded to the clustering and extent of the horizontal connections observed anatomically.
Intracortical injections of horseradish peroxidase conjugated with wheat-germ agglutinin (WGA-HRP) reveal a characteristic patchy staining pattern within the superficial layers of cat striate cortex. The patches consist of a dense accumulation of labeled neurons and axonal arborizations. We have investigated the tangential organization and the development of these intrinsic cortical connections by using a flat-mount preparation of area 17. The diameter of the patches varied from 200 to 400 micron, the center-to-center distance ranged from 400 to 800 micron, and the spread of patches extended further in the anterior-posterior than in the medial-lateral direction. The expression of these horizontal patchy connections is age- and experience-dependent. From ten days to six weeks of age patches are exuberant and on occasion fuse to beaded bands extending radially from the injection site. From 6 weeks onwards the number and the tangential spread of the patches decreases to one or two rows of isolated clusters. Long-term binocular deprivation disrupts this pattern of intrinsic connections nearly completely. We infer from these results that there is an inborn pattern of discrete horizontal connections in striate cortex which is shaped by visual experience and requires contour vision for its maintenance.
There are numerous hypotheses for the role of the axon collaterals of pyramidal cells. Most hypotheses predict that pyramidal cells activate specific classes of postsynaptic cells. We have studied the postsynaptic targets of two layer III pyramidal cells, that were of special interest because of their clumped axon arborization near, and also 0.4-1.0 mm from the cell body, in register in both layers III and V. 191 terminations from four sites (layers III and V, both in the column of the cell and in distant clumps) were analysed by electron microscopy. Only one bouton contacted a cell body and that was immunoreactive for GABA. The major targets were dendritic spines (84 and 87%), and the remainder were dendritic shafts. Of these 13 were classed as pyramidal-like (P), 8 smooth cell-like (S) and three could not be classified. Four of five S types, but none of the seven P types tested were immunoreactive for GABA, supporting the fine structural classification. The putative inhibitory cells therefore formed not more than 5% of the postsynaptic targets, and their activation could only take place through the convergence of pyramidal cells onto a select population of GABA cells. The results show that the type of pyramidal cells with clumped axons studied here make contacts predominantly with other pyramidal cells. Thus the primary role of both the intra and intercolumnar collateral systems is the activation of other excitatory cells.
We perceive the visual world as a unitary whole, yet one of the guiding principles of nearly a half century of neurophysiological research since the early recordings by Hartline (1938) has been that the visual system consists of neurons that are driven by stimulation within small discrete portions of the total visual field. These classical receptive fields (CRFs) have been mapped with the excitatory responses evoked by a flashed or moving stimulus, usually a spot or bar of light. Most of the visual neurons, in turn, are organized in a series of maps of the visual field, at least 10 of which exist in the visual cortex in primates as well as additional topographic representations in the lateral geniculate body, pulvinar and optic tectum (Allman 1977, Newsome & Allman 1980, Allman & Kaas 1984). It has been widely assumed that perceptual functions that require the integration of inputs over large portions of the visual field must be either collective properties of arrays of neurons representing the visual field, or features of those neurons at the highest processing levels in the visual system, such as the cells in inferotemporal or posterior parietal cortex that typically possess very large receptive fields and do not appear to be organized in visuotopic maps. These assumptions have been based on the results of the many studies in which receptive fields were mapped with conventional stimuli, presented one at a time, against a featureless background. However, unlike the neurophysiologist's tangent screen, the natural visual scene is rich in features, and there is a growing body of evidence that in many visual neurons stimuli presented outside the CRF strongly and selectively influence neural responses to stimuli presented within the CRF. These results suggest obvious mechanisms for local-global comparisons within visuotopically organized structures. Such broad and specific surround mechanisms could participate in many functions that require the integration of inputs over wide regions of the visual space such as the perceptual constancies, the segregation of figure from ground, and depth perception through motion parallax. In the first section of this paper, we trace the historical development of the evidence of response selectivity for visual stimuli presented beyond the CRF; in the second, examine the anatomical pathways that sub serve these far-reaching surround mechanisms; and in the third, explore the possible relationships between these mechanisms and perception.
The main factors limiting the performance of the peripheral parts of the visual system can be specified, and doing this clarifies the nature of the interpretive tasks that must be performed by the central parts of the system. It is argued that the critical factor that hinders development of better resolving power is the difficulty of confining light within the waveguide-like outer segment, and that for sensitivity this critical factor is the thermal decomposition of photosensitive pigments. Knowledge of these limits makes many surprising details of the eye intelligible. Understanding the difficulties posed by the narrow dynamic range of nerve fibres may give similar insight into the coding of the retinal image for transmission to the brain. Our level of understanding changes when we come to the visual cortex, for although we do not lack good anatomical and neurophysiological data, these do not make the principles of operation self-evident in the way that the structure of the eye immediately suggests that it is an image-forming device. The cortex converts the representation of the visual field that it receives into reliable knowledge of the world around us, and the trouble may be that we lack good models of how this can be done. A system that can respond to single quanta and resolve almost to the diffraction limit is unlikely to employ grossly inefficient methods for those higher functions upon which its whole utility depends, and so it is worth seeking out the limiting factors. The quality of human performance at certain higher perceptual tasks is high compared with the limit of reliable statistical inference; hence much of the sample of information available in a visual image must be effectively utilized. But there are strong limitations on the connectivity in the cortex, so that one is forced to consider how the relevant information can be collected together. Three stages of dealing with the visual image are proposed: the improvement of the cortical map in primary visual cortex by processes analogous to spatial and temporal interpolation; the detection of linking features in this map; and the concentration of this information by non-topographical mapping in adjacent visual areas.
An attempt is made to bring earlier circuit models of primary sensory cortical areas into better line with recent observations on (1) the distribution of excitatory feedback connexions in cortical tissue volume, (2) putative inhibitory interneurons and the distribution of inhibition in well defined space modules, and (3) the direct (monosynaptic) cortical target cells of the specific sensory afferents, and the modes of relay to secondary neurons. Even though the concept of cortical circuitry on larger 'integrative units' containing smaller modules (or fields) of specific (excitatory and inhibitory) neuronal actions, proposed in 1967 (ref. 34) and 1969 (ref. 35), had gross deficiencies in the light of newly emerging data, the basic idea of how to look at the functional organization of the cortical neuron network may still be useful as a conceptual framework for the functional interpretation of structural data.
Three-dimensional reconstructions of the orientation column system were obtained from the visual cortex of four cats using the deoxyglucose technique. One cat had normal visual experience, one was monocularly deprived and two had selective experience with vertical and horizontal contours, respectively. In areas 17 and 18 orientation columns form a remarkably regular system of equally spaced parallel bands whose trajectory is orthogonal to the borderline between areas 17 and 18. This topographic organization is resistant to manipulations of early visual experience.
The patterns of arborisation of apical dendrites of different varieties of pyramidal neurons in area 17 differ and are characteristic for each cell type. They appear to serve as a means of collating within one neuron information derived directly from several different laminae. These different patterns of apical dendrite arborisation provide dendritic links which relate closely to the laminar distribution of axons of the spiny stellate neurons as well as the pyramidal neurons themselves. The axons of spiny stellate neurons lying in laminae IVCβ and IVA (Lund, '73)—Which receive information from parvocellular geniculate layers — project heavily to the lower half of lamina III (IIIB) and to a narrow zone at the top of lamina V (VA); laminae IIIB and VA are in turn linked by a specific variety of pyramidal neuron, with basal dendritic field in lamina VI, whose apical dendrite has marked lateral branching only in laminae VA and IIIB (where it terminates). Pyramidal neurons with basal dendritic field in laminae VA (with vestigial apical dendrite) or in IIIB have recurrent axon projections to lamina IIIA and above (the descending axon projection of lamina IIIB pyramids is principally to lamina VA itself). The pyramidal neurons of laminae IIIA and above have axons which distribute in the same upper laminae as their dendtritic fields and a descending axon projection to lamina VB. Pyramidal neurons with basal dendritic field on lamina VB have an apical dendrite which, if not vestigal, arborises in IIIA or above; their axons in some cases project to the superior colliculus or may be exclusively, or in addition, recurrent, distributing collaterals within laminae VB, VI and in IIIA or above; one variety of pyramidal neuron with basal dentritic field in lamina VI makes a dentritic link with these same regions, its apical dendrite arborising first within lamina VB and then in lamina IIIA and above.
Axons of spiny stellate neurons of lamina IVCα (which receives the projection of the magnocellular layers of the lateral geniculate nucleus) as well as distributing widely within lamina IVCα also contribute to laminae IVB and VA; a link is again made by a specific variety of pyramidal neuron, with basal dendtritic field in lamina VI, which shows branching to its apical dendtrite only in laminae VA and as a terminal arborisation in IVCα. Another variety of pyramidal neuron with basal dendtric field in lamina VI has apical dendritic arborisation only in lamina IVB. The pyramidal neurons with basal dendritic field in lamina IVB and apical dendrite arborising in lamina IIIB and above, also contribute axonal collatetrals to lamina IIIA and above; their horizontal axon collaterals, together with the axons of spiny stellate neurons of laminae IVCα and IVB, form the horizontal fiber band of lamina IVB (to which the axons of laminae III and II pyramidal neurons do not contribute. The descending axon projection of the spiny stellate and pyramidal neurons of lamina IVB appears to be principally to lamina VI.
The pattern of branching of pyramidal neuron apical dendrites is therefore neither random nor a continuum of one basic pattern; instead it is a series of separate patterns, each spatially distributed in a highly specific and unique fashion relating to the patterns of projection of afferent information through the cortex.
We developed a flat-mount technique in order to visualize, without additional reconstruction, the system of orientation columns in the cat visual cortex by using 2-deoxyglucose-autoradiography.
Experimental animals were injected with 2-deoxyglucose and then stimulated for 45–60 minutes either with vertical or horizontal or oblique gratings alone or with vertical and horizontal gratings presented in alternation.
In both areas 17 and 18 stimulation with either vertical or horizontal or oblique stripes produced similar and highly ordered patterns of parallel bands of increased 2-deoxyglucose uptake that were perpendicular to the boundaries of the areas. In area 17 they occasionally extended without interruption from the 17/18 border on the top of the lateral gyrus to the monocular segment in the splenial sulcus. Superposition of serial sections revealed that these bands were present in all cortical layers and in precise register along lines orthogonal to the lamination. The center-to-center spacing of the bands was 1.0–1.1 mm in area 17 and 1.2–1.4 mm in area 18. Stimulation with alternating vertical and horizontal contours led to a pattern the general organization of which resembled that induced by a single orientation but the spacing of which was reduced by a factor of 0.5. This strongly supports the concept that orientation is mapped in a system of parallel bands and argues against a recently formulated hypothesis that iso-orientation bands extend like spokes from centers that lack orientation selectivity (Braitenberg and Braitenberg, Biol. Cybern, 33:179–186, '79).
Another characteristic feature, revealed by the flat-mount technique, was a periodic variation of 2-deoxyglucose uptake along the bands that gave them a beaded appearance. The mean center-to-center distance between adjacent beads on the same band was in the range of 0.9–1.2 mm and remained unchanged when horizontal and vertical gratings were presented in alternation. We propose that these beads reflect another columnar system whose features have yet to be determined.
Of the many possible functions of the macaque monkey primary visual cortex (striate cortex, area 17) two are now fairly well understood. First, the incoming information from the lateral geniculate bodies is rearranged so that most cells in the striate cortex respond to specifically oriented line segments, and, second, information originating from the two eyes converges upon single cells. The rearrangement and convergence do not take place immediately, however: in layer IVc, where the bulk of the afferents terminate, virtually all cells have fields with circular symmetry and are strictly monocular, driven from the left eye or from the right, but not both; at subsequent stages, in layers above and below IVc, most cells show orientation specificity, and about half are binocular. In a binocular cell the receptive fields in the two eyes are on corresponding regions in the two retinas and are identical in structure, but one eye is usually more effective than the other in influencing the cell; all shades of ocular dominance are seen. These two functions are strongly reflected in the architecture of the cortex, in that cells with common physiological properties are grouped together in vertically organized systems of columns. In an ocular dominance column all cells respond preferentially to the same eye. By four independent anatomical methods it has been shown that these columns have the from of vertically disposed alternating left-eye and right-eye slabs, which in horizontal section form alternating stripes about 400 mu m thick, with occasional bifurcations and blind endings. Cells of like orientation specificity are known from physiological recordings to be similarly grouped in much narrower vertical sheeet-like aggregations, stacked in orderly sequences so that on traversing the cortex tangentially one normally encounters a succession of small shifts in orientation, clockwise or counterclockwise; a 1 mm traverse is usually accompanied by one or several full rotations through 180 degrees, broken at times by reversals in direction of rotation and occasionally by large abrupt shifts. A full complement of columns, of either type, left-plus-right eye or a complete 180 degrees sequence, is termed a hypercolumn. Columns (and hence hypercolumns) have roughly the same width throughout the binocular part of the cortex. The two independent systems of hypercolumns are engrafted upon the well known topographic representation of the visual field. The receptive fields mapped in a vertical penetration through cortex show a scatter in position roughly equal to the average size of the fields themselves, and the area thus covered, the aggregate receptive field, increases with distance from the fovea. A parallel increase is seen in reciprocal magnification (the number of degrees of visual field corresponding to 1 mm of cortex). Over most or all of the striate cortex a movement of 1-2 mm, traversing several hypercolumns, is accompanied by a movement through the visual field about equal in size to the local aggregate receptive field. Thus any 1-2 mm block of cortex contains roughly the machinery needed to subserve an aggregate receptive field. In the cortex the fall-off in detail with which the visual field is analysed, as one moves out from the foveal area, is accompanied not by a reduction in thickness of layers, as is found in the retina, but by a reduction in the area of cortex (and hence the number of columnar units) devoted to a given amount of visual field: unlike the retina, the striate cortex is virtually uniform morphologically but varies in magnification. In most respects the above description fits the newborn monkey just as well as the adult, suggesting that area 17 is largely genetically programmed. The ocular dominance columns, however, are not fully developed at birth, since the geniculate terminals belonging to one eye occupy layer IVc throughout its length, segregating out into separate columns only after about the first 6 weeks, whether or not the animal has visual experience. If one eye is sutured closed during this early period the columns belonging to that eye become shrunken and their companions correspondingly expanded. This would seem to be at least in part the result of interference with normal maturation, though sprouting and retraction of axon terminals are not excluded.
The current source density (CSD) method in its one-dimensional approximation is used to analyze the field potentials in visual areas 18 and 17 of the cat, which were elicited by stimulating electrodes in the optic chiasm (OX), the optic radiation (OR) or in the respective cortical area itself. The CSD analysis reveals the basic pattern of excitatory postsynaptic activity.1.
In both visual areas the basic specific excitatory activity flows along three different intracortical pathways, all starting in layer IV: The first pathway relays activity from layer IV to supragranular pyramidal cells via strong, local connections to layer III and from there through long-distance connections to layer II. The second pathway conveys activity from layer IV to layer V, where it mainly contacts apical dendrites of layer VI pyramidal cells. This infragranular polysynaptic activity is not clearly resolvable into separate components, suggesting that it is conveyed by various groups of axons, among them long-distance horizontal connections. The third pathway has one synaptic relay within layer IV and then conveys activity to layer III. In addition, monosynaptic activity is revealed in layers VI and I.
2.
In A 18 one coherent, fast-conducting group of afferents induces this basic activity pattern. In A 17 no such fast conducting input is resolvable; the supragranular activity is induced by a small group of afferents with intermediate conduction velocity, which terminate in the upper part of layer IV. The infragranular activity is induced by afferents with slower and widely scattered conduction velocities, which terminate in the lower part of layer IV. The layer VI input is very prominent in A 17 and also has a wide latency scatter.
3.
The supragranular activity is more prominent in A 18 than in A 17 and the respective layers appear thicker, in accordance with anatomy. In A 17 the infragranular activity prevails and layers IV and VI appear very broad, again in accordance with anatomy.
4.
Comparison of the CSDs with the original evoked potentials shows that the surface evoked potentials over A 18 reflect the three dipolar sink/source distributions of the coherent monosynaptic activity in layer IV and of the two prominent polysynaptic activities in layers III and II. The widely scattered activity in the lower part of layer IV in A 17 and all infragranular activities in both areas generate smaller, partly closed-field potentials; those are not discernible from the strong far-field potentials which originate from the supragranular activity and — especially in A 17 —from farther distant events.
The location and retinotopic organization of visual areas in the cat cortex were determined by systematically mapping visual cortex in over 100 cats. The positions of the receptive fields of single neurons or small clusters of neurons were related to the locations of the corresponding recording sites in the cortex to determine the representations of the visual field in these cortical areas. In this report, the first of a series, we describe the organization of area 17. A single representation of the cat's entire visual field corresponds closely to the cytoarchitectonically defined area 17.
This area has the largest cortical surface area (380 mm2) and the highest cortical magnification factor (3.6 mm2/degree2 at area centralis) of all the cortical areas we have studied. There was perfect agreement between the borders of area 17 determined electrophysiologically and cytoarchitecturally. This area contains a first order transformation of the visual hemifield in which every adjacent point in the visual field is represented as an adjacent point in the cortex. Some variability exists among cats in the extent and retinotopic representation of the visual field in area 17.
1. Cells in area 17 of the cat visual cortex were studied with a view towards correlating receptive field properties with layering. A number of receptive field parameters were measured for all units, and nearly every unit was marked with a microlesion to determine accurately the layer in which it was found.
2. Cells were defined as simple or complex by mapping with stationary stimuli, using the criteria of Hubel & Wiesel (1962). Complex cells fell into two groups: those that showed summation for increased slit length (standard complex) and those that did not (special complex).
3. The simple cells were located in the deep part of layer 3, in layer 4, and in layer 6. This corresponds to the distribution of afferents from the dorsal layers of the lateral geniculate nucleus. In these cortical layers the simple cells differed primarily with respect to their receptive field size, cells in layer 4 having the smallest, layer 3 intermediate, and layer 6 the largest fields. Layer 4 was the only layer in which simple cells showed end‐inhibition (a reduction in response to slits extending beyond the excitatory portion of the receptive field).
4. The standard complex cells were found in all layers, but were quite scarce in layer 4. As with the simple cells, field size varied with layer: in layer 2+3 they had small to intermediate field sizes, in layer 5 intermediate, and in layer 6 very large. Layer 6 cells showed summation for slits of increased length up to very large values, and responded best when the slits were centred in the receptive field.
The only standard complex cells that showed end‐inhibition were those in layer 2+3, and these were similar to the layer 4 simple cells in terms of proportion of end‐inhibited units and degree of end‐inhibition.
5. The special complex cells, originally described by Palmer & Rosenquist (1974), were found in two tiers: the upper one at the layer 3/layer 4 border and the lower one in layer 5. They were different from the standard complex cells in having a high spontaneous activity, high velocity preference, and large fields which were similar in size (at a given eccentricity) from one cell to the next. Many showed reduced response to slits of increasing length, even for slits that did not extend beyond the borders of the responsive region.
6. Cells in layer 6 (the origin of the corticogeniculate projection) were antidromically activated from the lateral geniculate nucleus. The antidromically activated units included both simple and complex cells, and they had the long receptive fields characteristic of the overall population of cells in layer 6.
7. The results showed that there are different types of simple and complex cells, and that cells in different layers have different properties. Taken together with their differences in site of projection, this demonstrates that the anatomical lamination pattern is reflected in functional differences between cells in different layers.
The extent of the spread of axonal degeneration was investigated in the visual cortex of the cat after making small lesions restricted to the grey matter. Two series of experiments were undertaken. In the first, normal adult cats were used, and in the second, the cortex of the postlateral gyrus was isolated from its extrinsic afferents by surgical undercutting 3 months before making the lesions. The results were similar in the two series in most respects. 1. Horizontal fibres extended in considerable numbers for some 500 micrometer from the lesion, mainly in layers I, III/IV and V, a few reaching 2/3 mm. These fibres were better seen in the intact than in the isolated cortex. Their spread was usually asymmetrical, being greater posteromedially than anterolaterally. 2. Oblique axons ran downwards from the middle layers into layers V and VI, or upwards into layers I and II. 3. Axons arising from layers II to VI descended vertically into the white matter. Degeneration patterns after lesions in areas 17 and 18 were compared.
Biocytin, a naturally occurring low molecular weight analog of biotin, was evaluated as a neuroanatomical tract-tracing marker in the adult rat brain. Since it retains high-affinity binding to avidin, biocytin can be labelled with avidinylated visualization reagents. Iontophoretic or pressure injections resulted in filling of cell bodies and dendrites around the injection site and their efferent axonal processes and boutons. Retrogradely labelled neurons were occasionally observed at a distance but only with large injections. Anterograde tracing with biocytin is successful even in animals that are quite old, in contrast to lectins and HRP conjugates, and offers advantages in delivery, tissue processing, selection of light and/or electron microscopic labels, time to obtain results, and cost over many conventional tracers.
Small injections of peroxidase-labeled wheat germ agglutinin into cat area 18 gave rise to patches of labeled cells and axon terminals around the injection site. In EM sections it was found that the labeled cells had a pattern of synaptic inputs characteristic of spiny-dendrite neurons (pyramidal or spiny stellate cells). The labeled axon terminals formed type 1 (asymmetric) synapses, most of which were made onto dendritic spines. In other experiments injections of fluorescent beads were made into area 18, giving rise to a similar patchy distribution of labeled cells. The sections were then processed for immunocytochemical demonstration of gamma-aminobutyric acid (GABA). The bead-labeled cells in the patches were GABA negative. The findings suggest that the patchy projections mediate mutual excitation between groups of spiny-dendrite neurons.
Biocytin is a biotin-lysine complex of low molecular weight containing about 65% biotin, which retains a high affinity for avidin. Since the latter molecule has been conjugated to several histochemical markers, the use of biocytin as an intracellular marker was investigated. Electrodes were filled with a solution of 4-6% biocytin dissolved in 0.5 M KCl and 0.05 M Tris buffer, pH 7-7.6. Neurons were recorded intracellularly in the supraoptic nucleus of an explant preparation of the rat supraoptico-neurohypophysial system and injected for 1-20 min with either hyperpolarizing or depolarizing current. Following variable recovery times, the explants were fixed in either 10% formalin or 4% paraformaldehyde overnight, sectioned on a vibratome, and incubated with the avidin-biotin complex (ABC) or avidin which had been conjugated to fluorescein, rhodamine, Texas Red or horseradish peroxidase and containing 1% Triton-X 100. A high percentage of injected neurons were recovered using each of the labels with about equal success. Both negative or positive current injection could be used with little electrode clogging. Labeling with fluorescent conjugates was qualitatively similar to that of Lucifer Yellow, whereas labeling with avidin coupled to horseradish peroxidase or with ABC was qualitatively similar to filling neurons directly with horseradish peroxidase. The advantages of this technique are the ease of injection of biocytin and the versatility in allowing the investigator to choose among light-emitting and light-absorbing images.
The structural features of two physiologically‐characterised pyramidal neurons (PC 1 and PC 2 ) closely situated in layer 5b in the visual cortex (area 17) of a single cat were studied using a combination of electrophysiological and anatomical techniques. Both PC 1 and PC 2 had exceptionally large somata (30–40 μm in diameter). On the basis of this and other morphological features cell PC 1 was classified as a Meynert cell. PC 1 possessed a very large (2.75° × 4.50°) binocularly driven standard complex receptive field. PC 2 was also binocularly driven with a small, B‐type receptive field. Both cells had the same preference for the direction and orientation of visual stimuli. PC 1 and PC 2 could be antidromically activated from stimulating electrodes positioned above the dorsal lateral geniculate nucleus with a response latency indicating that these cells probably innervated the visual tectum or pretectum.
In addition to corticoefferent axons, the two neurons possessed extensive intracortical axon arbors that ramified extensively in layers 5 and 6 of the medial and lateral banks of the lateral gyrus in area 17. Axon collaterals from both PC 1 and PC 2 also innervated a small common target region in area 18. A total of 313 boutons from the axonal arbors of PC 1 and PC 2 were examined in the electron microscope. All of the identified synaptic junctions were found to establish Gray type 1 asymmetrical contacts. The combined ultrastructural data for both neurons indicated that 80% of boutons were onto dendritic spine heads, with 14%, 6%, and 1% onto small‐, medium‐, and large‐calibre dendritic shafts, respectively. The spectrum of postsynaptic targets showed little variation with respect to lamina, distance from somata, or cortical area. Other large pyramidal neurons in layer 5 and spiny neurons in layer 6 were identified as receiving synaptic input from either PC 1 or PC 2 .
Using a computer graphics system, rotations of the bouton distributions revealed the existence of a clustered innervation of layers 5 and 6 in areas 17 and 18 derived from the two identified neurons. The bouton distributions strongly resembled the tangential pattern described previously for the functional slab‐like organisation of the cortex.
The results provide a morphological basis for the clustered intrinsic connectivity of pyramidal cells in layers 5 and 6 of the cat visual cortex. Furthermore, the results indicate the widespread excitatory influence of large pyramidal neurons on other cells projecting subcortically to sites dealing with visually guided behavior.
The number of synapses per unit volume of tissue (NV) has been estimated in individual laminae of the binocular and monocular regions of area 17 in six adult cats by using a method of size-frequency distribution. Separate estimates were obtained for RA synapses (containing round vesicles associated with asymmetric membrane differentiations) and for FS synapses (containing flat vesicles associated with symmetric membrane differentiations). For the total cortical thickness, the NV of all synapses is not statistically different between binocular (286 million per mm1(3] and monocular (281 million) regions, nor is it different between the two regions for any of the laminae. Eighty-four percent of synapses are of the RA type. Of those, 79% are found on dendritic spines, 21% on dendritic trunks, 0.1% on somata. FS synapses represent 16% of the total, with 31% of them on spines, 62% on dendritic trunks, and 7% on somata. The ratio of RA to FS synapses is kept relatively constant throughout the layers. A two-way analysis of variance shows no difference in the NV of either RA or FS synapses in the two regions nor in the NV or RA synapses between cats. It does, however, clearly demonstrate (p less than 0.001) interindividual differences for FS synapses. These variations between individual cats may be due to differences in age, breed, or environmental factors. In contrast to the relative uniformity of the NV of synapses between regions, the number of each type under 1 mm2 of cortical surface is 33% higher in the binocular region. This is due mainly to the greater thickness of the binocular region.
Most neurons in cat striate visual cortex show inhibitory effects when moving contours are presented beyond the limits of classic receptive field regions. Facilitatory effects are also present in about 40% of simple cells. Here, we report a highly specific form of this facilitation, mediated only by neurons possessing both an orientation tuning matched to the test unit, and a receptive field position aligned with its long axis. This finding illustrates one of the intracortical interconnection schemes hypothesized by Mitchison and Crick (1982). Periodic clustering in long, intrinsic axons may signify a neuron seeking specific functional interactions like these across columnar systems in both the spatial and orientation domains.
1. Based on average response histograms to an optimal stimulus, binocular interaction field plots were obtained from twenty‐five simple neurones in the striate cortex of the cat. Each binocularly activated cell has two interaction fields, one for each eye. The binocular interaction field for one eye plots the changes in the amplitude of the response from the other eye as the two receptive fields of the binocularly activated cell are moved across one another, first into and then out of alignment in the plane of the optimal stimulus (tangent screen).
2. The binocular interaction field provides an important clue to the nature of the spatial organization of the excitatory and inhibitory regions of the monocular receptive field. The commonest type of receptive field organization has regions of inhibition (inhibitory side bands) to either side of the discharge centre in the direction at right angles to the optimal stimulus orientation. As well as inhibition, there are subliminal excitatory effects.
3. Binocular interaction fields differ with the various cell types, i.e. cells that are discharged only from the one eye, cells binocularly discharged with very weak or absent monocular responses and cells showing binocularly opposite direction selectivity.
4. Marked facilitation to an optimal stimulus occurs when the two receptive fields of a binocularly activated neurone are in accurate alignment. Facilitation switches to depression for very small degrees of receptive field misalignment in a direction at right angles to the optimal stimulus orientation. These observations are of importance in relation to mechanisms for binocular single vision and depth discrimination.
A series of simultaneous recordings from several striate cortex neurons were made in paralyzed, anesthetized cats. Recordings were obtained with one or two bundles of extra fine wires and originated from one and two cortical orientation columns. Standard PST histograms and, in some cases, response planes were used to analyse the neuronal receptive fields. Functional connectivity between neurons was assessed by cross-correlation of their spike trains. It was found that 61% of neuronal pairs found within a column shared the same input, either excitatory or inhibitory, Even if neurons in a pair belonged to two different columns separated by 1mm lateral distance, 40% of pairs still exhibited shared input coordination. This type of coordination could also encompass all combinations of simple and complex fields in the pair. Direct connections between neurons were found almost exclusively within columns: excitatory connections were found in 20% of cases and inhibitory in 8%. Direct connections were often accompanied by the other types of interactions. Only one example of excitatory and one of inhibitory direct connections were found between columns. In both cases preferred orientations were almost identical.
The number of neurons in individual laminae of area 17 was determined separately for both the binocular and the monocular, regions in the left hemi-sphere of six cats. The number of neurons/mm3 of tissue was obtained for each lamina by using the method of size-frequency distribution applied to neuronal nuclei. The number of neurons per unit of cortical surface could then be calculated from measurements of layer, thickness. The number of neurons/mm3 of tissue for trie total cortical thickness is on the order of 48,000 to 50,000 neurons, with no statistically significant differences be-tween binocular and monocular regions. There are no significant differences for any of the layers except layer IV, in which the numerical density is 20% higher in the monocular region. The thickness of the cortex and of many of its layers, however, do vary between the two regions. Consequently there are significant differences in the number of neurons under 1 mm2 the total cortical thickness there are significantly more (27%) neurons in the binocular (78,440) than in the monocular region (61,900). This overall difference is due to significant changes in layers II, IIIA, IVA, and especially in layers V and VIA where neurons are 40% more numerous in the binocular region. These findings could signify either that the binocular region contains additional interneurons specifically related to binocular interactions or that it has a greater number of neurons projecting to other cortical and subcortical areas, or both.
The horizontal organization of iso-orientation bands in the cat's visual cortex has been investigated with the 14C-Deoxyglucose method (Sokoloff et al. 1977). Our findings suggest that in areas 17 and 18 the network formed by the bands representing one orientation comprises three basic patterns. Most frequently seen is a set of 2-4 iso-orientation bands which, straight or curved, run parallel to each other through cortical volumes usually less than 3 mm across. Beyond that distance the individual bands either end abruptly or fuse with adjacent bands. Less frequently seen are circular or triangular arrangements, or small isolated patches. Near the 17/18 border the predominant direction of the bands appears to be orthogonal to the vertical meridian, whereas in regions of areas 17 and 18 more distant from the vertical meridian the direction of the bands is more variable. In spite of irregularities in their spatial arrangement, the distance between bands and bandwidth remain constant throughout most parts of areas 17 and 18. The principles underlying the spatial arrangement of iso-orientation bands in the cat's visual cortex have yet to be identified.
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. 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. 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. 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. 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. 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. 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 of the striate cortex, while many axons of the layer 5 pyramids did not. 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 l.g.n. input. The majority of cells in layers 2, 3, 4 and 6 had the same basic S-type receptive field structure. Only layer 5 contained a majority of cells with spatially overlapping on and off subfields (C- and B-type receptive fields).(ABSTRACT TRUNCATED AT 400 WORDS)
The latencies of excitatory and inhibitory post-synaptic potentials (e.p.s.p.s and i.p.s.p.s) evoked by electrical stimulation of afferents from the lateral geniculate nucleus were recorded in neurones of area 17 of the cat visual cortex. After application of an extrapolation procedure to compensate for the conduction time of the afferent axons, a histogram of latencies formed three distinct peaks. Potentials in each of these were interpreted as being mediated by mono-, di- and trisynaptic pathways. Characteristic laminar differences in the extracellular field potentials evoked from the lateral geniculate nucleus (l.g.n.) and in the antidromic activation of neurones from the l.g.n. and superior colliculus were used to determine the laminar position of recorded neurones. It was found that within a given layer, all cells maintained similar connexions with relay cells in the l.g.n. Cells in layers 3, 4, upper 5 and 6 were monosynaptically excited by geniculate afferents, while cells in layers 2 and lower 5 received only indirect excitation via other cortical neurones. Layer 3 cells were unique in receiving a prominent disynaptic e.p.s.p. in addition to the direct excitation from the l.g.n. Late, trisynaptic e.p.s.p. components were seen in many layer 5 and 6 cells. The orderly laminar arrangement of the connexions had the consequence that identified cortico-geniculate neurones were monosynaptically excited and cortico-collicular neurones di- and trisynaptically excited by geniculate afferents. Cortico-cortical neurones in layers 2 and 3 received di- or mono- plus disynaptic excitation, depending on laminar position. Post-synaptic inhibitory potentials were evoked in all impaled cells, following stimulation of the geniculo-cortical pathway. Except for a few layer 2 cells, this inhibition was mediated through disynaptic pathways of the feed-forward type. There was a good positive correlation between conduction times for monosynaptic e.p.s.p.s and disynaptic i.p.s.p.s in the same cells, suggesting that cortical neurones receive excitation and inhibition from the same type of geniculate afferents. The stimulating electrodes activated not only geniculo-cortical afferents, but antidromically activated cortical efferent neurones from their extracortical axons. These neurones possess intracortical collaterals, and care must be taken to distinguish the resulting potentials from those mediated by orthodromic activation of geniculate afferents. In doing so, evidence was obtained for excitatory connexions from layers 2 and 3 to layer 5, from layer 5 to layer 6, and from layer 6 to layer 4. Typical recurrent inhibition was not observed.(ABSTRACT TRUNCATED AT 400 WORDS)
Rockland and Lung [Rockland, K. S. & Lung, J. S. (1982) Science 215, 1532-1534] have recently observed that an injection of horseradish peroxidase into the striate cortex of the tree shrew produces a patchy distribution of label adjacent to the injection site. They proposed that this pattern might be due to populations of neurons with long-range cortico-cortical connections that are interspersed with populations having no such connections. We suggest here an alternative explanation. We can account for the pattern by supposing that the label is carrier by a system of oriented axons. We suppose that these axons link cells with similar orientation preferences and make their connections within a narrow strip of cortex whose direction is related to the orientation of the cells in question. We suggest that such connections could be involved in generating complex receptive fields from simple ones. Other possibilities are that they are used to generate very elongated receptive fields, inhibitory flanks, or end-stopping. We suggest a number of experimental tests of these ideas.
Intrinsic connections within the tree shrew (Tupaia glis) visual cortex (area 17) are organized in periodic stripelike patterns within layers I, II, and III. This anatomical network resembles the regularly organized stripes of 2-deoxyglucose accumulation seen after stimulation of alert animals with uniformly oriented lines. Such connections imply that widespread lateral interactions are superimposed on the retinotopic organization of area 17 and suggest alternative interpretations of cortical columns.
Research with texture pairs having identical second-order statistics has revealed that the pre-attentive texture discrimination system cannot globally process third- and higher-order statistics, and that discrimination is the result of a few local conspicuous features, called textons. It seems that only the first-order statistics of these textons have perceptual significance, and the relative phase between textons cannot be perceived without detailed scrutiny by focal attention.
Three-dimensional reconstructions of the orientation column system were obtained from the visual cortex of four cats using the deoxyglucose technique. One cat had normal visual experience, one was monocularly and two had selective experience with vertical and horizontal contours, respectively. In areas 17 and 18 orientation columns form a remarkably regular system of equally spaced parallel bands whose trajectory is orthogonal to the borderline between areas 17 and 18. This topographic organization is resistant to manipulations of early visual experience.
We have demonstrated previously that neurons in cat striate cortex, in response to their preferred stimuli, exhibit oscillatory responses in a frequency range of 40 - 60 Hz. Recently, we obtained evidence that such oscillatory responses can synchronize across columns. We have now performed an extensive analysis of this phenomenon for both unit and field potential responses. In addition, we studied the stimulus conditions leading to intercolumnar synchronization. We recorded both multi-unit activity and local field potentials from area 17 of adult cats with arrays of several electrodes. Interelectrode distances ranged from 0.4 to 12 mm. For all pairs of unit (n=200) and field potential (n=174) recordings, we computed auto- and cross-correlation functions. The modulation of the correlograms was quantified by fitting a damped sine wave (Gabor) function to the data. Cross-correlation analysis of the unit data revealed that in 90 out of 200 cases the recorded cells established a constant phase-relationship of their oscillatory responses. This occurred, on average, with no phase difference. If the receptive fields were nonoverlapping, we observed a synchronization primarily between cells with similar orientation preferences. Cells with overlapping receptive fields also showed a high incidence of synchronization if their orientation preferences were different. In this latter group, synchronization occurred even in cases where the stimulus was optimal for only one of the recording sites. Under conditions of monocular instead of binocular stimulation the oscillatory modulation of the responses was attenuated, but the cross-correlogram still indicated a significant interaction. Similar effects were seen with the application of stationary instead of moving stimuli. A synchronization of oscillatory field potential responses was observed in 136 out of 174 paired recordings. At all distances investigated, the probability of synchronization of field potential responses was independent of the orientation preferences of the cells. However, the strength of interaction decreased with increasing spatial separation. Control experiments showed that the synchronization of field potential responses was not due to volume conduction. The results demonstrate that oscillatory responses at separate cortical sites can transiently synchronize. The probability and strength of synchronization are dependent on the spatial separation of the recorded cells and their orientation preferences. In addition, the cross-columnar synchronization is influenced by features of the visual stimulus. It is suggested that this synchronization provides a mechanism for the formation of neuronal assemblies in the visual cortex.
The cells in the primary visual cortex possess numerous functional properties that are more complex and varied than those seen in the cortical input. These properties result from the network of intrinsic cortical connections running across the cortical layers and between cortical columns. In the current study we relate the long receptive fields that are characteristic of layer 6 cells to the input that these cells receive from layer 5. The axons of layer 5 pyramidal cells project over long distances within layer 6, enabling layer 6 cells to collect input from regions of cortex representing large parts of the visual field. When layer 5 was locally inactivated by injection of the inhibitory transmitter GABA, layer 6 cells lost sensitivity over the portion of their receptive fields corresponding to the inactivated region of layer 5. This suggests that the extensive convergence in the projection from layer 5 to layer 6 is responsible for generating the long receptive fields characteristic of the layer 6 cells.
The morphology and postsynaptic targets of GABA-containing boutons were determined in the striate cortex of cat, using a postembedding immunocytochemical technique at the electron microscopic level. Two types of terminals, both making symmetrical synaptic contacts, were GABA-positive. The first type (95% of all GABA-positive boutons) contained small pleomorphic vesicles, the second type (5%) contained larger ovoid vesicles. Furthermore, 99% of all cortical boutons containing pleomorphic vesicles were GABA positive, and all boutons with pleomorphic vesicles made symmetrical synaptic contacts. These results together with previously published stereological data (Beaulieu and Colonnier, 1985, 1987) were used to estimate the density of GABA-containing synapses, which is about 48 million/mm3 in the striate cortex. The postsynaptic targets of GABA positive boutons were also identified and the distribution was calculated to be as follows: 58% dendritic shafts, 26.4% dendritic spines, 13.1% somata and 2.5% axon initial segments. A total of 11% of the postsynaptic targets were GABA immunoreactive and therefore originated from GABAergic neurons. The results demonstrate that the majority of GABAergic synapses exert their action on the membrane of dendrites and spines rather than on the somata and axons of neurons.