Cortical pyramidal neuronal processes are present in each of the 6 neocortical layers, with somata typically localized to layers 2–6 (except in layer 4c of the primary visual cortex, which is comprised exclusively of spiny stellate cells). The somata of pyramidal neurons are typically triangular, with a broad base from which a single axon and a skirt of basilar dendrites emanate and an apex from which, most typically, a single apical trunk projects. The apical dendrite has three compartments –a main trunk, oblique branches, a tuft that ramifies in layer 1- each of which possesses unique structural, connectional, and functional characteristics which broaden the dynamic range of signal integration by the apical dendritic arbor as a whole. The different dendritic compartments receive and integrate distinct presynaptic excitatory and inhibitory inputs and possess distinct passive and active signal filtering and boosting capacities (Larkum et al., 2001, 2009; London and Hausser, 2005; Losonczy and Magee, 2006; Losonczy et al., 2008; reviews: Spruston, 2008; Kubota et al., 2015, 2016). Differences in the lengths, diameters, and branching pattern of the dendritic arbor confer significant variability in cable properties and therefore the spatial distribution of electrical signals and degree of summation of synaptic inputs, which determine the temporal pattern of both forward and backward propagating action potentials (Mainen and Sejnowski, 1996; Koch and Segev, 2000; Euler and Denk, 2001; Vetter et al., 2001; Krichmar et al., 2002; Ascoli, 2003; reviews: Stuart et al., 1997; Waters et al., 2005). Thus, by virtue of their different somatodendritic compartments, pyramidal neurons act as coincidence detectors possessing a wide dynamic range for integration of temporally and spatially unique synaptic signals. Computational modeling studies suggest that even minor differences in branching characteristics can exert a major influence on signal processing by neurons. For example, even modest variations in branch point angles can transform the electrical coupling between oblique dendrites and the main apical shaft dendrite from fully coupled to fully compartmentalized (Ferrante et al., 2013).

Integration of synaptic inputs is also significantly shaped by active properties, including the number and distribution of a wide variety of transmembrane ion channels (reviews: Migliore and Shepherd, 2002; Magee and Johnston, 2005; Johnston and Narayanan, 2008). Over 20 different types of sodium, calcium, and potassium channels are distributed -some uniformly and some non-uniformly- across a given dendrite and confer distinct boosting and/or dampening of local signals (reviews: Migliore and Shepherd, 2002; Magee and Johnston, 2005; Johnston and Narayanan, 2008; Nusser, 2012). The complex interplay of intrinsic ionic and synaptic conductances with passive properties determined by dendritic morphology can effectively alter the cable properties of the dendritic tree (Segev and London, 2000; Bekkers and Häusser, 2007; review: Nusser, 2012) resulting in a variable and finely-tunable integrative and signaling capacity in pyramidal neurons.

Dendritic spines -principal recipients of glutamatergic synapses- also play a key role in the electrical and biochemical signaling in dendrites (reviews: Matus and Shepherd, 2000; Nimchinsky et al., 2002; Kasai et al., 2003; Bourne and Harris, 2008). While there is a continuum of spine morphology at steady state, and morphology can vary dynamically in response to synaptic activity (Lendvai et al., 2000; Zuo et al., 2005; for review: Lüscher et al., 2000; Wefelmeyer et al., 2016), spines can be broadly classified as being “thin,” “stubby,” “mushroom,” or “filopodia” (review: Bourne and Harris, 2008). Just as with dendrites, spine structural properties underlie functional physiological signaling; thus spine and synapse structure is largely determinative of the strength, stability and function of excitatory glutamatergic synapses (Tong and Jahr, 1994; Baude et al., 1995; Murthy et al., 1997, 2001; Nusser et al., 1998; Matsuzaki et al., 2001; Li et al., 2005; Germuska et al., 2006). Thus, quantification of the distribution of spine subtypes as well as of synapses on pyramidal neurons is essential for understanding the integrative capacities of these neurons in distinct brain areas and species.

Dendritic arbor size and spine density on pyramidal neurons differ markedly across functionally distinct cortical areas in the rhesus monkey and the human brain (Cajal SRy, 1894, 1995; Conel, 1941, 1967; DeFelipe et al., 2002; Jacobs, 2002; Elston, 2003; Elston and Fujita, 2014; Mohan et al., 2015). In the rhesus monkey, the size, and complexity of the dendritic arbors of L3 pyramidal neurons increases dramatically from primary visual cortex (V1) to higher-order lateral prefrontal cortex, and this increase in overall size is accompanied by a significantly higher spine density (Elston, 2000, 2002, 2003; Elston et al., 2001; Amatrudo et al., 2012). In addition to this caudal to rostral gradient observed in the primate, pyramidal neurons in some analogous cortical areas increase in size or “scale” from the rodent to the macaque (Elston and DeFelipe, 2002; Elston and Zeitsch, 2005; Ballesteros-Yanez et al., 2006; Elston, 2007; Wen et al., 2009; Elston and Manger, 2014; reviews: Wittenberg and Wang, 2008; DeFelipe, 2011).

To gain a deeper understanding of comparative morphological features of cortical neurons in rodents and primates, as well as their functional relevance, we have used in vitro whole-cell patch-clamp recordings and cell filling in a series of systematic studies to characterize the detailed structural, neuro-chemical and functional properties of L3 pyramidal neurons in the primary visual and frontal association cortices of mice and of rhesus monkeys (Amatrudo et al., 2012; Luebke et al., 2015; Medalla and Luebke, 2015; Gilman et al., 2016; Hsu et al., in press). A unique feature of these studies, which are summarized below, is that neurons were assessed both morphologically and physiologically at high resolution and in an identical manner across different brain areas in the two species allowing direct and comprehensive comparisons.

While frontal cortical L3 pyramidal neurons scale significantly in size from the mouse to the monkey, no such scaling exists with L3 pyramidal neurons in V1 (Gilman et al., 2016). That dendritic scaling occurs in frontal but not visual L3 pyramidal neurons provides interesting insight into potentially differential capabilities of these neurons in the two species. As frontal cortical pyramidal neurons increase in size across phylogeny, the opportunity for convergence of diverse inputs is increased, as is their integrative and computational dynamic range. As discussed above, integration and filtering of input signals occurs as a function of number of dendritic branch points and both diameter and the geometric features –notably length and diameter- of dendritic segments (Rall, 1962, 1964). Scaling and cable theory predict that monkey LPFC neurons filter input signals to a greater extent than mouse FC neurons due to their greater dendritic length and equivalent dendritic diameters. Consistent with this, there is a higher frequency of spontaneous EPSCs in mouse FC vs. monkey LPFC (Gilman et al., 2016). On the other hand, action potential firing and other intrinsic properties are largely preserved between mouse and monkey frontal neurons, suggesting significant roles for non-passive properties (e.g., active conductances and synaptic inputs; Nusser, 2012) that should be examined in future studies. V1 neurons in the monkey possess the lowest number and density of spines while spine densities do not significantly differ between monkey LPFC and mouse FC and mouse V1 neurons, though given their much smaller size, the mean numbers of spines on pyramidal neurons in the mouse is much lower than in monkey LPFC (Gilman et al., 2016). The ultrastructural properties of excitatory synapses vary across species in that the L2-3 neuropil of monkey LPFC contains a significantly higher proportion of perforated postsynaptic densities and, on average, larger spine volume compared to L2-3 neuropil of mouse FC. Finally, both presynaptic and postsynaptic entities are significantly smaller in the mouse FC than in the monkey LPFC.

L3 pyramidal neurons in mouse FC and V1 are virtually identical in their dendritic, spine, and excitatory synapse structure as well as in their physiological properties. These findings are in line with the high degree of cytoarchitectural and functional homogeneity across mouse cortical areas, compared to the highly specialized cortical areas of the primate brain. The similarity in the structural and biophysical properties of mouse V1 and FC neurons suggests that relatively similar temporal signaling dynamics may exist within these areas. In mice, both the primary sensory area V1 and the multimodal FC contain cellular and synaptic features consistent with a highly excitable circuit, being comprised of small and electrically compact output neurons, and abundant spines with relatively small excitatory synapses. Thus, in broad terms mouse cortical neurons (and presumably the dynamic networks of which they are a part) are well suited for rapid synaptic transmission with a high degree of input-output fidelity but relatively low dynamic range (review: Olshausen and Field, 2004; Vogels et al., 2005; Panzeri et al., 2010).

In contrast to the small and much less differentiated rodent neocortex, a larger and more specialized brain, such as that of the rhesus monkey, requires functionally distinct cortical areas to have different levels of excitability, filtering, and integration of inputs (Luebke et al., 2010; Barbas, 2015). The preponderance of morphological and electrophysiological data predict that in the monkey, synaptic integration at the cellular and network levels differ between V1, a primary sensory area for unimodal representation, and LPFC, a high-order area for complex multimodal processing (review: Schummers et al., 2004; Fuster, 2015). V1 neurons are small, compact and highly excitable, properties that enable them to respond optimally to small, fast synaptic inputs and for building a network with a limited dynamic range but well-suited for signal transformations with relatively high input-output fidelity (review: Olshausen and Field, 2004; Vogels et al., 2005; Panzeri et al., 2010). Compared to the primary sensory V1, the multimodal association LPFC in monkey is comprised of cellular and synaptic features—large and electrotonically complex L3 output neurons with many spines- consistent with more powerful and longer-lasting inputs. These features are optimal for facilitating sustained activation, coincidence detection, and spike-timing-dependent plasticity, all important for integrative functions such as decision making and integration of sensorimotor information (review: Constantinidis and Wang, 2004; Sjostrom et al., 2008). The function of the LPFC is to integrate multimodal information from a wide array of cortical and subcortical afferents in order to perform sophisticated executive tasks (review: Miller and Cohen, 2001; Luebke et al., 2010). A relatively larger dynamic range of integration of information conferred by larger neurons with more numerous synapses is required in a high-order area such as LPFC, while it would be disadvantageous in V1 where more rapid signal transformations are required.

In terms of their fundamental structural properties -dendrite, spine, and synapse morphology- there are some striking and many subtle differences between L3 pyramidal neurons in the mouse and the rhesus monkey and between cortical areas in the rhesus monkey but not in the mouse. In the mouse, where L3 pyramidal neurons are structurally the same in these two brain areas, a uniform prototypical cortical pyramidal neuron may be generalizable from one area to another, at least in terms of size, dendritic structure, and intrinsic membrane, synaptic, and action potential firing properties. In the rhesus monkey such a prototypical neuron does not exist- cortical areas differ markedly from each other at the individual pyramidal cell and network levels. Data such as those summarized here are important for understanding how signaling within neuronal networks differs between rodents and primates and for how these neurons and networks may contribute to species-specific functional capacities. These findings however do not answer the question of which, if any, of these particular differences are necessary and sufficient for differentiating neuronal network behavior in different brain areas and species. The answer to this question remains to be determined and is difficult to predict, particularly in light of in silico predictions that neuronal networks display emergent behavior that may not depend on details of individual neuron structure and function (review: Yuste, 2015). For now, since we do not know which of myriad details about individual neurons are key for network function, the tendency for premature simplification should be avoided (DeFelipe et al., 2016). Our understanding of neuronal diversity in all of its complexity is nascent, but thanks to advances in molecular, genetic and neuroanatomical tools we are on the verge of a new era in which the great diversity of neuronal types will be cataloged and lead to more nuanced and comprehensive insights into the mechanisms of cortical areal specialization.

The author confirms being the sole contributor of this work and approved it for publication.