What do we know about laminar connectivity?

What do we know about laminar connectivity?

Accepted Manuscript What do we know about laminar connectivity? Kathleen S. Rockland PII: S1053-8119(17)30601-8 DOI: 10.1016/j.neuroimage.2017.07.0...

5MB Sizes 0 Downloads 10 Views

Accepted Manuscript What do we know about laminar connectivity? Kathleen S. Rockland PII:

S1053-8119(17)30601-8

DOI:

10.1016/j.neuroimage.2017.07.032

Reference:

YNIMG 14197

To appear in:

NeuroImage

Received Date: 12 March 2017 Revised Date:

13 July 2017

Accepted Date: 15 July 2017

Please cite this article as: Rockland, K.S., What do we know about laminar connectivity?, NeuroImage (2017), doi: 10.1016/j.neuroimage.2017.07.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

WHAT DO WE KNOW ABOUT

SC

Kathleen S. Rockland

RI PT

LAMINAR CONNECTIVITY?

AC C

EP

TE D

M AN U

Keywords: anterograde, distributed, feedback, feedforward, inputs, interneurons, projections

[email protected] Department of Anatomy&Neurobiology Boston University School of Medicine 72 East Concord St. Boston, MA. 02118

ACCEPTED MANUSCRIPT

WHAT DO WE KNOW ABOUT

AC C

EP

TE D

M AN U

SC

Kathleen S. Rockland

RI PT

LAMINAR CONNECTIVITY?

[email protected] Department of Anatomy&Neurobiology Boston University School of Medicine 72 East Concord St. Boston, MA. 02118

1

ACCEPTED MANUSCRIPT

M AN U

SC

RI PT

Abstract In this brief review, I attempt an overview of the main components of anatomical laminar-level connectivity. These are: extrinsic outputs, excitatory and inhibitory intrinsic connectivity, and intrinsic inputs. Supporting data are biased from the visual system of nonhuman primates (NHPs), but I have drawn as much as possible from a broader span in order to treat the important issue of area-specific variability. In a second part, I briefly discuss laminar connectivity in the context of network organization (feedforward/feedback cortical connections, and the major types of corticothalamic connections). I also point out anatomical issues in need of clarification, including more systematic, whole brain coverage of tracer injections; more data on anterogradely labeled terminations; more complete, area-specific quantitative data about projection neurons, and quantitative data on terminal density and convergence. Postsynaptic targets are largely unknown, but their identification is essential for understanding the finer analysis and principles of laminar patterns. Laminar resolution MRI offers a promising new tool for exploring laminar connectivity: it is potentially fast and macro-scale, and allows for repeated investigation under different stimulus conditions. Conversely, anatomical resolution, although detailed beyond the current level of MRI visualization, offers a rich trove for experimental design and interpretation of fMRI activation patterns. 1. Introduction

EP

TE D

Connectivity is investigated at multiple levels. Laminar connectivity incorporates microcircuitry and, especially as it can point to general principles, informs network architecture. Laminar resolution MRI, with its potential for global visualization and faster throughput, can be a welcome new tool in testing our current ideas of connectivity at the laminar level. After a short discussion of methods, the first part of the review is organized as a set of comments on extrinsic projecting neurons, excitatory and inhibitory intrinsic connections, extrinsic inputs, and white matter neurons. These various components have been schematized at a generic level (Fig. 1); but further understanding of how they contribute to laminar activation patterns will, I suggest, require a next generation of maps, perhaps customized to connectivity in different stimulus conditions.

AC C

The second part reviews laminar points of feedforward/feedback and corticothalamic connections, in the context of network architecture. Since this article is directed to the Neuroimage community, results are mainly drawn from experiments in nonhuman primates (NHPs) as being most relevant to human brain organization. A separate article could easily be devoted to species comparisons; for example, modality specializations (visual for primate vs. somatosensory and olfactory for rodents); relative cell packing density (“gray cell coefficient”); almost total lack of inhibitory interneurons in the rodent thalamus; and consequences of lissencephalic vs. gyrencephalic brains (e.g., DeFelipe, 2011; Laramee and Boire, 2015; Luebke, 2017). Similarly, a separate article could easily be devoted regional variability, which is only touched on in section 5. To a large extent, examples are from visual cortex, but with some comparisons to point out the difficulty, at least at this time in the field, of describing a common plan.

2

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 1. A generic summary of inputs, outputs, and intrinsic excitatory connections at the laminar level, for a nonprimary area of NHP cerebral cortex (reproduced with permission from Shipp, 2007, see references).

2. How is Anatomical Connectivity Investigated?

EP

TE D

2.1 Tracers Much of what we know about laminar connectivity is derived form relatively large anterograde or retrograde tracer injections (i.e., diameter > 250µm) in NHPs or other experimental animals (for one recent review, see Nassi et al., 2015). The combined results from decades of retrograde tracer injections in various target structures have demonstrated to a first approximation a basic plan for neocortical areas, in which supragranular neurons, in layers 2 and 3, project cortically; and infragranular layers 5 and 6 contain more heterogeneous subpopulations, with intermixed groups of pyramidal neurons projecting subcortically or cortically, from both layers (see Figs. 1, 2 and section 3.1). That infragranular layers have neurons that project cortically, as well as subcortically projecting neurons, is a fact that is often over-looked in the interests of simplification.

AC C

Retrograde tracers have gone through several generations of increasing sensitivity (horseradish peroxidase, fluorescent dyes such as Fast Blue and Diamidino Yellow, alexa-conjugated cholera toxin, and multiple viral based vectors). These provide clear visualization of labeled cell bodies, but typically only limited dendritic fills. This hinders finer anatomical characterization that could be inferred from differences in the dendritic arbors. Since injections are not cell-type specific, tracer identified populations of projection neurons are likely to include masked subtypes, which will need to be characterized by combination with other techniques; for example, see Briggs et al., 2016, where five distinct subtypes of corticogeniculate neurons are distinguished on the basis of dendritic arborizations, in a virus-mediated approach.

3

ACCEPTED MANUSCRIPT

SC

RI PT

Figure 2. Schematic summary illustrating the major targets of pyramidal cells in different layers (monkey primary somatosensory cortex). Note that subpopulations of neurons in the infragranular layers project cortically as well as subcortically (Reproduced from Jones, 1986; see references).

AC C

EP

TE D

M AN U

Anterograde tracers have had more dramatic technical improvements. The first widely used technique, autoradiography with tritiated amino acids, was a welcome improvement over silverdegeneration techniques, but provided only low resolution images that did not distinguish well between axons of passage, preterminal axons, and terminations. Significant improvement came with biocytin and kidney bean lectin, which were in turn superseded by biotinylated dextran amine (BDA; Fig. 3). These three all produced Golgi-like images, which permitted clear visualization of axons, axon branches, and terminations. This made feasible the identification of bilaminar terminations, the elucidation of spatial patterns, and the discrimination of numbers and morphological types of terminations. Multiple virus based anterograde tracers are now available.

4

Figure 3. A) and, progressively higher magnification (from double asterisk), C) and D) Projections from temporal cortex to the depth of the superior temporal sulcus, anterogradely labeled in Golgi-like detail by BDA. At the higher magnification in D), terminal boutons and segments of preterminal axons are readily identified. Note also that terminations extend into lower layer 3. B) For comparison, a closely adjacent Nisslstained section. Since this is at the junction of the lower bank and sulcal depth, the section

ACCEPTED MANUSCRIPT

also serves to demonstrate the change in laminar proportions in the sulcal depth; namely, expanded layer 1, and compressed layers 4, 5, and 5. L. 4 = layer 4. Scale bar = 200µm (A and B), 100µm (C), 50µm (D).

RI PT

Full neuron maps of the total dendritic and axonal arborizations are still difficult to achieve. Classical Golgi stains remain useful; and immunohistochemical markers (e.g., SMI32 or parvalbumin) will often result in detailed dendritic fills. Intra- or juxtacellular fills reliably produce high resolution fills, but are labor intensive techniques, with low yields. Intracellular biocytin labeling reliably fills dendrites and proximal portions of the axon, but is largely restricted to in vitro experiments. 3. Laminar connectivity

AC C

EP

TE D

M AN U

SC

3.1 Extrinsic projecting neurons Neurons in the supragranular layers project cortically, those in layer 4 do not project extrinsically, and those in the infragranular layers project to cortical and subcortical targets (Figs. 1, 2). Is more detail necessary? I suggest yes, especially if we can incorporate features such as systematic quantification across species and areas. This has been started for macaque cortical connections (Markov et al., 2014), but important extensions would be to include corticosubcortical connections and, where available, pyramidal cell subtypes. Some issues useful for a next generation of laminar maps are: 1. Supragranular neurons never project to the thalamus, colliculus, pons, or other subcortical targets; but some layer 3 neurons in frontal motor areas project to the striatum (Haber, 2016; Shipp, 2017), along with the more numerous corticostriatal neurons in layer 5. The supragranular corticostriatal component is more abundant in rodents (Haber, 2016). Whether there is reciprocal connectivity between the layer 3 and layer 5 corticostriatal populations has not been investigated. Corticostriatal connections are considered “intratelencephalic” (Shepherd, 2013; and Table 1 in Shipp, 2017), like the corticocortical. This provides a rational explanation for the layer 3 corticostriatal component, although why these are relatively few in number and limited to only a few areas in NHPs has not been explained.

5

Figure 4. Pyramidal neurons in anterior temporal (aTE) cortex retrogradely labeled by injection in posterior temporal cortex of a Golgi-like tracer (adenoviral vector expressing EGFP under the control of a neuron-specific promoter, synapsin I). Labeled neurons are mainly in layers 3 and 5 (that is, from this association area, “deviant” from the classic feedback in early sensory areas). Those deeper in layer 3 (solid arrow) extend basal dendrites into layer 4, but not those located more superfically in layer 3 (hollow arrow). Inset is at higher magnification, from solid arrow. Dashed lines denote border between gray and white matter (WM). L. 4 = layer 4. Scale bar = 250µm (125µm for higher magnification inset)

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

2. Layer 4 neurons do not project extrinsically; and because of this, layer 4 is often described as an interlaminar intrinsic relay zone. In fact, however, like all areas, there are abundant dendrites of passage (basal dendrites from extrinsically projecting neurons in deep layer 3 and apical dendrites of neurons in the infragranular layers) that invade and potentially receive inputs in layer 4 proper. A special case, related to issues of nomenclature, is the stria of Gennari in primary visual cortex. This zone consists of myelinated intrinsic connections intermingling with large spiny stellate and small pyramidal neurons, both projecting to extrastriate areas. Because this strata intervenes between geniculo-recipient layers 4C and 4A, it is often designated layer 4B (Brodmann nomenclature, as reviewed in Balaram and Kaas, 2014), despite the resulting anomaly of an extrinsically projecting “layer 4.” An alternate nomenclature (of Hassler), intended to avoid this confusion, considers the stria as “layer 3C,” even though the thin overlying layer 3 (aka layer 4A of Brodmann) is thalamo-recipient. Even when designated as “layer 4B,” the stria should be viewed as apart from “layer 4” in the usual usage. 3. Layer 1 neurons are primarily inhibitory and do not project extrinsically (but see section 3.2, point 6). 4. Both supra- and infragranular layers contain a heterogeneous mix of projecting neurons. Depending on the areas, layers 2 and upper 3 contain corticoamygdalar and cortical feedback neurons; lower layer 3 contains feedforward and callosally projecting neurons; layer 5 contains corticocollicular, corticostriatal, some corticothalamic, cortical feedback, and corticocortical neurons; layer 6 contains corticothalamic, corticoclaustral, and cortical feedback neurons. In the association cortices, layers 5 and 6 both contain feedforward as well as feedback neurons, albeit projecting to different areas (Saleem et al., 2000; Lavenex et al., 2004; Borra et al., 2010; Mohedano-Moriano et al., 2015). Knowing the proportion of each subpopulation would ideally require multiple retrograde tracers in the same animal, and has not yet been established in NHPs (but for an exhaustive tracer study in rats, see Gabbott et al., 2005). 5. In the primary sensory areas, there are no cortical feedback projecting neurons in layer 6. This seems self-evident but is worth stating, as one more of many distinguishing features of the primary areas. 6. There are multi-laminar origins: a) Neurons in layers 2, 3A, 5, and 6 give rise to feedback cortical projections, in area-specific combinations. Some of these in the infragranular layers, but seemingly not those in the supragranular, are a neurochemically distinct subpopulation, which uses synaptic zinc, an activity-related neuromodulator (Ichinohe et al., 2010). b) Neurons in layers 5 and 6 project to association thalamic nuclei (mediodorsal or pulvinar; Rouiller and Welker, 2000). c) Neurons in the supra- and/or infragranular layers project to the amygdala (Stefanacci and Amaral, 2000; Morecraft et al., 2007; Hoistad and Barbas, 2008; Cho et al., 2013). 7. Single neurons can project to multiple cortical areas (Rockland, 2015) and even to distinct structures; for example, there are collateralized branches from temporal cortical neurons to both the striatum and deep amygdaloid nuclei, and to the medial basal amygdala and perirhinal cortex or area TG (Cheng et al., 1997). Neurons in the inferior parietal lobule can branch to both the presubiculum, area TF, and the perirhinal/entorhinal border (Ding et al., 2000). The issue of collateralization is relatively under-investigated: double or triple retrograde tracers are hard to position optimally and may well fail to involve equivalent target zones. Single axon visualization – arguably the gold standard for this type of question – requires difficult inter- or juxtacellular fills with subsequent labor intensive serial section reconstruction.

6

ACCEPTED MANUSCRIPT

RI PT

Summary. General schematics illustrate that neurons in the supragranular layers project cortically (with the exception of some corticostriatal neurons in frontal and a few other areas), layer 4 and layer 1 neurons do not project extrinsically, and a mix of neurons in the infragranular layers project to subcortical and cortical targets. There is substantial areal variability of a quantitative nature; for example, early visual areas project only sparsely to the striatum and primary visual cortex not at all. As referenced in section 5 (below), the proportion of corticothalamic neurons in layers 5 vs. 6 is area dependent, as is the density of corticoamygdalar projecting neurons. The issue of inter-areal variability is important, and currently poses a distinct challenge for attempts to formulate cross-area, general rules.

AC C

EP

TE D

M AN U

SC

3.2 Excitatory intrinsic connections Local connections are instrumental in functional inter- and intralaminar integration. It has repeatedly been shown that intrinsic excitatory synapses substantially outnumber extrinsic connections such as the geniculocortical in area V1 and corticocortical from V1 to MT (Anderson et al., 1998; da Costa and Martin, 2011). These can be grouped in several categories. 1. Long (“horizontal”) collaterals of pyramidal cells in layers 3 and 5 seemingly occur in all neocortical areas, and in both NHPs and cats, where they form a patch-like pattern. These extend for a radius of 2.0-3.0mm from the parent soma, as best visualized by intracellular fills (see Fig. 5 from Ojima et al., 1991, a detailed in vivo study in cat auditory cortex). In tangential sections, the collaterals can be seen to have a spoke-like (“daisy”) configuration (Fig. 6, from Koestinger et al., 2017). The range and average number of collaterals per neuron are not known, and it is not known if all -- or most? -- of the pyramidal neurons have horizontal collaterals. Collaterals extend mainly in layers 3 and 5, with a preponderance of synapses in the home layer of the parent soma (cat auditory cortex: Ojima et al., 1991; monkey prefrontal: Melchitzky et al., 1998). In rodents, divergent collaterals exist but seem not to form terminal patches. Some pyramidal neurons (proportion unknown, but likely to be area-dependent) do not project extrinsically, but only have local, intrinsic connections. Figure 5. Intracellularly injected pyramidal cell (soma, in red) at the border of layers 2 and 3 of cat auditory cortex. Schematic representation of a) intrinsic axonal collaterals (1-7) in a coronal (radial) plane of section and b) bouton distribution, as projected onto a tangential plane. c) Schematic representation of the collaterals as projected onto a tangential plane (at left). Small coronal inset (at right) is for general orientation. Scale bars = 500µm. Reproduced with permission from Oxford University Press, Ojima et al., 1991 (see references).

The collaterals form prominent synaptic patches in NHPs and cats, which have a complicated relationship to the functional architecture in the sensory cortices (Fig. 6). Very recent experiments, combining light and electron microscopic approaches, indicate that for a single neuron, some of the collaterals are myelinated, and others not (Koestinger et al., 2017). This

7

ACCEPTED MANUSCRIPT

intriguing finding suggests complex combinatorial capacities for this system (“Thus, instead of treating the lateral connections as a single homogeneous network, the real clue to its structure and function may lie in its heterogeneity of connections,” Martin et al., 2014.)

SC

RI PT

Figure 6. Intracellularly injected pyramidal cell (soma and dendrites in red in A) with abundant intrinsic collaterals (cat visual cortex). A: coronal view. B: Tangential view to demonstrate the “spokelike” pattern of intrinsic collaterals, here shown in relation to optically imaged orientation domains. Scale bars = 500µm. Reproduced with permission from Brain Structure Function, Koestinger et al., 2017 (see references).

AC C

EP

TE D

M AN U

2. There are long, divergent collaterals from pyramidal neurons in layer 6. In area V1, the special class of large Meynert cells has an exceptionally dispersed collateral tree, greater (~8.0mm) than that of the better studied layer 3 axon collateral “daisies.” This is mainly confined within layer 6 but some cells have additional branches in layer 4B or even layer 1 (Rockland and Knutson, 2001). Corticoclaustral neurons in layer 6 are reported to have widespread collaterals within layer 6 (for cat: Katz, 1987; for rat: Thomson, 2010). 3. Corticothalamic neurons in layer 6 send widespread collateral branches to layer 4 (in area V1: Anderson et al., 1993) or layer 3 (in area V2: Lund et al., 1981). Whether this is true for feedback cortical neurons is not known. 4. Neurons in layer 4B of area V1 have long collaterals with multiple patchy terminations (Rockland and Lund, 1983; Lund, 1988). 5. Neurons in layer 4 (which are small pyramids, except for the spiny stellate neurons in the primary sensory areas) have both horizontal intralaminar and vertical interlaminar axons. This organization has been investigated in detail by Golgi studies (Lund, 1988) and layer-specific tracer injections (Casagrande and Kaas, 1994), but remains overall relatively under-investigated in NHPs. Highly selective connections from layer 4 to layer 3 pyramidal neurons have been defined in vitro on the basis of the interconnected subnetworks of layer 3 neurons (in rat: Yoshimura et al., 2005). 6. According to one recent study, a distinctive class of calretinin-positive (non-GABAergic) neurons in layer 1 (“subpial fan cell”) has a widespread arborization in layer 1 (NHP prefrontal and other areas: Gabbott, 2016). Layer 1 otherwise contains abundant inhibitory neurons, with predominantly local innervation territories. Summary. There is an elaborate cortical-wide network of both “horizontal” (i.e., spoke-like) intralaminar and vertical, interlaminar excitatory connections. How this interacts with extrinsic afferents and the equally elaborate network of intrinsic inhibitory connections remains a major research question. From this perspective alone, however, the simplifying concept of layer 4 as a “relay” for interlaminar connections can seem wholly inadequate. 3.3 Inhibitory intrinsic connections

8

ACCEPTED MANUSCRIPT

SC

RI PT

The literature for inhibitory interneurons is vast in terms of the diversity of subtypes, laminarand area-specific distribution, and, to some extent, postsynaptic targets. Fewer data are available for NHPs, as compared with rodents, and of this, much of the fine detail on laminar connectivity still derives from Golgi studies in young monkeys (Lund, 1988 and Fig. 7). This, along with several quantitative immunohistochemical analyses, abundantly demonstrates the diversity of these neurons and, importantly, their variability across areas (Gabbott and Bacon, 1996; DeFelipe et al., 1999). As a very general number, inhibitory neurons will be about 20% of the total neurons per area; and of those, the parvalbumin subpopulation is the most abundant. A helpful recent review of inhibitory circuits compares a model of surround modulation in area V1 of NHPs with experimental data on inhibitory circuits (mainly derived from NHPs; Angelucci et al., 2017). Without going into details of microcircuitry – most of which are unknown in NHPs – in this section, I will again only list a few points I thought might be of particular relevance to neuroimaging.

TE D

M AN U

Layers 4A, 3 (intr.)

Figure 7. Diagram summarizing the interneuron varieties associated with laminae 4A and 3B as described in a Golgi study of area V1 (NHP). Note diversity and different combinations of interlaminar innervations. Reproduced with permission from Wiley Online Library, Lund and Yoshioka, 1991 (see references).

AC C

EP

1. Inhibitory subtypes have a) very local, b) vertical interlaminar, or c) horizontal (that is, spokelike) intralaminar axon terminal fields. Large, parvalbumin positive basket cells extend myelinated axons for 2.0-3.0mm from the parent soma, in a spoke-like pattern reminiscent of the patchy pyramidal cell collaterals (in cat, Kisvarday, 1992; Karube et al., 2016). At least some large basket cells arborize in both infra- and supragranular layers, thus offering one of several substrates for interlaminar inhibitory interactions. 2. Available evidence indicates that inhibitory interneurons participate in complex and specific microcircuits. One example is the calretinin-positive population in NHP area V1 (reported as about 14% of the total GABAergic population in area V1; Meskenaite, 1997). These form synapses locally with other interneurons in layers 1-3, but also - in area V1 but not V2 - have a basket-like innervation of pyramidal neurons in layers 5 and 6 (and see, for monkey prefrontal cortex: Melchitzky et al., 2005). A second example, from an in vitro study of two layer 1 interneurons in rat (with recordings from up to eight postsynaptic locations) demonstrates highly selective disinhibitory and inhibitory influences - by neurogliaform or single bouquet cells - with a “menagerie” of other cortical interneurons and with pyramidal neurons in layers 5 and 3 (Lee et al., 2015). The circuitry is suggestive of multiple, parallel postsynaptic routings potentially available for different tasks (Larkum, 2013: “It now remains to be discovered through which L1projecting pathways and under which circumstances the different circuits are activated”). 3. Interneuron subtypes and the number of each subtype are laminar-, area-, and species specific. Area-specificity has been demonstrated repeatedly in NHPs: for three medial prefrontal areas (Gabbott and Bacon, 1996), and for temporal-occipital areas (DeFelipe et al., 1999). A

9

ACCEPTED MANUSCRIPT

RI PT

quantitative immunohistochemical study of human postmortem brain (Blazquez-Llorca et al., 2010) reported a complementary high-low density, across nineteen areas, of chandelier axon terminations (targeting pyramidal cell axon initial segments) vs. complex basket cell formations (targeting pyramidal cell soma). That is, chandelier and basket cells contribute distinctively to cortical circuits (anatomically and physiologically), depending on the cortical area or layer in which they are situated. 4. A small subpopulation of neurons co-localizing nitric oxide and neuropeptide Y (respectively, vasodilator and vasoconstrictor) has been implicated in neurovascular control (e.g., Cauli et al., 2004; and see references in Rockland and Nayyar, 2012). This may figure among the mechanisms underlying the BOLD signal (e.g., Goense et al., 2016).

SC

Summary. There is an impressive (aka bewildering) diversity of inhibitory neurons and their associated microcircuitry. This may betoken a very high degree of specificity, not yet fully appreciated, but required in conjunction with the complexity of the external and internal environments, and serving to enhance plasticity and flexibility of response.

TE D

M AN U

3.4 Extrinsic inputs Thalamocortical and corticocortical inputs are perhaps the best investigated of multiple cortical inputs, and are themselves from multiple nuclei and areas. There are also callosal, amygdalocortical, and claustrocortical inputs, in addition to the neuromodulatory serotonin, noradrenaline, cholinergic, and dopaminergic inputs. The colliculi and striatal nuclei do not project to the cortex; and the hippocampus (CA1 or subiculum) only to perirhinal cortex and medial frontal, among the neocortical areas. A long history of cortical inactivation experiments, by reversible cooling or, more recently, photoinhibition techniques (Diester et al., 2011; Han et al., 2011), has provided data on how inputs contribute to normal response properties (e.g., V1 to V2: Girard et al., 1992; Payne et al., 1996; Schmid et al., 2009).

AC C

EP

The laminar organization of inputs is commonly visualized at a global level after bulk injections of anterograde tracers in the source area (section 2). These, however, are averages of multiple individual arbors which single axon analysis and reconstructions show to be of different sizes and shapes (Rockland, 1997). For inputs, the knowledge gap is greater than for data based on retrograde tracers, and include the problems of 1) unknown postsynaptic dendritic targets, 2) convergence of multiple inputs in the same layer (i.e., thalamic and cortical), and 3) hard to investigate temporal interactions. Temporal correlations have been identified as an important influence on response variability in the target area (Gomez-Laberge et al., 2016, and discussion in Self and Roelfsema, 2017). Laminar preferences are not stereotyped across areas. Thus, amygdalocortical connections target the layer 1, 2 border in area V1, but are more widely distributed in both the supra- and infragranular layers in temporal areas (Freese and Amaral, 2005). This is an instance where variability may be explainable by a systematic rule; that is, laminar organization is shown to vary according to the “distance” between source and target: typically, connections are denser and involve more layers if source and target are “closer.” The distance rule was first formulated from observations that retrogradely labeled feedback-projecting neurons from area V2 to V1 were strongly bilaminar, but those from MT to V1 were concentrated in layer 6 (see Markov et al.,

10

ACCEPTED MANUSCRIPT

2014; Wang and Kennedy, 2016). It can also apply in the anterograde direction, as in the case of amygdalocortical connections.

Figure 8. Diagram summarizing a schematic microcircuit for feedback connections to area V1. Note that feedback connections from area V2 and other extrastriate areas potentially contact infragranular neurons at multiple sites (yellow), including basal dendrites in layers 5 or 6 and distal dendrites in layer 1. Red = approximate sites of geniculocortical terminations. Reproduced with permission from Wiley Online Library, Rockland and Virga, 1989 (see references).

EP

TE D

M AN U

SC

RI PT

Other points important to consider include: 1. To say that a connection terminates in layer 4 is a convenient shorthand, but often is not the whole truth. For example, “feedforward” cortical input, in the visual areas, also extends into layer 3 (e.g., Rockland and Pandya, 1979, among many others; and see Figs. 1 and 3). Thalamocortical inputs from the pulvinar (Lund et al., 1981; Rockland et al., 1999; Shipp, 2003) and from mediodorsal thalamus (Erickson and Lewis, 2014) terminate preferentially in a “middle zone,” consisting of upper layer 4 and lower layer 3. Single axon visualization clearly shows that one pulvinocortical axon can have multiple arbors, some preferentially targeting layer 4 and some targeting layer 3 (Rockland et al., 1999). 2. The middle zone consisting of upper 4 and overlying layer 3 is also distinguishable by heightened enzyme activity for cytochrome oxidase (Lund et al., 1981; Casagrande and Kaas; 1994; Levitt et al., 1995; Rockland, 1996; Erickson and Lewis, 2004). 3. Multiple studies have demonstrated amygdalocortical inputs to layer 1 or the layer 1,2 border (temporal and occipital cortex: Freese and Amaral, 2005; cingulate motor cortex: Morecraft et al., 2007; frontal cortex, Miyashita et al., 2007; Timbie and Barbas, 2014). This zone also receives convergent input from cortical areas, association thalamus, and the inhibitory zona incerta (Mitrofanis, 2005).

AC C

4. As shown most clearly by single axon visualization, inputs can target pairs of layers; namely, layers 4 and 6 (geniculocortical: Blasdel and Lund, 1983; Freund et al, 1989; V1 to MT: Rockland, 1989; Anderson et al., 1998)); layers 1 and 6, or layers 1 and 5 (cortical feedback: Rockland and Virga, 1989; and Fig. 8); layers 3 and 4 (thalamocortical and corticocortical: point 1, above). Multi-laminar terminations are a relatively under-investigated area, and suffer from small sample sizes, relative lack of multiple distinguishable tracers, and limited coverage of cortical areas in NHP. Summary. The detailed organization of the complete set of cortical connections, even to area V1, is not well established. Currently, results are correlations from single tracers in different experiments; but a more accurate investigation would use two or three distinguishable tracer injections in one experiment. A provocative comparison is the intricate, multi-laminar

11

ACCEPTED MANUSCRIPT

organization of eight association area inputs to subdivisions of entorhinal cortex (Fig. 9; from Insausti and Amaral, 2008; but also by correlation across experiments).

TE D

M AN U

SC

RI PT

Figure 9. Summary diagram of the laminar organization of inputs to NHP entorhinal cortex (subfields indicated at top, layers at the left, and eight anterogradely labeled regions at bottom (OF, orbitofrontal; MF, medial prefrontal; I, insular; STS, superior temporal sulcus; PR, perirhinal; PH, parahippocampal; AC, anterior cingulated; RSP, retrosplenial cortex). Colors denote intricate mulit-laminar pattern (not necessarily stratification), as revealed by anterograde tracer injections in the several source regions. Reproduced with permission from Wiley Online Library, Insausti and Amaral, 2008 (see references).

AC C

EP

3.5 White matter neurons (WMNs) The vast majority of connectivity studies focus on gray matter to gray matter, and tend to ignore a mixed population of excitatory and inhibitory neurons in the superficial and deep white matter (see Mortazavi et al., 2016, and the references therein). These neurons are not densely distributed, and are unlikely to be visualized by MRI. They are relevant to the imaging community because 1) at least some are integrated into the overlying gray matter (Fig. 10), 2) at least the nitric oxide+ subpopulation (~5%) is involved in neurovascular coupling, and 3) they contribute to the dense axo-dendritic neuropil in the superficial white matter that could impact on technical aspects of DTI (Reveley et al., 2015). WMNs are phylogenetically conserved, and have been reported in rodents, carnivores, and cetaceans, as well as human and NHPs (see Introduction, Mortazavi et al., 2016). The density and subtype distribution is species- and areaspecific. In humans, the density of superficial WMNs is reported greatest for frontal cortex area 10 (~ 2,660 /mm3) and least for visual and temporal areas (areas 17, 20, 24: ~1,770/mm3), with the majority of neurons being excitatory. Of the inhibitory subpopulation, calretinin+ neurons are the most numerous in human (Garcia-Marin et al., 2010).

12

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

TE D

Figure 10. Dendrites cross freely in both directions over the border between layer 6 and the white matter (WM). A) and (higher magnification) B) Neurons in the superficial WM labeled by the enzyme nicotinamide adenine dinucleotide phosphate diaphorase (NADPH-d) extend dendrites (double asterisk) into the overlying gray matter. C) and D) Golgi-like labeling with a retrogradely transported adenoviral vector (see Fig. 4) shows that neurons in layer 5 can extend dendrites into the underlying WM. These are neurons in ventral temporal cortex, labeled by an injection in lateral surface area TE. Inset in C) is higher magnification, from the double asterisk, to show detail. E) is higher magnification (from the double asterisk in D) to show dendritic spines along the WM portion (single asterisk). Dashed lines in C) and D) indicate upper border of WM. Scale bar = 100µm (A-D), 50µm (E).

EP

4. What Laminar Patterns Can Say About Network Architecture

AC C

4.1 Feedforward/feedback cortical projections As summarized many times, feedforward (FFD) and feedback (FBK) extrinsic cortical connections are anatomically definable as separate types by several criteria (and see Fig. 1; and Table 1 in Adams et al., 2013). 1) Distinct layers of origin: layer 3 (FFD), or layers 2, 3A, and 6 (FBK); 2) Distinct layers of termination: layer 4 (FFD), or layers 1 and 6 (FBK); 3) Largely distinct postsynaptic targets (see below); and 4) Spatially distinct patterns of arborization: relatively small arbors (diameter < 0.5mm, FFD), or relatively divergent arbors ( extent >1.0mm, FBK). Dense layer 1 “feedback” projections are an influential counterpoint to the idea of a predominantly feedforward, layer 4-biased “relay” cortical architecture; and evidence from retrograde tracers (injected in V4 and V1, to look for double labeled neurons in V2) so far indicates that FFD and FBK projecting neurons are indeed separate populations (Markov et al., 2014; Shipp, 2016); but the separation of inputs - FFD input to layer 4 and FBK to layer 1 - is 13

ACCEPTED MANUSCRIPT

TE D

M AN U

SC

RI PT

somewhat deceptive, and may be less compelling at the level of individual axons and postsynaptic targets (Fig. 8). The laminar dissociation is more blurred in rodents (Coogan and Burkhalter, 1993; Berezovskii et al., 2011); and, as frequently discussed, applies only approximately outside of the sensory areas (e.g., Shipp et al., 2013; Barbas, 2015; and see Fig. 4). With continued investigations, particularly with anterograde tracers, important modifications are likely to be identified across the sensory modalities; for example, a systematic recent analysis of laminar connectivity patterns of primate auditory areas found both similarities and differences with the standard model of visual cortical organization (Hackett et al., 2014, and Fig. 11).

Figure 11.Summary of laminar terminations of anterogradely labeled connections within the auditory pathway. (Reproduced from Frontiers in Neuroscience, Hackett et al., 2014; see references).

AC C

EP

In interpreting laminar activity patterns, several other features of FFD and FBK systems need to be kept in mind. One is that there is extensive convergence and presumably flexible interaction with other systems. As listed in section 3.4, these include callosal inputs (homo- and heterotopic), inputs from the amgydala and from multiple cortical areas, both FFD and FBK. FBK terminations converge at the laminar level with inhibitory terminations from the zona incerta in layer 1 (Mitrofanis, 2005), or, in area V1, with pulvinocortical terminations in layer 1, and geniculocortical terminations to layers 1 and 6 (Blasdel and Lund, 1983; Freund et al., 1989). The detailed laminar organization, input convergence, and relative synaptic weights (for example, to area V1 from V2, V4, MT/V5, V3, and TEO) are not established. In addition, layer 6 FBK neurons have local collaterals which would interact with the extensive systems of withinarea intrinsic connectivity; for example, with the extensive system of pyramidal cell collaterals in layer 5 (see section 3.2, and discussion in Stettler et al., 2002; Angelucci et al., 2017). Intracelluar fills demonstrate dense local projections from corticothalamic layer 6 neurons to layer 4 (Anderson et al., 1993); data for layer 6 FBK neurons in specific are still lacking. How the several layers, individually and in different combinations, contribute to various stimuli aspects might be addressable with higher resolution MRI and stimulation paradigms.

14

ACCEPTED MANUSCRIPT

RI PT

Second, FFD and FBK connections are often discussed as reciprocating. However, single axon investigations demonstrate a spatial asymmetry, in that FDK axons typically have a divergent territory in layer 1 (>1.0mm), but FFD connections have more delimited arbors (0.25-0.50mm). Thus, more neurons in a wider sector are contacted in the FBK direction (see further discussion in Shipp, 2016). Third, important information is lacking about postsynaptic targets. Although distal apical dendrites in layer 1 are often cited as a prime target of FBK projections, a layer 5 neuron potentially receives FBK input on both distal and basal dendrites (Fig. 8), and similarly for those layer 6 neurons when apical dendrites extend to layer 1.

M AN U

SC

With continuing work, one can expect further, functionally significant subdivision of FFD/FBK types of connections (i.e., extending beyond a simple FFD/FBK dichotomy). On the down side, this may lead to discrepancy between increasingly sensitive anatomical resolution, and what can be achievable, at least initially, with laminar specific MRI; but better anatomical resolution should inform experimental design and interpretation of the MRI images. For example, anatomical connections in the early visual pathway have long been known to include, in addition to areal pairwise connections, poorly understood bypass connections -- from area V1 to V4, and area V2 to TEO -- thought to subserve a coarse information, rapid processing route (Nakamura et al., 1993). For FBK, in addition to pairwise area connections, there is also a subset of branched connections to multiple earlier visual areas (i.e., area TEO projections to area V4, V2, and V1; Rockland et al., 1994). The proportion of these auxiliary projections is unknown.

AC C

EP

TE D

4.2 Type 1 and Type 2 Corticothalamic Projections Cortical connections to the pulvinar and other association thalamic nuclei are distinguished as two types, in part on the basis of their bilaminar origin in layers 5 and 6. Despite species differences in thalamic organization, the distinction is conserved in rodents and carnivores as well as NHPs (Rouiller and Welker, 2000). Along with other distinguishing features, this has led to the influential distinction of driving vs. modulatory corticothalamic connections (originating respectively from layers 5 and 6). In summary (and see Table 1 in Sherman, 2017), a small number of large layer 5 pyramidal cells terminate in the thalamus with one to several small arbors (diameter ~ 0.25mm), having a small number (~200) of large boutons (often ~3.0µm) at proximal locations on the postsynaptic thalamic neurons. These are thought to be collaterals from corticopontine neurons. In contrast, a large number of layer 6 neurons terminate with divergent axon arborizations (>1.0mm), having a large number (~1,000) of small, stalked boutons (reminiscent of dendritic spines) at distal locations (Rockland, 1996; Shipp, 2003; Sherman, 2017). Corticothalamic neurons in layer 6 have distinct gene expression profiles from cortical FBK neurons and, area dependent, are often deeper in the layer (Fig. 1; and Watakabe et al., 2007). The assortment of synaptic arrangements actually comprises a wide organizational diversity, appropriate to increased computational capabilities. As discussed in Bickford, 2016, there are abundant examples where large “driver” cortical terminations converge with ascending “driverlike” terminations from the superior colliculus or trigeminal nucleus, possibly as a means of reporting relative timing between sensory events and ongoing cortical activity. Moreover, as noted for FFD/FBK connections, there is clear evidence of morphological diversity within each type (Kultas-Ilinsky et al., 2003), supporting the possibility of something more complex than any simple duality. The importance of the marked spatial differences (focused vs. divergent axonal arborization) has also not been explained within the driver/modulatory framework. 15

ACCEPTED MANUSCRIPT

RI PT

Projections between cortical areas and the thalamus are reciprocal. At the laminar level, there is some dissociation, where dense, topographically focused terminations (from “core” thalamocortical neurons) target upper layer 4 and lower layer 3 (Levitt et al., 1995; Erickson and Lewis, 2004), while a more divergent subpopulation (“matrix”), targets layer 1. Postsynaptic targets are largely unknown. An interesting, but enigmatic clue to organization is that thalamocortical axons visualized at the single axon level in the early visual pathway have multiple arbors, which frequently target several different layers, including the combination of layers 4 and 1 (Rockland et al., 1999). 5. Regional variability

AC C

EP

TE D

M AN U

SC

Anatomical connectivity is not uniform, either across or within areas. Callosal connections typically do not extend throughout an area. In the early visual pathway, differences related to field visual periphery are well-known for areas V1, V4, and MT/V5. There is a striking laminar difference, still unexplained, in the feedback connections from area MT/V5 to V1. That is, in the central field representation in V1, these terminate in layers 4B and 6; but in the peripheral field representation, terminations are in layers 1, 4B, and 6 (Ungerleider Desimone, 1986; Shipp and Zeki, 1989; Rockland and Knutson, 2000). The topographic switch point has not been identified, nor is it known if there are topographic effects in laminar terminations in other visual areas.

Figure 12. A), and higher magnification, B) (from the single asterisk) and C) (from the double asterisk) of feedback neurons in area V2 labeled in Golgi-like detail by a retrogradely transported adenovirus (see Fig. 3). Neurons are in upper layer 3 and layer 6. Images in B) and C) have been rotated, so that the pia surface is at the top. Arrow points to the V1/V2 border. LS = lunate sulcus. Scale bar = 500µm (A), 100µm (B, C).

Both areas V1 and MT/V5 have different connections as a function of visual field eccentricity; namely, there is a wider set of connections to non-central regions. There are inputs from nonvisual areas (putatively multisensory) to the peripheral representation in area V1 (auditory: Falchier et al., 2002; Rockland and Ojima, 2003; parietal: Borra and Rockland, 2011). The peripheral representation in area MT/V5 receives inputs from retrosplenial cortex, prostriata, area V1 and MT-satellite areas, but less from extrastriate areas. Those, by contrast, target more central visual field representations in MT/V5 (in marmoset: Palmer and Rosa, 2006). Areas V2

16

ACCEPTED MANUSCRIPT

RI PT

and V4 provide direct peripheral field inputs to dorsal stream extrastriate visual areas; area V3A receives projections only from the peripheral representation of area MT/V5; and area V4 receives projections, if at all from V1, only in the foveal representation (reviewed in Ungerleider et al., 2008). The direct projections preferentially to peripheral field representations have been discussed as a means to rapidly activate circuits for spatial vision and spatial attention (Ungerleider et al., 2008).

SC

Corticothalamic neurons in layer 6 are more numerous than those in layer 5, but the proportions are not uniform across areas (Erickson and Lewis, 2004; Xiao et al., 2009). Bilaminar projections, from the supra- and/or infragranular layers, project to the amygdala, but again in proportions that vary across areas (Stefanacci and Amaral, 2000; Morecraft et al., 2007; Hoistad and Barbas, 2008; Cho et al., 2013).

M AN U

Depth position within the different layers also influences connectivity. Projection neurons are not monolayer. In layer 3, cortical FFD neurons can distribute over a 200-400µm wide zone. Neurons in the deeper strata will likely have basal dendrites invading layer 4 and sharing layer 4 inputs, but those in the middle and upper strata will not (see Figs. 1, 2, and 12B).

Figure 13. Retrogradely labeled neurons in aTE (same case as illustrated in Fig. 4). As a result of the sulcal deformation, apical dendrites converge (*) and presumably share common inputs, even through parent somas (hollow arrows) are spatially non-adjacent. Scale bar = 125µm

AC C

EP

TE D

Something as simple as the well-known sulcal and gyral deformations (Welker, 1990) also impacts on laminar connectivity. Layer 4 and deeper layers are relatively compressed in sulcal depths and expanded in gyral crowns; and the density of both inputs and projection neurons will be differentially influenced relative to non-sulcus adjoining cortex, even considered as the same area (e.g., Fig. 1 in Rockland, 1997). Infragranular neurons are likely to have an exaggerated slant at gyral crowns (compare Fig. 12B and 12C); and neurons in the sulcal depths will have a greater convergence of apical dendrites. Sulci usually have an accentuated layer 1, which will result in a larger zone of layer 1 input on distal apical dendrites (Fig. 13).

17

ACCEPTED MANUSCRIPT

5. Conclusions

RI PT

“What do we know about laminar connectivity”? A lot, at the global laminar level, but not so much at the level of circuitry. How does the intricate microcircuitry contribute to global laminar processes? Signals like neural oscillations, although they often have a specific laminar pattern, are hard to relate closely to underlying anatomy: Which cortical layers generate rhythms? Is the generation equivalent across sensory systems? What is the mix between excitatory and inhibitory local circuits (e.g., Besserve et al., 2015; Buzsaki and Schomberg, 2015; Haegens et al., 2015; Michalareas et al., 2015)?

M AN U

SC

The laminar organization of connections incorporates a high degree of alternative or parallel processing. Which microcircuits are recruited can be assumed to be highly stimulus- and taskdependent (Raichle, 2015; Zeki, 2016). The anatomy is not stereotyped across areas, and differs both quantitatively and qualitatively. At the network level, the anatomical organization arguably is only a partial fit with a strict hierarchical organization. In one alternative proposal, the [visual] brain is seen as “constituted of multiple, parallel and asynchronously operating task- and stimulus-dependent hierarchies” (Zeki, 2016). The primary areas may function at a different level of abstraction (Christophel et al., 2017).

TE D

Anatomical connections need to be understood in the context of polysynaptic, distributed processing. This will become increasingly possible to investigate; for example, with experiments combining targeted inactivation with functional imaging (e.g., Grayson et al., 2016). Finally, the relationship of vascularization in relation to cellular layers and inputs can be expected to figure more prominently with the increased use of laminar visualization in fMRI (see discussion in Huber et al., 2014; Adams et al., 2015; Goense et al., 2016). As MRI experiments move to laminar resolution, we can hope for an enriched dialogue between imaging results and anatomical substrates. Global, stimulus driven, cross-area activity patterns will be a welcome complement to the anatomical toolbox. Connectional anatomy, in turn, can be a rich source for interpretation and formulation of network architectures.

EP

Conflict of interest The author declares no competing financial interests.

AC C

Acknowledgements I would like to thank Andrew Chang for help with manuscript preparation, and Dr. Christina Tognoni for assistance with figures. KSR was partially supported in the writing by MH106796 and MH107456. References

Adams, D.L., Piserchia, V., Economides, J.R., Horton, J.C., 2015. Vascular supply of the cerebral cortex is specialized for cell layers but not columns. Cereb. Cortex 25, 3673–3681. doi:10.1093/cercor/bhu221 Adams, R.A., Shipp, S., Friston, K.J., 2013. Predictions not commands: active inference in the motor system. Brain Struct Funct 218, 611–643. doi:10.1007/s00429-012-0475-5

18

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Anderson, J.C., Martin, K.A., Whitteridge, D., 1993. Form, function, and intracortical projections of neurons in the striate cortex of the monkey Macacus nemestrinus. Cereb. Cortex 3, 412–420. Anderson, J.C., Binzegger, T., Martin, K.A., Rockland, K.S., 1998. The connection from cortical area V1 to V5: a light and electron microscopic study. J. Neurosci. 18, 10525–10540. Angelucci, A., Bijanzadeh, M., Nurminen, L., Federer, F., Merlin, S., Bressloff, P.C., 2017. Circuits and mechanisms for surround modulation in visual cortex. Annu. Rev. Neurosci. 40, 425–451. doi:10.1146/annurev-neuro-072116-031418 Balaram, P., Kaas, J.H., 2014. Towards a unified scheme of cortical lamination for primary visual cortex across primates: insights from NeuN and VGLUT2 immunoreactivity. Front Neuroanat 8, 81. doi:10.3389/fnana.2014.00081 Barbas, H., 2015. General cortical and special prefrontal connections: principles from structure to function. Annu. Rev. Neurosci. 38, 269–289. doi:10.1146/annurev-neuro-071714033936 Berezovskii, V.K., Nassi, J.J., Born, R.T., 2011. Segregation of feedforward and feedback projections in mouse visual cortex. J. Comp. Neurol. 519, 3672–3683. doi:10.1002/cne.22675 Besserve, M., Lowe, S.C., Logothetis, N.K., Schölkopf, B., Panzeri, S., 2015. Shifts of gamma phase across primary visual cortical sites reflect dynamic stimulus-modulated information transfer. PLoS Biol. 13, e1002257. doi:10.1371/journal.pbio.1002257 Bickford, M.E., 2015. Thalamic circuit diversity: modulation of the driver/modulator framework. Front Neural Circuits 9, 86. doi:10.3389/fncir.2015.00086 Blasdel, G.G., Lund, J.S., 1983. Termination of afferent axons in macaque striate cortex. J. Neurosci. 3, 1389–1413. Blazquez-Llorca, L., García-Marín, V., DeFelipe, J., 2010. GABAergic complex basket formations in the human neocortex. J. Comp. Neurol. 518, 4917–4937. doi:10.1002/cne.22496 Borra, E., Rockland, K.S., 2011. Projections to early visual areas v1 and v2 in the calcarine fissure from parietal association areas in the macaque. Front Neuroanat 5, 35. doi:10.3389/fnana.2011.00035 Borra, E., Ichinohe, N., Sato, T., Tanifuji, M., Rockland, K.S., 2010. Cortical connections to area TE in monkey: hybrid modular and distributed organization. Cereb. Cortex 20, 257–270. doi:10.1093/cercor/bhp096 Briggs, F., Kiley, C.W., Callaway, E.M., Usrey, W.M., 2016. Morphological substrates for parallel streams of corticogeniculate feedback originating in both V1 and V2 of the macaque monkey. Neuron 90, 388–399. doi:10.1016/j.neuron.2016.02.038 Buzsáki, G., Schomburg, E.W., 2015. What does gamma coherence tell us about inter-regional neural communication? Nat. Neurosci. 18, 484–489. doi:10.1038/nn.3952 Casagrande, V.A., Kaas, J.H., 1994. The afferent, intrinsic, and efferent connections of primary visual cortex in primates, in: Peters, A., Rockland, K.S. (Eds.), Cerebral Cortex, Vol. 10: Primary Visual Cortex in Primates. Plenum Press, New York, NY, pp. 201–260. Cauli, B., Tong, X.-K., Rancillac, A., Serluca, N., Lambolez, B., Rossier, J., Hamel, E., 2004. Cortical GABA interneurons in neurovascular coupling: relays for subcortical vasoactive pathways. J. Neurosci. 24, 8940–8949. doi:10.1523/JNEUROSCI.3065-04.2004

19

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Chen, M., Yan, Y., Gong, X., Gilbert, C.D., Liang, H., Li, W., 2014. Incremental integration of global contours through interplay between visual cortical areas. Neuron 82, 682–694. doi:10.1016/j.neuron.2014.03.023 Cheng, K., Saleem, K.S., Tanaka, K., 1997. Organization of corticostriatal and corticoamygdalar projections arising from the anterior inferotemporal area TE of the macaque monkey: a Phaseolus vulgaris leucoagglutinin study. J. Neurosci. 17, 7902–7925. Cho, Y.T., Ernst, M., Fudge, J.L., 2013. Cortico-amygdala-striatal circuits are organized as hierarchical subsystems through the primate amygdala. J. Neurosci. 33, 14017–14030. doi:10.1523/JNEUROSCI.0170-13.2013 Christophel, T.B., Klink, P.C., Spitzer, B., Roelfsema, P.R., Haynes, J.-D., 2017. The distributed nature of working memory. Trends Cogn. Sci. (Regul. Ed.) 21, 111–124. doi:10.1016/j.tics.2016.12.007 Coogan, T.A., Burkhalter, A., 1993. Hierarchical organization of areas in rat visual cortex. J. Neurosci. 13, 3749–3772. da Costa, N.M., Martin, K.A.C., 2011. How thalamus connects to spiny stellate cells in the cat’s visual cortex. J. Neurosci. 31, 2925–2937. doi:10.1523/JNEUROSCI.5961-10.2011 DeFelipe, J., 2011. The evolution of the brain, the human nature of cortical circuits, and intellectual creativity. Front Neuroanat 5, 29. doi:10.3389/fnana.2011.00029 DeFelipe, J., 2015. The anatomical problem posed by brain complexity and size: a potential solution. Front Neuroanat 9, 104. doi:10.3389/fnana.2015.00104 DeFelipe, J., González-Albo, M.C., Del Río, M.R., Elston, G.N., 1999. Distribution and patterns of connectivity of interneurons containing calbindin, calretinin, and parvalbumin in visual areas of the occipital and temporal lobes of the macaque monkey. J. Comp. Neurol. 412, 515–526. Diester, I., Kaufman, M.T., Mogri, M., Pashaie, R., Goo, W., Yizhar, O., Ramakrishnan, C., Deisseroth, K., Shenoy, K.V., 2011. An optogenetic toolbox designed for primates. Nat. Neurosci. 14, 387–397. doi:10.1038/nn.2749 Ding, S.L., Van Hoesen, G., Rockland, K.S., 2000. Inferior parietal lobule projections to the presubiculum and neighboring ventromedial temporal cortical areas. J. Comp. Neurol. 425, 510–530. Erickson, S.L., Lewis, D.A., 2004. Cortical connections of the lateral mediodorsal thalamus in cynomolgus monkeys. J. Comp. Neurol. 473, 107–127. doi:10.1002/cne.20084 Falchier, A., Clavagnier, S., Barone, P., Kennedy, H., 2002. Anatomical evidence of multimodal integration in primate striate cortex. J. Neurosci. 22, 5749–5759. doi:20026562 Freese, J.L., Amaral, D.G., 2005. The organization of projections from the amygdala to visual cortical areas TE and V1 in the macaque monkey. J. Comp. Neurol. 486, 295–317. doi:10.1002/cne.20520 Freund, T.F., Martin, K.A., Soltesz, I., Somogyi, P., Whitteridge, D., 1989. Arborisation pattern and postsynaptic targets of physiologically identified thalamocortical afferents in striate cortex of the macaque monkey. J. Comp. Neurol. 289, 315–336. doi:10.1002/cne.902890211 Gabbott, P.L.A., 2016. “Subpial fan cell” - a class of calretinin neuron in layer 1 of adult monkey prefrontal cortex. Front Neuroanat 10, 28. doi:10.3389/fnana.2016.00028 Gabbott, P.L., Bacon, S.J., 1996. Local circuit neurons in the medial prefrontal cortex (areas 24a,b,c, 25 and 32) in the monkey: II. quantitative areal and laminar distributions. J. Comp.

20

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Neurol. 364, 609–636. doi:10.1002/(SICI)1096-9861(19960122)364:4<609::AIDCNE2>3.0.CO;2-7 Gabbott, P.L.A., Warner, T.A., Jays, P.R.L., Salway, P., Busby, S.J., 2005. Prefrontal cortex in the rat: projections to subcortical autonomic, motor, and limbic centers. J. Comp. Neurol. 492, 145–177. doi:10.1002/cne.20738 García-Marín, V., Blazquez-Llorca, L., Rodriguez, J.R., Gonzalez-Soriano, J., DeFelipe, J., 2010. Differential distribution of neurons in the gyral white matter of the human cerebral cortex. J. Comp. Neurol. 518, 4740–4759. doi:10.1002/cne.22485 Girard, P., Salin, P.A., Bullier, J., 1992. Response selectivity of neurons in area MT of the macaque monkey during reversible inactivation of area V1. J. Neurophysiol. 67, 1437– 1446. Goense, J., Bohraus, Y., Logothetis, N.K., 2016. fMRI at high spatial resolution: implications for BOLD-models. Front Comput Neurosci 10, 66. doi:10.3389/fncom.2016.00066 Gómez-Laberge, C., Smolyanskaya, A., Nassi, J.J., Kreiman, G., Born, R.T., 2016. Bottom-up and top-down input augment the variability of cortical neurons. Neuron 91, 540–547. doi:10.1016/j.neuron.2016.06.028 Grayson, D.S., Bliss-Moreau, E., Machado, C.J., Bennett, J., Shen, K., Grant, K.A., Fair, D.A., Amaral, D.G., 2016. The rhesus monkey connectome predicts disrupted functional networks resulting from pharmacogenetic inactivation of the amygdala. Neuron 91, 453– 466. doi:10.1016/j.neuron.2016.06.005 Haber, S.N., 2016. Corticostriatal circuitry. Dialogues Clin Neurosci 18, 7–21. Hackett, T.A., de la Mothe, L.A., Camalier, C.R., Falchier, A., Lakatos, P., Kajikawa, Y., Schroeder, C.E., 2014. Feedforward and feedback projections of caudal belt and parabelt areas of auditory cortex: refining the hierarchical model. Front Neurosci 8, 72. doi:10.3389/fnins.2014.00072 Haegens, S., Barczak, A., Musacchia, G., Lipton, M.L., Mehta, A.D., Lakatos, P., Schroeder, C.E., 2015. Laminar profile and physiology of the α rhythm in primary visual, auditory, and somatosensory regions of neocortex. J. Neurosci. 35, 14341–14352. doi:10.1523/JNEUROSCI.0600-15.2015 Han, X., Chow, B.Y., Zhou, H., Klapoetke, N.C., Chuong, A., Rajimehr, R., Yang, A., Baratta, M.V., Winkle, J., Desimone, R., Boyden, E.S., 2011. A high-light sensitivity optical neural silencer: development and application to optogenetic control of non-human primate cortex. Front Syst Neurosci 5, 18. doi:10.3389/fnsys.2011.00018 Höistad, M., Barbas, H., 2008. Sequence of information processing for emotions through pathways linking temporal and insular cortices with the amygdala. Neuroimage 40, 1016– 1033. doi:10.1016/j.neuroimage.2007.12.043 Huber, L., Goense, J., Kennerley, A.J., Ivanov, D., Krieger, S.N., Lepsien, J., Trampel, R., Turner, R., Möller, H.E., 2014. Investigation of the neurovascular coupling in positive and negative BOLD responses in human brain at 7 T. Neuroimage 97, 349–362. doi:10.1016/j.neuroimage.2014.04.022 Ichinohe, N., Matsushita, A., Ohta, K., Rockland, K.S., 2010. Pathway-specific utilization of synaptic zinc in the macaque ventral visual cortical areas. Cereb. Cortex 20, 2818–2831. doi:10.1093/cercor/bhq028 Insausti, R., Amaral, D.G., 2008. Entorhinal cortex of the monkey: IV. topographical and laminar organization of cortical afferents. J. Comp. Neurol. 509, 608–641. doi:10.1002/cne.21753

21

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Jones, E.G., 1986. Connectivity of the primate sensory-motor cortex, in: Jones, E.G., Peters, A. (Eds.), Cerebral Cortex, Vol. 5: Sensory-Motor Areas and Aspects of Cortical Connectivity. Plenum Press, New York, NY, pp. 113–183. Karube, F., Sári, K., Kisvárday, Z.F., 2016. Axon topography of layer 6 spiny cells to orientation map in the primary visual cortex of the cat (area 18). Brain Struct Funct 222, 1401–1426. doi:10.1007/s00429-016-1284-z Katz, L.C., 1987. Local circuitry of identified projection neurons in cat visual cortex brain slices. J. Neurosci. 7, 1223–1249. Kennedy, H., Bullier, J., 1985. A double-labeling investigation of the afferent connectivity to cortical areas V1 and V2 of the macaque monkey. J. Neurosci. 5, 2815–2830. Kisvárday, Z.F., 1992. GABAergic networks of basket cells in the visual cortex. Prog. Brain Res. 90, 385–405. Koestinger, G., Martin, K.A.C., Roth, S., Rusch, E.S., 2017. Synaptic connections formed by patchy projections of pyramidal cells in the superficial layers of cat visual cortex. Brain Struct Funct. doi:10.1007/s00429-017-1384-4 Kultas-Ilinsky, K., Sivan-Loukianova, E., Ilinsky, I.A., 2003. Reevaluation of the primary motor cortex connections with the thalamus in primates. J. Comp. Neurol. 457, 133–158. doi:10.1002/cne.10539 Laramée, M.-E., Boire, D., 2014. Visual cortical areas of the mouse: comparison of parcellation and network structure with primates. Front Neural Circuits 8, 149. doi:10.3389/fncir.2014.00149 Larkum, M.E., 2013. The yin and yang of cortical layer 1. Nat. Neurosci. 16, 114–115. doi:10.1038/nn.3317 Lavenex, P., Suzuki, W.A., Amaral, D.G., 2004. Perirhinal and parahippocampal cortices of the macaque monkey: Intrinsic projections and interconnections. J. Comp. Neurol. 472, 371– 394. doi:10.1002/cne.20079 Lee, A.J., Wang, G., Jiang, X., Johnson, S.M., Hoang, E.T., Lanté, F., Stornetta, R.L., Beenhakker, M.P., Shen, Y., Julius Zhu, J., 2015. Canonical organization of layer 1 neuron-led cortical inhibitory and disinhibitory interneuronal circuits. Cereb. Cortex 25, 2114–2126. doi:10.1093/cercor/bhu020 Levitt, J.B., Yoshioka, T., Lund, J.S., 1995. Connections between the pulvinar complex and cytochrome oxidase-defined compartments in visual area V2 of macaque monkey. Exp Brain Res 104, 419–430. Luebke, J.I., 2017. Pyramidal Neurons Are Not Generalizable Building Blocks of Cortical Networks. Front Neuroanat 11, 11. doi:10.3389/fnana.2017.00011 Lund, J.S., 1988. Anatomical organization of macaque monkey striate visual cortex. Annu. Rev. Neurosci. 11, 253–288. doi:10.1146/annurev.ne.11.030188.001345 Lund, J.S., Yoshioka, T., 1991. Local circuit neurons of macaque monkey striate cortex: III. neurons of laminae 4B, 4A, and 3B. J. Comp. Neurol. 311, 234–258. doi:10.1002/cne.903110206 Lund, J.S., Hendrickson, A.E., Ogren, M.P., Tobin, E.A., 1981. Anatomical organization of primate visual cortex area VII. J. Comp. Neurol. 202, 19–45. doi:10.1002/cne.902020104 Markov, N.T., Vezoli, J., Chameau, P., Falchier, A., Quilodran, R., Huissoud, C., Lamy, C., Misery, P., Giroud, P., Ullman, S., Barone, P., Dehay, C., Knoblauch, K., Kennedy, H., 2014. Anatomy of hierarchy: feedforward and feedback pathways in macaque visual cortex. J. Comp. Neurol. 522, 225–259. doi:10.1002/cne.23458

22

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Martin, K.A.C., Roth, S., Rusch, E.S., 2014. Superficial layer pyramidal cells communicate heterogeneously between multiple functional domains of cat primary visual cortex. Nat Commun 5, 5252. doi:10.1038/ncomms6252 Mejias, J.F., Murray, J.D., Kennedy, H., Wang, X.-J., 2016. Feedforward and feedback frequency-dependent interactions in a large-scale laminar network of the primate cortex. Sci Adv 2, e1601335. doi:10.1126/sciadv.1601335 Melchitzky, D.S., Sesack, S.R., Pucak, M.L., Lewis, D.A., 1998. Synaptic targets of pyramidal neurons providing intrinsic horizontal connections in monkey prefrontal cortex. J. Comp. Neurol. 390, 211–224. Melchitzky, D.S., Eggan, S.M., Lewis, D.A., 2005. Synaptic targets of calretinin-containing axon terminals in macaque monkey prefrontal cortex. Neuroscience 130, 185–195. doi:10.1016/j.neuroscience.2004.08.046 Meskenaite, V., 1997. Calretinin-immunoreactive local circuit neurons in area 17 of the cynomolgus monkey, Macaca fascicularis. J. Comp. Neurol. 379, 113–132. Michalareas, G., Vezoli, J., van Pelt, S., Schoffelen, J.-M., Kennedy, H., Fries, P., 2016. Alphabeta and gamma rhythms subserve feedback and feedforward influences among human visual cortical areas. Neuron 89, 384–397. doi:10.1016/j.neuron.2015.12.018 Mitrofanis, J., 2005. Some certainty for the “zone of uncertainty”? Exploring the function of the zona incerta. Neuroscience 130, 1–15. doi:10.1016/j.neuroscience.2004.08.017 Miyashita, T., Ichinohe, N., Rockland, K.S., 2007. Differential modes of termination of amygdalothalamic and amygdalocortical projections in the monkey. J. Comp. Neurol. 502, 309–324. doi:10.1002/cne.21304 Mohedano-Moriano, A., Muñoz-López, M., Sanz-Arigita, E., Pró-Sistiaga, P., Martínez-Marcos, A., Legidos-Garcia, M.E., Insausti, A.M., Cebada-Sánchez, S., Arroyo-Jiménez, M.D.M., Marcos, P., Artacho-Pérula, E., Insausti, R., 2015. Prefrontal cortex afferents to the anterior temporal lobe in the Macaca fascicularis monkey. J. Comp. Neurol. 523, 2570–2598. doi:10.1002/cne.23805 Morecraft, R.J., McNeal, D.W., Stilwell-Morecraft, K.S., Gedney, M., Ge, J., Schroeder, C.M., van Hoesen, G.W., 2007. Amygdala interconnections with the cingulate motor cortex in the rhesus monkey. J. Comp. Neurol. 500, 134–165. doi:10.1002/cne.21165 Mortazavi, F., Wang, X., Rosene, D.L., Rockland, K.S., 2016. White matter neurons in young adult and aged rhesus monkey. Front Neuroanat 10, 15. doi:10.3389/fnana.2016.00015 Nakamura, H., Gattass, R., Desimone, R., Ungerleider, L.G., 1993. The modular organization of projections from areas V1 and V2 to areas V4 and TEO in macaques. J. Neurosci. 13, 3681–3691. Nassi, J.J., Cepko, C.L., Born, R.T., Beier, K.T., 2015. Neuroanatomy goes viral! Front Neuroanat 9, 80. doi:10.3389/fnana.2015.00080 Ojima, H., Honda, C.N., Jones, E.G., 1991. Patterns of axon collateralization of identified supragranular pyramidal neurons in the cat auditory cortex. Cereb. Cortex 1, 80–94. Palmer, S.M., Rosa, M.G.P., 2006. A distinct anatomical network of cortical areas for analysis of motion in far peripheral vision. Eur. J. Neurosci. 24, 2389–2405. doi:10.1111/j.14609568.2006.05113.x Payne, B.R., Lomber, S.G., Villa, A.E., Bullier, J., 1996. Reversible deactivation of cerebral network components. Trends Neurosci. 19, 535–542. Raichle, M.E., 2015. The brain’s default mode network. Annu. Rev. Neurosci. 38, 433–447. doi:10.1146/annurev-neuro-071013-014030

23

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Reveley, C., Seth, A.K., Pierpaoli, C., Silva, A.C., Yu, D., Saunders, R.C., Leopold, D.A., Ye, F.Q., 2015. Superficial white matter fiber systems impede detection of long-range cortical connections in diffusion MR tractography. Proc. Natl. Acad. Sci. U.S.A. 112, E2820-2828. doi:10.1073/pnas.1418198112 Rockland, K.S., 1989. Bistratified distribution of terminal arbors of individual axons projecting from area V1 to middle temporal area (MT) in the macaque monkey. Vis. Neurosci. 3, 155–170. Rockland, K.S., 1996. Two types of corticopulvinar terminations: round (type 2) and elongate (type 1). J. Comp. Neurol. 368, 57–87. doi:10.1002/(SICI)10969861(19960422)368:1<57::AID-CNE5>3.0.CO;2-J Rockland, K.S., 1997. Elements of cortical architecture: hierarchy revisited, in: Rockland, K.S., Kaas, J.H., Peters, A. (Eds.), Cerebral Cortex, Vol. 12: Extrastriate Cortex in Primates. Plenum Press, New York, NY, pp. 243–293. Rockland, K.S., 2015. About connections. Front Neuroanat 9, 61. doi:10.3389/fnana.2015.00061 Rockland, K.S., 2016. Anatomy of the cerebral cortex, in: Cechetto, D., Weishaupt, N. (Eds.), The Cerebral Cortex in Neurodegenerative and Neuropsychiatric Disorders. Academic Press, Cambridge, MA, pp. 3–36. Rockland, K.S., Pandya, D.N., 1979. Laminar origins and terminations of cortical connections of the occipital lobe in the rhesus monkey. Brain Res. 179, 3–20. Rockland, K.S., Lund, J.S., 1983. Intrinsic laminar lattice connections in primate visual cortex. J. Comp. Neurol. 216, 303–318. doi:10.1002/cne.902160307 Rockland, K.S., Virga, A., 1989. Terminal arbors of individual “feedback” axons projecting from area V2 to V1 in the macaque monkey: a study using immunohistochemistry of anterogradely transported Phaseolus vulgaris-leucoagglutinin. J. Comp. Neurol. 285, 54– 72. doi:10.1002/cne.902850106 Rockland, K.S., Knutson, T., 2000. Feedback connections from area MT of the squirrel monkey to areas V1 and V2. J. Comp. Neurol. 425, 345–368. Rockland, K.S., Knutson, T., 2001. Axon collaterals of Meynert cells diverge over large portions of area V1 in the macaque monkey. J. Comp. Neurol. 441, 134–147. Rockland, K.S., Ojima, H., 2003. Multisensory convergence in calcarine visual areas in macaque monkey. Int J Psychophysiol 50, 19–26. Rockland, K.S., Nayyar, N., 2012. Association of type I neurons positive for NADPH-diaphorase with blood vessels in the adult monkey corpus callosum. Front Neural Circuits 6, 4. doi:10.3389/fncir.2012.00004 Rockland, K.S., Saleem, K.S., Tanaka, K., 1994. Divergent feedback connections from areas V4 and TEO in the macaque. Vis. Neurosci. 11, 579–600. Rockland, K.S., Andresen, J., Cowie, R.J., Robinson, D.L., 1999. Single axon analysis of pulvinocortical connections to several visual areas in the macaque. J. Comp. Neurol. 406, 221–250. Rouiller, E.M., Welker, E., 2000. A comparative analysis of the morphology of corticothalamic projections in mammals. Brain Res. Bull. 53, 727–741. Saleem, K.S., Suzuki, W., Tanaka, K., Hashikawa, T., 2000. Connections between anterior inferotemporal cortex and superior temporal sulcus regions in the macaque monkey. J. Neurosci. 20, 5083–5101.

24

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Schmid, M.C., Panagiotaropoulos, T., Augath, M.A., Logothetis, N.K., Smirnakis, S.M., 2009. Visually driven activation in macaque areas V2 and V3 without input from the primary visual cortex. PLoS ONE 4, e5527. doi:10.1371/journal.pone.0005527 Self, M.W., Roelfsema, P.R., 2017. Paying attention to the cortical layers. Neuron 93, 9–11. doi:10.1016/j.neuron.2016.12.032 Shepherd, G.M.G., 2013. Corticostriatal connectivity and its role in disease. Nat. Rev. Neurosci. 14, 278–291. doi:10.1038/nrn3469 Sherman, S.M., 2017. Functioning of circuits connecting thalamus and cortex. Compr Physiol 7, 713–739. doi:10.1002/cphy.c160032 Shipp, S., 2003. The functional logic of cortico-pulvinar connections. Philos. Trans. R. Soc. Lond., B, Biol. Sci. 358, 1605–1624. doi:10.1098/rstb.2002.1213 Shipp, S., 2007. Structure and function of the cerebral cortex. Curr. Biol. 17, R443-449. doi:10.1016/j.cub.2007.03.044 Shipp, S., 2016. Neural elements for predictive coding. Front Psychol 7, 1792. doi:10.3389/fpsyg.2016.01792 Shipp, S., 2017. The functional logic of corticostriatal connections. Brain Struct Funct 222, 669– 706. doi:10.1007/s00429-016-1250-9 Shipp, S., Zeki, S., 1989. The organization of connections between areas V5 and V1 in macaque monkey visual cortex. Eur. J. Neurosci. 1, 309–332. Shipp, S., Adams, R.A., Friston, K.J., 2013. Reflections on agranular architecture: predictive coding in the motor cortex. Trends Neurosci. 36, 706–716. doi:10.1016/j.tins.2013.09.004 Sincich, L.C., Horton, J.C., 2003. Independent projection streams from macaque striate cortex to the second visual area and middle temporal area. J. Neurosci. 23, 5684–5692. Stefanacci, L., Amaral, D.G., 2000. Topographic organization of cortical inputs to the lateral nucleus of the macaque monkey amygdala: a retrograde tracing study. J. Comp. Neurol. 421, 52–79. Stettler, D.D., Das, A., Bennett, J., Gilbert, C.D., 2002. Lateral connectivity and contextual interactions in macaque primary visual cortex. Neuron 36, 739–750. Thomson, A.M., 2010. Neocortical layer 6, a review. Front Neuroanat 4, 13. doi:10.3389/fnana.2010.00013 Timbie, C., Barbas, H., 2014. Specialized pathways from the primate amygdala to posterior orbitofrontal cortex. J. Neurosci. 34, 8106–8118. doi:10.1523/JNEUROSCI.5014-13.2014 Ungerleider, L.G., Desimone, R., 1986. Cortical connections of visual area MT in the macaque. J. Comp. Neurol. 248, 190–222. doi:10.1002/cne.902480204 Ungerleider, L.G., Galkin, T.W., Desimone, R., Gattass, R., 2008. Cortical connections of area V4 in the macaque. Cereb. Cortex 18, 477–499. doi:10.1093/cercor/bhm061 Wang, X.-J., Kennedy, H., 2016. Brain structure and dynamics across scales: in search of rules. Curr. Opin. Neurobiol. 37, 92–98. doi:10.1016/j.conb.2015.12.010 Watakabe, A., Ichinohe, N., Ohsawa, S., Hashikawa, T., Komatsu, Y., Rockland, K.S., Yamamori, T., 2007. Comparative analysis of layer-specific genes in mammalian neocortex. Cereb. Cortex 17, 1918–1933. doi:10.1093/cercor/bhl102 Welker, W., 1990. Why does cerebral cortex fissure and fold? A review of determinants of gyri and sulci, in: Jones, E.G., Peters, A. (Eds.), Comparative Structure and Evolution of Cerebral Cortex. Plenum Press, New York, NY, pp. 3–136.

25

ACCEPTED MANUSCRIPT

RI PT

Xiao, D., Zikopoulos, B., Barbas, H., 2009. Laminar and modular organization of prefrontal projections to multiple thalamic nuclei. Neuroscience 161, 1067–1081. doi:10.1016/j.neuroscience.2009.04.034 Yoshimura, Y., Dantzker, J.L.M., Callaway, E.M., 2005. Excitatory cortical neurons form finescale functional networks. Nature 433, 868–873. doi:10.1038/nature03252 Zeki, S., 2016. Multiple asynchronous stimulus- and task-dependent hierarchies (STDH) within the visual brain’s parallel processing systems. Eur. J. Neurosci. 44, 2515–2527. doi:10.1111/ejn.13270 Figure Legends

SC

Figure 1. A generic summary of inputs, outputs, and intrinsic excitatory connections at the laminar level, for a nonvisual area of NHP cerebral cortex (reproduced with permission from Shipp, 2007, see references).

M AN U

Figure 2. Schematic summary illustrating the major targets of pyramidal cells in different layers (monkey primary somatosensory cortex). Note that subpopulations of neurons in the infragranular layers project cortically as well as subcortically (Reproduced from Jones, 1986; see references).

TE D

Figure 3. A) and, progressively higher magnification (from double asterisk), C) and D) Projections from temporal cortex to the depth of the superior temporal sulcus, anterogradely labeled in Golgi-like detail by BDA. At the higher magnification in D), terminal boutons and segments of preterminal axons are readily identified. Note also that terminations extend into lower layer 3. B) For comparison, a closely adjacent Nissl-stained section. Since this is at the junction of the lower bank and sulcal depth, the section also serves to demonstrate the change in laminar proportions in the sulcal depth; namely, expanded layer 1, and compressed layers 4, 5, and 5. L. 4 = layer 4. Scale bar = 200µm (A and B), 100µm (C), 50µm (D).

AC C

EP

Figure 4. Pyramidal neurons in anterior temporal (aTE) cortex retrogradely labeled by injection in posterior temporal cortex of a Golgi-like tracer (adenoviral vector expressing EGFP under the control of a neuron-specific promoter, synapsin I). Labeled neurons are mainly in layers 3 and 5 (that is, from this association area, “deviant” from the classic feedback in early sensory areas). Those deeper in layer 3 (solid arrow) extend basal dendrites into layer 4, but not those located more superfically in layer 3 (hollow arrow). Inset is at higher magnification, from solid arrow. Dashed lines denote border between gray and white matter (WM). L. 4 = layer 4. Scale bar = 250µm (125µm for higher magnification inset) Figure 5. Intracellularly injected pyramidal cell (soma, in red) at the border of layers 2 and 3 of cat auditory cortex. Schematic representation of a) intrinsic axonal collaterals (1-7) in a coronal (radial) plane of section and b) bouton distribution, as projected onto a tangential plane. Scale bars = 500µm. Reproduced with permission from Oxford University Press, Ojima et al., 1991 (see references). Figure 6. Intracellularly injected pyramidal cell (soma and dendrites in red in A) with abundant intrinsic collaterals (cat visual cortex). A: coronal view. B: Tangential view to demonstrate the

26

ACCEPTED MANUSCRIPT

“spokelike” pattern of intrinsic collaterals, here shown in relation to optically imaged orientation domains. Scale bars = 500µm. Reproduced with permission from Brain Structure Function, Koestinger et al., 2017 (see references).

RI PT

Figure 7. Diagram summarizing the interneuron varieties associated with laminae 4A and 3B as described in a Golgi study of area V1 (NHP). Note diversity and different combinations of interlaminar innervations. Reproduced with permission from Wiley Online Library, Lund and Yoshioka, 1991 (see references).

SC

Figure 8. Diagram summarizing a schematic microcircuit for feedback connections to area V1. Note that feedback connections from area V2 and other extrastriate areas potentially contact infragranular neurons at multiple sites (yellow), including basal dendrites in layers 5 or 6 and distal dendrites in layer 1. Red = approximate sites of geniculocortical terminations. Reproduced with permission from Wiley Online Library, Rockland and Virga, 1989 (see references).

M AN U

Figure 9. Summary diagram of the laminar organization of inputs to NHP entorhinal cortex (subfields indicated at top, layers at the left, and eight anterogradely labeled regions at bottom (OF, orbitofrontal; MF, medial prefrontal; I, insular; STS, superior temporal sulcus; PR, perirhinal; PH, parahippocampal; AC, anterior cingulated; RSP, retrosplenial cortex). Colors denote intricate mulit-laminar pattern (not necessarily stratification), as revealed by anterograde tracer injections in the several source regions. Reproduced with permission from Wiley Online Library, Insausti and Amaral, 2008 (see references).

EP

TE D

Figure 10. Dendrites cross freely in both directions over the border between layer 6 and the white matter (WM). A) and (higher magnification) B) Neurons in the superficial WM labeled by the enzyme nicotinamide adenine dinucleotide phosphate diaphorase (NADPH-d) extend dendrites (double asterisk) into the overlying gray matter. C) and D) Golgi-like labeling with a retrogradely transported adenoviral vector (see Fig. 4) shows that neurons in layer 5 can extend dendrites into the underlying WM. These are neurons in ventral temporal cortex, labeled by an injection in lateral surface area TE. Inset in C) is higher magnification, from the double asterisk, to show detail. E) is higher magnification (from the double asterisk in D) to show dendritic spines along the WM portion (single asterisk). Dashed lines in C) and D) indicate upper border of WM. Scale bar = 100µm (A-D), 50µm (E).

AC C

Figure 11.Summary of laminar terminations of anterogradely labeled connections within the auditory pathway. (Reproduced from Frontiers in Neuroscience, Hackett et al., 2014; see references). Figure 12. A), and higher magnification, B) (from the single asterisk) and C) (from the double asterisk) of feedback neurons in area V2 labeled in Golgi-like detail by a retrogradely transported adenovirus (see Fig. 3). Neurons are in upper layer 3 and layer 6. Images in B) and C) have been rotated, so that the pia surface is at the top. Arrow points to the V1/V2 border. LS = lunate sulcus. Scale bar = 500µm (A), 100µm (B, C).

27

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Figure 13. Retrogradely labeled neurons in aTE (same case as illustrated in Fig. 4). As a result of the sulcal deformation, apical dendrites converge (*) and presumably share common inputs, even through parent somas (hollow arrows) are spatially non-adjacent. Scale bar = 125µm

28