Brain samples from the left hemisphere of 46 individuals representing eight haplorhines species were used for this study and included New World monkeys (black-capped squirrel monkeys and tufted capuchins), Old World monkeys (moor macaques, pigtailed macaques, and olive baboons), African great apes (western lowland gorillas and common chimpanzees), and humans (see Table 1 for details). Brain weight was recorded for every specimen either at the time of necropsy or shortly afterwards. All individuals were adults and had no evidence of neuropathological alterations based on inspection of gross specimens. Sexes were balanced within species as much as possible given the opportunistic nature of brain collection. Human brain samples were provided by the Northwestern University Alzheimer's Disease Center Brain Bank. The human subjects exhibited no evidence of cognitive changes before death and all received a score of zero for the CERAD senile plaque grade (Mirra et al., 1991) and the Braak and Braak neurofibrillary tangle stage (Braak and Braak, 1991). Non-human primate brains were acquired from American Zoo and Aquarium-accredited zoos or research institutions and were maintained in accordance with each institution's animal care and use guidelines. All brains except those of the moor macaques were immersion-fixed in 10% buffered formalin for at least 7 days, transferred to a 0.1 M phosphate buffered saline solution containing 0.1% sodium azide, and stored at 4°C until processing. The moor macaques were perfused transcardially with 4% paraformaldehyde as part of unrelated experiments (Duan et al., 2003). Postmortem interval (PMI) did not exceed 17 h for any of the specimens. While all four cortical areas were available for the majority of specimens, some individuals were not represented for every area. All available materials were analyzed for neuron densities and axon length densities. A minimum of four individuals per species (except three for squirrel monkeys) was used to quantify varicosity densities.

Prior to sectioning, specimens were cryoprotected in a series of sucrose solutions (10, 20, and 30%) until saturated. Brains were frozen in dry ice and cut in the coronal plane into 40 μm-thick sections using a Leica SM2000R freezing sliding microtome. Sections were placed into individual microcentrifuge tubes containing a freezer storage solution (30% of each distilled water, ethylene glycol, and glycerol and 10% 0.244 M phosphate-buffered saline) and numbered sequentially. Sections were stored at −20°C until further processing. Every tenth section was stained for Nissl substance with a solution of 0.5% cresyl violet. Nissl-stained sections were used to identify the cortical areas of interest based on cytoarchitecture.

The identification of each cortical area was performed using descriptions from previous parcellations in a wide range of primate species, including galagos, New World monkeys, Old World monkeys, apes, and humans (Bailey et al., 1950; Rosabal, 1967; Preuss and Goldman-Rakic, 1991b; Watanabe-Sawaguchi et al., 1991; Vogt et al., 1995; Petrides and Pandya, 1999; Paxinos et al., 2000; Sherwood et al., 2003, 2004; Petrides, 2005; Schenker et al., 2008; Spocter et al., 2010). Because the cytoarchitecture of the areas of interest have not been described for all species analyzed in the present study, reports from a diverse array of primates were used to identify the cortical areas using Nissl-stained sections.

Broca's area (area 44) and its homolog have been described in humans (Amunts et al., 1999; Petrides, 2005), chimpanzees (Sherwood et al., 2003), and other great apes (Schenker, 2007). There has been considerable controversy over whether monkeys possess a homolog of human area 44. Petrides et al. (2005) addressed this question using both electrophysiological recordings and quantitative cytoarchitecture in macaque monkeys, demonstrating the presence of a cortical area that is topologically and cytoarchitecturally comparable to that of humans and which is involved in orofacial muscle movement. Coudé et al. (2011) also recently reported that voluntary vocalizations, but not emotional vocalizations, evoked the firing of neurons within the macaque homolog of Broca's area. Area 44 is characterized in each species as being dysgranular with a poorly developed layer IV and large pyramidal cells in the lower portion of layers III and V.

Wernicke's area and its homolog (area 22 or Tpt) is well described for humans (Fullerton and Pandya, 2007), chimpanzees (Spocter et al., 2010), macaques (Preuss and Goldman-Rakic, 1991a; Lewis and Van Essen, 2000; Fullerton and Pandya, 2007; Gannon et al., 2008), and galagos (Preuss and Goldman-Rakic, 1991a). Area Tpt is located in the caudal portion of the superior temporal gyrus and is readily distinguished by its well-developed layer II, a columnar appearance of pyramidal neurons in layers III and V and a thick layer IV (Galaburda and Pandya, 1982; Preuss and Goldman-Rakic, 1991a; Fullerton and Pandya, 2007; Gannon et al., 2008).

Both areas 10 and 24 are present in all haplorhine primate species (Brodmann, 1909; Allman et al., 2001; Semendeferi et al., 2001; Burman et al., 2006; Passingham and Wise, 2012). The anterior cingulate cortex (area 24) is involved in visceral, emotional, and cognitive processes including control of heart rate, blood pressure, vocalizations, and facial expressions (Smith, 1945; Jürgens and Ploog, 1970; Jürgens, 1998). This area also regulates emotional self-control (Damasio et al., 2000), and is activated during cognitively demanding tasks as well as intense drive states in humans (Paus et al., 1993, 1998; Bartels and Zeki, 2000; Bush et al., 2000). The cytoarchitecture of area 24 has been described in humans (Vogt et al., 1995; Öngür and Price, 2000; Öngür et al., 2003), macaques (Hof and Nimchinsky, 1992; Carmichael and Price, 1994; Dombrowski et al., 2001; Vogt et al., 2005), and in a study that compared the cytoarchitecture of humans to that of macaques (Petrides and Pandya, 1994). Area 24 is located in the anterior cingulate gyrus (Carmichael and Price, 1994; Vogt et al., 1995, 2005; Öngür et al., 2003) and the region sampled for this study was anterior to the genu of the corpus callosum, which is agranular.

Area 10 is involved in verbal and non-verbal episodic memory, reward-based decision making, and planning future actions (Buckner, 1996; Okuda et al., 1998; Daffner et al., 2000; Lepage et al., 2000; Burgess et al., 2007). Lesions restricted to area 10 in humans impair the ability to extract meaning from events, and limit creative thinking, artistic expression, and disrupt the ability to plan future actions. Semendeferi et al. (2001) assessed the homology of area 10 among humans, apes, and macaque monkeys. The authors found that area 10 of the frontal pole in human, bonobo, chimpanzee, and orangutan was homologous with the orbital sector of the frontal pole in gibbons. The orbitolateral component of the frontal pole in macaques most closely resembled area 10 of the other species (Semendeferi et al., 2001). Area 10 also has been described in the hamadryas baboon (Watanabe-Sawaguchi et al., 1991), squirrel monkey (Rosabal, 1967), and marmoset (Burman et al., 2006). Area 10 is a granular cortex with a thin layer II, wide layer III, and a sharp border between layer VI and white matter (Semendeferi et al., 2001).

Every tenth section from each individual and cortical area was processed for NPY immunohistochemistry using the avidin-biotin-peroxidase method as described previously (Raghanti et al., 2013). Briefly, sections were pretreated for antigen retrieval by incubation in 0.05% citraconic acid (pH 7.4) at 85–90°C for 30 min. Endogenous peroxidase was quenched and sections were preblocked, pretreated, and incubated in a polyclonal primary antibody (Abcam, Cambridge, MA, catalog # ab30914) at a dilution of 1:5000 for 24 h at room temperature followed by 24 h at 4°C (Massoner et al., 2013; Raghanti et al., 2013). This antibody is derived from within residues 1 to the C-terminus of pig NPY. NPY has a high degree of sequence homology among vertebrates, with 22 of its 36 positions identical in all sequenced species (Larhammar et al., 1987; Larhammar, 1996), and 100% sequence homology in human, rabbit, rat, guinea pig, and alligator (O'Hare et al., 1988; Larhammar, 1996). After incubation in primary antibody, sections were incubated in a biotinylated secondary antibody (1:200) followed by the avidin-peroxidase complex (PK-6100, Vector Laboratories, Burlingame, CA). A 3,3′-diaminobenzidine-peroxidase (DAB) substrate with nickel enhancement was used as the chromogen (SK-4100, Vector Laboratories).

Quantitative analyses were performed using computer-assisted stereology with a system including an Olympus BX-51 photomicroscope equipped with a digital camera and StereoInvestigator software version 10 (MBF Biosciences, Williston, VT). Sampling parameters for each species and variable were determined using subsampling techniques to obtain a coefficient of error below 0.08 (Slomianka and West, 2005). Sampling areas (i.e., layers III and V-VI of each area) were outlined at low magnification (4×, N.A. 0.13). Total axon length was measured utilizing the SpaceBalls probe under Koehler illumination at 60x using the following equation: L = 2 * (v/a) * (Σ is) * 1/asf * 1/ssf * 1/tsf, where v/a is the ratio of sampling frame volume to probe surface area, Σis is the sum of the number of intersections between fibers and sampling hemispheres, asf (area sampling fraction; the fraction of the total area sampled) is the area of the counting frame divided by the total area of the reference space, section sampling fraction (ssf) is the number of sections analyzed divided by the total number of sections through the reference space, and tissue sampling fraction (tsf) is the sampling box height divided by mean mounted section thickness (Calhoun et al., 2004). Mean mounted section thickness was measured at every fifth sampling site. To measure axon length density (ALv), total axon length was divided by the planimetric measurement of the reference volume sampled.

Axonal varicosities, or synaptic boutons, are enlarged areas of an axon that contain release sites for neurotransmitters (Sutoo et al., 2001; Shepherd and Raastad, 2003), and were quantified as a putative measure of neuron-specific innervation to differentiate between axons that are directly innervating cells of a specific area vs. those that are en passant. Varicosities were defined as having a transverse diameter greater than 0.5 μm (Mechawar et al., 2000). Varicosity density (Vv) was quantified using the optical disector probe with a 100× (N.A. 1.4) objective lens. Vv was calculated as the sum of varicosities counted divided by the product of the disector and the volume of the disector (e.g., Sherwood et al., 2007). In addition to varicosity density, mean spacing of varicosities was calculated for each cortical layer and area for each species using the calculated estimates of NPY-ir Vv and ALv. Varicosity spacing has received attention as a putative measure of synaptic density (Shepherd et al., 2002; Shepherd and Raastad, 2003). Here, we aimed to evaluate the variability of varicosity spacing (i.e., mean spacing of varicosities in μm) to determine if this system is globally conserved across cortical regions within a species and among species. Nissl-stained sections were used to obtain neuron density (Nv) with a 60× (N.A. 1.35) objective lens. Nv was calculated in the same manner as Vv.

The variables used to compare cortical NPY innervation among species were ALv/Nv, Vv/Nv, and mean varicosity spacing. The ratio of NPY-ir axon and varicosity density to neuron density allowed for comparison among species with divergent brain sizes. Factorial analyses of variance (ANOVA) with repeated-measures design were used to analyze differences among species for NPY-ir ALv/Nv, Vv/Nv, and mean varicosity spacing. Cortical area and layer were within-subjects factors and species was the between-subjects factor. Repeated-measures ANOVAs also were used to evaluate the uniformity of varicosity spacing within each species with both layer and area as within-subjects factors. Fisher's LSD post hoc analyses were employed to evaluate significant findings. Nonparametric Spearman's correlation coefficient was used to assess whether PMI affected the intensity of immunohistochemical staining in humans. Exact PMIs were not known for other species. Additionally, Spearman's correlation was used to test the effect of age on cortical NPY innervation within each species and the relationship between ALv/Nv and Vv/Nv. Independent t-tests were used to examine differences between sexes. Standard linear regression was used to determine if differences in cortical ALv or Vv were associated with variation in brain mass or encephalization quotient. Brain masses were obtained for each specimen before processing. Because body mass was not available for all individuals, encephalization quotients were taken from Jerison (1973). Neither body mass nor encephalization quotients were available for moor macaques and were omitted from these analyses. Finally, an intraclass correlation coefficient was used to assess inter-observer reliability for quantitative stereological data. Periodically, inter-observer reliability was evaluated utilizing an audit feature within the StereoInvestigator software. The level of significance (α) was set at 0.05 for all statistical tests.

We examined the effect of PMI on our measures of cortical NPY in humans, where the most precise data on PMI were available. In humans, PMI was not correlated with NPY-ir ALv/Nv in layers III or V-VI in area 10 (Spearman's rho = −0.23, p = 0.66; −0.12, p = 0.83), area 24 (Spearman's rho = −0.21, p = 0.74; 0.15, p = 0.81), or area 22 (Spearman's rho = 0.56, p = 0.61; 0.31, p = 0.61). PMI was not correlated with NPY ALv/Nv in layer III of area 44 (Spearman's rho = 0.56, p = 0.32), but a significant correlation existed within layers V-VI (Spearman's rho = −0.81, p = 0.03). Given that 87.5% of possible correlations were not significant and PMI did not exceed 17 h for any individual, PMI is not considered a confounding variable for this study.

Independent samples t-tests were used to determine possible differences in NPY-ir ALv/Nv or Vv/Nv between males and females in each species except squirrel monkeys, as only females were available from this species. No sex differences were detected in capuchins, pigtailed macaques, baboons, chimpanzees, or gorillas. In humans, females exhibited a denser ALv/Nv relative to males only for layers V-VI of area 22 [t₍₃₎ = −3.51, p = 0.04]. There was also a sex difference for layer III of area 44 in moor macaques [t₍₄₎ = −3.32, p = 0.03], with females having a denser ALv/Nv. Given these results, the overall effect of sex on cortical NPY innervation was not substantial.

To determine if there was a significant effect of age on NPY-ir ALv/Nv and Vv/Nv, each species was analyzed separately, for a total of 64 correlation analyses per measure. Age was not correlated with NPY ALv/Nv in any layer or area for non-human species. In humans, only layer III of area 22 showed an effect of age (Spearman's rho = −0.95, p = 0.01). For Vv/Nv, two significant correlations were found for pigtailed macaques (layer III of area 44, Spearman's rho = −0.90, p < 0.01; layers V-VI of area 22, p = −1.00, p < 0.01) and one for capuchins (layer III of area 10, Spearman's rho = 1.00, p < 0.01). Age within the present was not a significant variable as only 0.02% (ALv/Nv) and 0.05% (Vv/Nv) of the possible correlations were significant.

NPY-ir ALv and Vv were each regressed on brain mass and encephalization quotient. There was a significant relationship between ALv and brain mass [F_{(1, 343)} = 28.48, p < 0.01, R² = 0.08], but not Vv and brain mass [F_{(1, 289)} = 1.66, p = 0.20, R² = 0.006). Neither measure was significant when regressed on encephalization quotient [ALv: F_{(1, 296)} = 2.70, p = 0.10, R² = 0.009; Vv: F_{(1, 250)} = 0.05, p = 0.83, R² = 0.00].

Across the entire primate sample, NPY-ir ALv/Nv was significantly correlated with Vv/Nv in layers III and V-VI in all cortical areas (Spearman's rho, all p's < 0.05, Figure 4). In some individuals, not every cortical area was represented due to tissue damage during brain extraction or unavailability of tissue. Therefore, analyses were restricted to individuals in which all four cortical areas were represented (see Table 1). Initial repeated-measures ANOVAs revealed no differences among species within the phylogenetic groups of New World monkeys, Old World monkeys, and hominids (great apes and humans) for ALv/Nv (Figure 5), Vv/Nv (Figure 6), and varicosity spacing (Figure 7). Thus, species were collapsed into these categories for further analysis.

A repeated-measures ANOVA was used to analyze ALv/Nv among New World monkeys, Old World monkeys, and hominids (Figure 8). There were significant main effects of phylogenetic group [F_{(2, 34)} = 10.34, p < 0.01], cortical area [F_{(3, 102)} = 8.24, p < 0.01], and layer[F_{(1, 34)} = 14.75, p < 0.01]. Additionally, the interaction between layer and phylogenetic group was significant [F_{(2, 34)} = 6.32, p < 0.01]. The interaction terms of area by phylogenetic group, area by layer, and the three-way interaction of area by layer by phylogenetic group were not significant (all p's > 0.05). Fisher's LSD post-hoc analyses revealed that hominids had significantly higher ALv/Nv relative to both New World monkeys and Old World monkeys (all p's < 0.01). No differences were detected between the monkey groups (p = 0.50).

The repeated measures ANOVA for Vv/Nv (Figure 9) revealed significant main effects of phylogenetic group [F_{(2, 28)} = 24.73, p < 0.01] and area [F_{(3, 84)} = 4.94, p < 0.01]. The main effect of layer [F_{(1, 28)} = 0.87, p = 0.36] and interaction terms were not significant (all p's > 0.05). Fisher's LSD post-hoc analyses demonstrated that hominids had significantly higher Vv/Nv relative to both New World monkeys and Old World monkeys (all p's < 0.05). No differences were detected between the monkey groups (p = 0.56).

For varicosity spacing (Figure 10), there were significant main effects of phylogenetic group [F_{(2, 28)} = 17.35, p < 0.01], area [F_{(3, 84)} = 18.19, p < 0.01] and layer [F_{(1, 28)} = 76.19, p < 0.01]. The interaction of area by phylogenetic group [F_{(6, 27)} = 3.07, p < 0.05], layer by phylogenetic group [F_{(2, 27)} = 10.53, p < 0.01], and layer by area [F_{(3, 27)} = 10.47, p < 0.01] were significant. The three-way interaction of phylogenetic group by area by layer was not statistically significant [F_{(6, 84)} = 1.97, p = 0.08]. Post-hoc analyses showed that hominids had smaller inter-varicosity spacing relative to both New and Old World monkeys in area 22 (p = 0.03 and 0.001, respectively). Hominids also had a lower varicosity spacing relative to Old World monkeys in area 44 (p = 0.01). Furthermore, hominids displayed a lower mean varicosity spacing in layer III relative to the Old World monkeys (p = 0.02). No other significant differences were detected.

Varicosity spacing was analyzed within each species to determine variability among cortical areas and layers using repeated-measures ANOVAs. A summary of main effects and interaction are listed in Table 2.

For squirrel monkey, post-hoc tests showed that area 22 had higher varicosity spacing relative to the other cortical areas, with area 10 having the lowest varicosity spacing (all p's < 0.05). No significant differences between layers were detected (all p's > 0.05). Capuchin varicosity spacing was higher in area 22 relative to the other areas and was lower in layers V-VI relative to layer III within areas 44 and 22 (all p's > 0.05).

Post-hoc analyses for moor macaques showed that areas 44 and 22 had higher varicosity spacing relative to areas 10 and 24, and layers V/VI displayed lower varicosity spacing than layer III in areas 24 and 44 (all p's < 0.05). In pigtailed macaques, there was increased spacing of varicosities in area 22 relative to area 10 (p = 0.04). Distance between varicosities was lower for layer V-VI relative to layer III (p = 0.02). For baboon, area 22 had higher varicosity spacing relative to areas 10 and 24 (p = 0.02 for each). Post-hoc analyses did not detect a significant difference between layers III and V/VI.

Post-hoc analyses for gorillas did not detect a significant difference between layers (p = 0.13). No significant differences were detected in either the chimpanzee or human.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.