Abstract

In primates, different cortical areas send axons of different diameters into comparable tracts, notably the corpus callosum (Tomasi S, Caminiti R, Innocenti GM. 2012. Areal differences in diameter and length of corticofugal projections. Cereb Cortex. 22:1463–1472). We now explored if an area also sends axons of different diameters to different targets. We find that the parietal area PEc sends thicker axons to area 4 and 6, and thinner ones to the cingulate region (area 24). Areas 4 and 9, each sends axons of different diameters to the nucleus caudatus, to different levels of the internal capsule, and to the thalamus. The internal capsule receives the thickest axon, followed by thalamus and nucleus caudatus. The 2 areas (4 and 9) differ in the diameter and length of axons to corresponding targets. We calculated how diameter determines conduction velocity of the axons and together with pathway length determines transmission delays between different brain sites. We propose that projections from and within the cerebral cortex consist of a complex system of lines of communication with different geometrical and time computing properties.

Introduction

Axons originating in different cortical areas of the macaque monkey differ in their diameter, with the thickest axons originating in primary motor and somatosensory areas and the thinnest ones in prefrontal and temporal areas (Tomasi et al. 2012). Areal differences in axon diameters were found for both callosal and for cortico-thalamic axons, but not for intra-area axons. The diameter of outgoing axons, therefore, appears to be a feature characterizing cortical areas, in addition to the classical size and distribution of cell bodies, degree of myelination, topography of connections, neuronal response properties, etc. Axon diameter determines conduction velocity and together with axon length generates temporal delays, whereby different brain sites communicate. It was proposed that conduction delays may constrain processing speed, and that there may be a hierarchy of processing speeds among cortical areas, with premotor, motor, and primary somatosensory areas on the lead (Tomasi et al. 2012). Areas with, on average, thicker axons also had an increased spectrum of axon diameters and of conduction delays. These differences appear to be exaggerated in evolution in parallel with the increased volume of the brain and might enrich cortical dynamics (Caminiti et al. 2009).

It was unclear if diameter relates not only to the area of origin of the axons or but also to their sites of termination. This question was tackled in the present study by comparing the diameter of axons originating in the posterior parietal cortex and terminating in motor, premotor, and cingulate cortices. We also compared axons originating in area 4, and in prefrontal cortex area 9, and terminating in the caudate, and thalamus or continuing into the internal capsule.

Materials and Methods

Two monkey brains injected with biotinylated dextran amine (BDA) in the dorsal part of area 4, presumably the proximal arm/trunk representation, were at our disposal (Tomasi et al. 2012); one of them was also injected in prefrontal area 9. This allowed to reduce the number of animals as much as possible since in studies with primates the repetition of experiments is restricted by ethical considerations and administrative rules; the projection from the parietal cortex was studied in a third animal. An additional monkey injected in area PEc was prepared for electron microscopy (EM), and intracortical axons were measured in the parietal cortex. The surgical methods were reported in detail in Tomasi et al. (2012). Briefly, deeply anesthetized animals received three to five 0.3–0.5 µL injections of BDA (Invitrogen, Carlsbad, CA, USA) MW 10 000 (10% in 0.01 M phosphate buffer, PB) in area 4, 9, and PEc/PE border (area 5). The monkeys were males ranging between 6 and 8 years of age and weighing between 6 and 10 kg.

For EM, one animal was perfused 5 min with heparinized buffered saline followed by 2 L of 3% paraformaldehyde (PAF) and 1% glutaraldehyde in 0.9% NaCl in 0.15 M PB, following clamping of the descending aorta. The skull and the dura were opened, and the head was stored in 3% PAF in 0.9% NaCl in 0.15 M overnight. The following day, the corpus callosum was cut on the midline. Parasagittal blocks were cut from each hemisphere, containing the injection site and the labeled fibers. Eighty-micrometer-thick sections were obtained at the vibratome and reacted with BDA according to Wang et al. (2010). Selected sections were osmicated and flat embedded in Araldite, to be sectioned for EM. Semithin sections were cut on the ultramicrotome with a glass knife, to outline the area of interest; ultrathin sections (from silver to gold) were cut with a diamond knife, counterstained with uranyl acetate and lead citrate, and observed with a JEM-1010 transmission electron microscope (JEOL, Tokyo, Japan) equipped with a Mega-View-III digital camera and a Soft-Imaging System (SIS, Münster, Germany) for the computerized acquisition of the images, at the magnification of 10 or 20K, for measuring myelinated and unmyelinated fibers, respectively. The diameter of BDA-labeled axons identified in the gray matter near the injection site was measured in 4 different specimens, using the software ImageJ (NIH, USA) by drawing a line along the shortest axis of the axon profile. BDA-labeled axons were identified under the EM by the greater electron density of the cytoplasm, except for mitochondria, particularly in the vicinity of vesicular and other membranes.

For light microscopy (LM) sections from brains cryoprotected in 30% (wt/vol) sucrose in PB saline were cut frozen at 34 µm and reacted with BDA. Alternating sections were stained with Cresyl Violet and the Gallyas method for myelin as in Tomasi et al. (2012).

Axons originating in area PEc/PE were sampled and measured as in Tomasi et al. (2012). The axons profiles at the upper or lower surface of the sections were approximated to circles whose size was incremented in 0.09-µm steps. Axons originating in areas 4 and 9 were sampled and measured with the following strategy using the Neurolucida software (MBF Biosciences, Williston, VT, USA). Since each monkey had received injections at different locations, a preliminary sketchy reconstruction of the pathway was performed at ×345 magnification (2 cm = 58 μm) to separate the different projections. Areas of interest were circumscribed and within these areas 80–119 axon segments traversing the section were drawn at ×3448 magnification (2 cm = 5.8 μm) over their entire length, that is, between 5 and 70 µm faithfully following the changes in diameter that the axon undergoes in its trajectory. The Neuroexplorer software (MBF Biosciences) calculated the diameter of the axonal segment at the surface of the section as well as the average diameter of the axonal segment. The 2 types of measurements are similar but not identical. The average diameter can be affected by apparent thinning of the axon due to incomplete visualization in the depth of the section. The diameter of the axon at the surface of the section more closely approximates the measurements performed with the circle and, as the latter, is affected by whether the section fell by chance on a thin or on a varicose portion of the axon. Eventually, the latter measurement at the surface of the section was retained.

The significance of differences was tested with the Mann–Whitney 2-tailed U-test. Lengths of tracts were measured as curvilinear segments with one of the Neuroexplorer software options (MBF Biosciences).

Correction for Shrinkage

Shrinkage in this material consisted of 2 components. The first component is due to the perfusion–fixation and was estimated by comparing the distance between the sites of injection at the time of the stereotaxic measurements and on the sections in 4 animals. This component varied between 1.1 and 1.4 (mean 1.2). The second component is the differential shrinkage between the BDA and the cresyl violet or myelin staining. This component was negligible. Therefore, we applied a 1.3 correction for shrinkage to the results on the distribution of axon diameters and conduction velocities of axons originating in area 4 in order to test if the shrinkage of the histological material might explain the differences between our anatomical results and some electrophysiological data in the literature (see Discussion).

Results

Sites of Injection and Axonal Trajectories

The sites of injections are shown in Figure 1. The core of the injections was between 500 µm and 2 mm in diameter, restricted to the gray matter. The location was based on cytoarchitectonic criteria and on comparisons with Paxinos et al. (2000) atlas. The axonal trajectories correspond to those reported by Schmahamann and Pandya (2009) for similar injection locations (see Discussion). They can also be appreciated in Figures 3 and 5 and, therefore, they will not be described in detail.

Figure 1.

Location and extent of BDA injection sites. Top: Area PEc injection in experiment CCT1; middle: Area 9 injection in CCT2; and bottom: Area 4 injection in CCT5. Each injection site is shown in outlines of the section and in photomicrographs of BDA-reacted material. Vertical segments mark the boundaries of area PEc close to the border with PE (top panel), 9 L and 46 (and the 9/46 transition between the two, not labeled; middle panel), and area 4 (bottom panel); the outlines of the injection site are also projected on the CV-stained section where the large pyramids of layer 5 can be appreciated and the injection track is marked by an arrow. Calibration bars are 1 mm in the section outlines and 500 µm in the photomicrographs.

Figure 1.

Location and extent of BDA injection sites. Top: Area PEc injection in experiment CCT1; middle: Area 9 injection in CCT2; and bottom: Area 4 injection in CCT5. Each injection site is shown in outlines of the section and in photomicrographs of BDA-reacted material. Vertical segments mark the boundaries of area PEc close to the border with PE (top panel), 9 L and 46 (and the 9/46 transition between the two, not labeled; middle panel), and area 4 (bottom panel); the outlines of the injection site are also projected on the CV-stained section where the large pyramids of layer 5 can be appreciated and the injection track is marked by an arrow. Calibration bars are 1 mm in the section outlines and 500 µm in the photomicrographs.

Electron Microscopic Analysis of Axon Diameters

The limited resolution of the LM could introduce biases in our assessment of axon diameters. For certain tracts, notably the corpus callosum, our estimates could be compared with EM data in the literature, and this possible source of error could be avoided (discussed in Tomasi et al. 2012). Unfortunately, a comparison of our measurements with EM data is not available for all tracts, and it is not feasible to analyze all the corticofugal tracts of the monkey with combined tract-tracing and EM techniques. In this study, one monkey was injected with BDA in the PEc subdivision of area 5, and the diameter of the labeled local intracortical axons was measured in EM photomicrographs. The choice was dictated by 2 considerations. The first is that the local intracortical axons are likely to be the thinnest since many of them span much shorter distances than the other corticofugal axons. The second is that the diameter of local intracortical axons was found to be surprisingly similar in the 5 areas (9, 9/46, 6, 6-F4, and 4) where they were studied in 3 different animals (Tomasi et al. 2012). If a comparable spectrum of axon diameters were to be found in the new experiment, it could confirm that 1) the LM analysis is not seriously biased by missing small axons, and that 2) the intracortical axons are indeed of similar diameter across different cortical areas. Examples of BDA-labeled myelinated and unmyelinated axons are shown in Figure 2. Sixty-six myelinated axons were measured and their mean diameter was 0.76 μm [±0.29 standard deviation (SD); median 0.71], of which the axoplasm was 0.53 μm (±0.24 SD; median 0.49). The axoplasm/axon ratio was, as expected, almost exactly 0.7. Seventy-five unmyelinated axons were measured and their mean diameter was 0.44 μm (±0.16 SD; median 0.4; marginally smaller, P < 0.05, than the axoplasm of the myelinated axons). Since BDA appears not to diffuse into the myelin sheath, the comparison between EM and LM data was restricted to the axoplasm. The EM values we found for the axoplasm of BDA-labeled axons fall within the range of intracortical axon diameters previously measured with LM (0.48–0.66 μm; Tomasi et al. 2012), and it is nearly identical to the axon diameters found in the same study (0.48–0.51 μm) in areas 9, 9/46, 6, and 6-F4, while those from area 4 were slightly thicker (0.54 and 0.66 μm).

Figure 2.

Photomicrographs of 3 myelinated (A, B, D) and 1 unmyelinated (C) BDA-labeled axons. Calibration bars are 0.5 μm.

Figure 2.

Photomicrographs of 3 myelinated (A, B, D) and 1 unmyelinated (C) BDA-labeled axons. Calibration bars are 0.5 μm.

Cortico-cortical Projections from Area PEc

The cortico-cortical axons originating in area PEc (Fig. 3) course in the superior longitudinal fasciculus [SLF; the subdivision SLF I of Schmahamann and Pandya (2009)]. They terminated at 3 separate locations in areas 4, 24, and 6. In area 4, the axons ended in its caudal part, in 3 columns. In area 6, the axons ended on the convexity of the superior frontal gyrus in the dorso-caudal division of premotor cortex, cytoarchitectonically corresponding to area F2 of Belmalih et al. (2007). At the same rostrocaudal levels, axons also terminated at the fundus and on the dorsal bank of the cingulate sulcus. This site of termination is characterized by a superficial band of tangentially coursing, myelinated axons that acquire a deeper location and lose their identity in the more dorsal cortex. The termination corresponds to a subdivision of area 24 in the proximity to the fundus of the sulcus, but it extends more dorsally and might reach area 6/32 according to Paxinos et al. (2000). At roughly corresponding locations, Picard and Strick (1996) described cingulate motor areas, notably CMAd.

Figure 3.

Location of axon measurement sites (white polygons) in the SLF (bottom panel), and at the entrance of axons in area 4 (top left), areas 24 and 6 (top right) along the sketched axonal pathways in superposed outlines of sequential, nonconsecutive, sections. Calibration bars are 2000 μm.

Figure 3.

Location of axon measurement sites (white polygons) in the SLF (bottom panel), and at the entrance of axons in area 4 (top left), areas 24 and 6 (top right) along the sketched axonal pathways in superposed outlines of sequential, nonconsecutive, sections. Calibration bars are 2000 μm.

Axons were measured in the SLF I before they distribute to their sites of termination and then at their entrance into the gray matter (Table 1 and Figs 3 and 4). Axons in the SLF were 1.03 µm on average (median 0.8). Those to areas 4 and 6 had similar diameters (means 0.95 and 0.96 µm; medians 0.8 and 0.87, respectively), while those to area 24 were thinner (mean 0.77 µm, median 0.73). Mann-Whitney U-test reported highly significant differences (P < 0.001) between the axons to area 6 and those to area 24, marginally not significant differences (P ≤ 0.05) for the other comparisons.

Table 1

Axons from area PEc projecting to other cortical areas

 SLF to area 4 to area 24 to area 6 
Mean 1.03 0.95 0.77 0.96 
Median 0.80 0.80 0.73 0.87 
Mode 0.51 0.44 0.66 0.58 
SD 0.75 0.57 0.21 0.36 
Sample variance 0.56 0.33 0.05 0.13 
Kurtosis 4.87 5.86 2.10 2.03 
Skewness 2.09 2.04 1.40 1.21 
Range 4.16 3.50 1.16 1.97 
Minimum 0.29 0.29 0.44 0.44 
Maximum 4.45 3.79 1.60 2.41 
Count 273 207 203 164 
 SLF to area 4 to area 24 to area 6 
Mean 1.03 0.95 0.77 0.96 
Median 0.80 0.80 0.73 0.87 
Mode 0.51 0.44 0.66 0.58 
SD 0.75 0.57 0.21 0.36 
Sample variance 0.56 0.33 0.05 0.13 
Kurtosis 4.87 5.86 2.10 2.03 
Skewness 2.09 2.04 1.40 1.21 
Range 4.16 3.50 1.16 1.97 
Minimum 0.29 0.29 0.44 0.44 
Maximum 4.45 3.79 1.60 2.41 
Count 273 207 203 164 
Figure 4.

Bar diagram of axon diameters at different locations along the parieto-frontal projection. S35 denotes the section where measurements were performed.

Figure 4.

Bar diagram of axon diameters at different locations along the parieto-frontal projection. S35 denotes the section where measurements were performed.

The axons labeled by the BDA injection in area PEc were a subpopulation of the myelin-stained SLF axons. Profiles of cross-sectioned myelinated axons were sampled in SLF (10 921; preliminary work for another study). Their average near-neighbor distance was only 6.3 µm but 33 µm for the BDA-labeled axons, suggesting that only 1 in 5 axons was BDA labeled. Myelinated axons were also significantly thicker (1.3 µm on average) than the BDA-labeled axons. This is not surprising since at the locations measured SLF I contain also axons originating in area 4 and elsewhere (see Schmahamann and Pandya 2009).

Descending Projections from Areas 4 and 9

Axons were sampled at 4 locations: 1) at their entrance into the dorsal aspect of the caudate, 2) in the proximal part of the internal capsule, dorsal to the entrance into the thalamus, 3) at entrance into the thalamus, in proximity to their principal thalamic nuclei, that is, VL and MD, where a few retrogradely labeled neurons were also seen, and 4) in the distal part of the internal capsule. This location corresponded to the entrance into the cerebral peduncle for the axons from area 4; the axons from area 9 were sampled more dorsally just ventral to the thalamus while coursing toward the zona incerta, since the axons reaching the cerebral peduncle were few and difficult to separate from those labeled by the more caudal injection in area 4. In addition, in 1 experiment, the main tract streaming from the injection site was sampled dorsal to the separation of axons directed to the corpus callosum, internal capsule, and caudate (Fig. 5).

Figure 5.

Location of axon measurement sites (white polygons) at different locations in the projection from area 4, along the sketched axonal pathways in superposed outlines of sequential, nonconsecutive, sections (experiment CCT2). Insets show axonal labeling and selection of axonal segments at the entrance into caudate and traced axons in proximal internal capsule in experiment CCT5.

Figure 5.

Location of axon measurement sites (white polygons) at different locations in the projection from area 4, along the sketched axonal pathways in superposed outlines of sequential, nonconsecutive, sections (experiment CCT2). Insets show axonal labeling and selection of axonal segments at the entrance into caudate and traced axons in proximal internal capsule in experiment CCT5.

To test the consistency of measurements across studies, we resampled axons to the proximal internal capsule and the results (Table 2 and Fig. 6) well correspond to those of earlier measurements (Tomasi et al. 2012). The mean value of axon diameters from area 4 to the proximal internal capsule is 0.74 and 0.91 µm (medians 0.6 and 0.7 µm) in the 2 experiments, compared with the 0.68 and 1.01 µm (medians 0.59 and 0.89 µm) of the previous estimates in the same animals (Tomasi et al. 2012). For area 9, we found an average diameter of 0.59 µm (median 0.50) versus 0.51 (mean; 0.46 median) of the previous estimates. The small differences might suggest the existence of local heterogeneities in the axonal composition of the various projections, which further studies might document.

Table 2

Size of axons originating from different injection sites and projecting to nucleus caudatus, internal capsule and to the thalamus

Motor area 4 (μm) 
 CCT2  To caudatus To ic prox To thalamus To ic dist 
  Mean  0.40 0.74 0.63 1.40 
  Median  0.40 0.60 0.50 1.20 
  Mode  0.40 0.40 0.40 1.60 
  SD  0.19 0.54 0.33 0.85 
  Sample variance  0.04 0.29 0.11 0.73 
  Kurtosis  15.55 6.75 2.43 0.70 
  Skewness  3.51 2.23 1.38 0.99 
  Range  1.20 3.20 1.80 3.50 
  Minimum  0.20 0.20 0.20 0.30 
  Maximum  1.40 3.40 2.00 3.80 
  Count  80 87 80 84 
 CCT5 Main tract To caudatus To ic prox To thalamus To ic dist 
  Mean 0.79 0.49 0.91 0.69 1.37 
  Median 0.50 0.40 0.70 0.70 1.00 
  Mode 0.30 0.40 0.40 0.40 1.20 
  SD 0.67 0.25 0.66 0.32 1.10 
  Sample variance 0.45 0.06 0.44 0.10 1.21 
  Kurtosis 4.18 5.67 15.90 0.87 5.86 
  Skewness 1.81 2.07 3.06 1.01 2.11 
  Range 3.70 1.50 5.00 1.50 6.60 
  Minimum 0.20 0.20 0.20 0.30 0.20 
  Maximum 3.90 1.70 5.20 1.80 6.80 
  Count 105 119 108 104 101 
Prefrontal area 9 μm 
 CCT2  To caudatus To ic prox To thalamus To ic dist 
  Mean  0.36 0.59 0.44 0.54 
  Median  0.40 0.50 0.40 0.45 
  Mode  0.40 0.40 0.40 0.40 
  SD  0.11 0.25 0.13 0.27 
  Sample variance  0.01 0.06 0.02 0.07 
  Kurtosis  2.22 1.21 5.16 3.50 
  Skewness  0.62 1.14 1.95 1.64 
  Range  0.60 1.20 0.80 1.40 
  Minimum  0.10 0.30 0.20 0.20 
  Maximum  0.70 1.50 1.00 1.60 
  Count  80 103 100 100 
Motor area 4 (μm) 
 CCT2  To caudatus To ic prox To thalamus To ic dist 
  Mean  0.40 0.74 0.63 1.40 
  Median  0.40 0.60 0.50 1.20 
  Mode  0.40 0.40 0.40 1.60 
  SD  0.19 0.54 0.33 0.85 
  Sample variance  0.04 0.29 0.11 0.73 
  Kurtosis  15.55 6.75 2.43 0.70 
  Skewness  3.51 2.23 1.38 0.99 
  Range  1.20 3.20 1.80 3.50 
  Minimum  0.20 0.20 0.20 0.30 
  Maximum  1.40 3.40 2.00 3.80 
  Count  80 87 80 84 
 CCT5 Main tract To caudatus To ic prox To thalamus To ic dist 
  Mean 0.79 0.49 0.91 0.69 1.37 
  Median 0.50 0.40 0.70 0.70 1.00 
  Mode 0.30 0.40 0.40 0.40 1.20 
  SD 0.67 0.25 0.66 0.32 1.10 
  Sample variance 0.45 0.06 0.44 0.10 1.21 
  Kurtosis 4.18 5.67 15.90 0.87 5.86 
  Skewness 1.81 2.07 3.06 1.01 2.11 
  Range 3.70 1.50 5.00 1.50 6.60 
  Minimum 0.20 0.20 0.20 0.30 0.20 
  Maximum 3.90 1.70 5.20 1.80 6.80 
  Count 105 119 108 104 101 
Prefrontal area 9 μm 
 CCT2  To caudatus To ic prox To thalamus To ic dist 
  Mean  0.36 0.59 0.44 0.54 
  Median  0.40 0.50 0.40 0.45 
  Mode  0.40 0.40 0.40 0.40 
  SD  0.11 0.25 0.13 0.27 
  Sample variance  0.01 0.06 0.02 0.07 
  Kurtosis  2.22 1.21 5.16 3.50 
  Skewness  0.62 1.14 1.95 1.64 
  Range  0.60 1.20 0.80 1.40 
  Minimum  0.10 0.30 0.20 0.20 
  Maximum  0.70 1.50 1.00 1.60 
  Count  80 103 100 100 

ic prox: proximal internal capsule; ic dist: distal internal capsule.

Figure 6.

Bar diagram of mean axon diameters (and SDs) at different locations in the projections from areas 4 and 9.

Figure 6.

Bar diagram of mean axon diameters (and SDs) at different locations in the projections from areas 4 and 9.

The axons originating from the motor cortex area 4 and directed to the thalamic nucleus VL are significantly thicker than those originating from the prefrontal cortex area 9 and directed to the nucleus MD (0.63 and 0.69 in 2 experiments vs. 0.44 µm; P < 0.001). The difference in favor of area 4 is even larger for axons directed to the distal internal capsule (1.4 and 1.37 vs. 0.54 µm; P < 0.001). Differences in the projection from areas 4 and 9 to the caudate (0.4 and 0.49 vs. 0.36 µm) and to the proximal internal capsule (0.74 and 0.91 vs. 0.59 µm) were just below the significance level (P < 0.05) in 1 experiment, but were significant in the other.

In conclusion, motor cortex area 4 sends the thickest axons toward the cerebral peduncle (distal internal capsule; 1.4 and 1.37 µm in the 2 experiments), thinnest axons to the caudate nucleus (0.4 and 0.49 µm), while the thalamic nucleus VL receives axons between those two (0.63 and 0.69 µm). These differences are statistically significant at the 0.05–0.001 level.

The distribution of different diameter axons to different targets could be implemented in 2 different ways, that is, collateralization of the axons or parallel axonal systems. Collateralization predicts that the axons within a tract would become thinner in the proximal-to-distal direction, while the parallel projection hypothesis allows for thicker axons distally (Fig. 7). The results of the present measurements are clearly in favor of the second possibility. Indeed, the projection from area 4 (Fig. 8) contains progressively thicker axons in the proximal-to-distal direction, while the contingent of thin axon progressively decreases, consistent with the hypothesis that these axons have terminated at their proximal targets, caudate and thalamus, in particular.

Figure 7.

The collateralization and parallel pathway hypotheses and their predictions.

Figure 7.

The collateralization and parallel pathway hypotheses and their predictions.

Figure 8.

Axon diameter distributions show a progressive proximo-distal shift toward thicker axons and disappearance of thinner axons compatible the parallel pathway hypothesis (Fig. 7).

Figure 8.

Axon diameter distributions show a progressive proximo-distal shift toward thicker axons and disappearance of thinner axons compatible the parallel pathway hypothesis (Fig. 7).

Consequences for Conduction Velocities and Delays

From the estimated diameters, we calculated the conduction velocities and from the length of the pathways the approximate delay to the target (Table 3 and Fig. 9). Velocity (V) was computed as previously (Tettoni et al. 1998; Caminiti et al. 2009; Tomasi et al. 2012) as V = (5.5/0.7) × diameter. As confirmed above, the factor 0.7 is the average axoplasm-to-myelin ratio, and it corrects for the fact that BDA visualizes only the axoplasm, without percolating into the myelin sheath.

Table 3

Conduction velocities path lengths and delays

 Diameter median (µm) Velocity (m/s) Path (µm) Delay (µs) 
Cortico-cortical axons 
 PEc to area 4 0.8 6.29 12 041 1916 
 PEc to area 24 0.73 5.74 20 201 3522 
 PEc to area 6 0.87 6.84 23 001 3365 
Descending axons 
 CCT2 area 9 to MD 0.4 3.14 25 934 8252 
 CCT2 area 4 to VL 0.5 3.93 19 615 4993 
 CCT5 area 4 to VL 0.7 5.50 20 306 3692 
 CCT2 area 9 to caudate 0.4 3.14 10 586 3368 
 CCT2 area 4 to caudate 0.4 3.14 13 475 4288 
 CCT5 area 4 to caudate 0.4 3.14 12 293 3911 
 CCT2 area 9 to distal ic 0.45 3.54   
 CCT2 area 4 to distal ic 1.2 9.43   
 CCT5 area 4 to distal ic 7.86   
 Diameter median (µm) Velocity (m/s) Path (µm) Delay (µs) 
Cortico-cortical axons 
 PEc to area 4 0.8 6.29 12 041 1916 
 PEc to area 24 0.73 5.74 20 201 3522 
 PEc to area 6 0.87 6.84 23 001 3365 
Descending axons 
 CCT2 area 9 to MD 0.4 3.14 25 934 8252 
 CCT2 area 4 to VL 0.5 3.93 19 615 4993 
 CCT5 area 4 to VL 0.7 5.50 20 306 3692 
 CCT2 area 9 to caudate 0.4 3.14 10 586 3368 
 CCT2 area 4 to caudate 0.4 3.14 13 475 4288 
 CCT5 area 4 to caudate 0.4 3.14 12 293 3911 
 CCT2 area 9 to distal ic 0.45 3.54   
 CCT2 area 4 to distal ic 1.2 9.43   
 CCT5 area 4 to distal ic 7.86   
Figure 9.

Schematic representation of diameters and estimated median conduction delays for connections studied, in this paper and in Tomasi et al. (2012). Thickness of arrow segments is proportional to median diameter and conduction velocity; Length of arrows is proportional to pathway length. Projections from area 4 and 9 are the means of two animals. Delays are calculated on conduction velocities and pathway lengths (Table 3 and Table 1 in Tomasi et al. 2012). CC, Corpus Callosum; areas are denoted according to Paxinos et al. (2000).

Figure 9.

Schematic representation of diameters and estimated median conduction delays for connections studied, in this paper and in Tomasi et al. (2012). Thickness of arrow segments is proportional to median diameter and conduction velocity; Length of arrows is proportional to pathway length. Projections from area 4 and 9 are the means of two animals. Delays are calculated on conduction velocities and pathway lengths (Table 3 and Table 1 in Tomasi et al. 2012). CC, Corpus Callosum; areas are denoted according to Paxinos et al. (2000).

The median axon diameter from area PEc to area 4 being 0.8 µm in diameter, it is expected to conduct at approximately 6.3 m/s. The distance (D) from the injection to the site where the diameters were sampled was found to be of approximately 12 mm. The delay to target (T = D/V) should be in the order of 1.9 ms. The projection to area 24, being 0.73 µm in diameter, is expected to conduct at 5.7 m/s over a distance of 20.2 mm generating delays in the order of 3.5 ms. The median axon to area 6 being 0.87 µm in diameter, its conduction velocity 6.8 m/s, and the path to target approximately 23 mm the delay to target can be expected to be 3.3 ms. Therefore, it appears that the information from area PEc should reach area 4 in advance of area 24 or 6.

The conduction velocity from area 4 to VL was 3.9 and 5.5 m/s in CCT2 and CCT5, respectively, generating delays of 5 and 3.7 ms over conduction distances of 19.6 and 20.3 mm. In contrast, the conduction velocity from area 9 to MD was 3.1 m/s, generating delays of 8.3 ms. These findings confirm that delays to their principal thalamic nuclei are greater for prefrontal than motor projections (Tomasi et al. 2012). The estimated conduction velocities to the caudate turned out to be the similar in all experiments (3.1 m/s). The estimated conduction delays ranged between 3.4 and 4.3 ms.

Finally, considerable differences in the conduction velocities of axons to the distal internal capsule were found between axons originating in area 4 (7.9 and 9.5 m/s in the 2 experiments) and those originating in area 9 (3.5 m/s). It should be stressed that the conduction delays are rough estimates since axons defasciculate near their targets; more precise estimates would require the serial reconstruction of individual axons to their sites of termination, which was impossible on the present material.

Discussion

Goals and Methodological Considerations

In this study, as in previous ones (Caminiti et al. 2009; Tomasi et al. 2012), we have chosen to undertake a light microscopic analysis of axonal projections originating in the cerebral cortex of the macaque monkey with focus on axon diameters and length of the projections. This work has 3 long-term goals. One is to explore cortical connectivity at the cellular resolution, when this is missing in the current database of brain structure. The second is to allow predictions on the conduction velocity and timing of interaction between the different structures connected to the cerebral cortex, which is relevant for the speed of information processing in the brain and for cortical dynamics (Roxin et al. 2005; Roberts and Robinson 2008; Caminiti et al. 2009; Panzeri et al. 2010). The third is to allow translation from the structure of the monkey brain to that of the human and/or of other primates in view of applications to human misconnectivity syndromes and to the evolution of the primate brain (Caminiti et al. 20092010). Comparisons of cellular monkey data with the human are timely since noninvasive imaging techniques are approaching the level of resolution required to analyze the axonal diameters in fiber tracts (e.g. Barazany et al. 2009; Dyrby et al. 2012).

The LM analysis is bound to miss axons with diameters below the level of optical resolution, that is 0.2–0.3 µm, but it appears to be the only approach possible, in view of the difficulty of performing systematic EM analyses of labeled axons in primates, particularly in humans, and in protected primate species. The error intrinsic to the LM analysis of the monkey material can be estimated by comparison with EM data in the literature when they exist (e.g. the corpus callosum; discussed in Tomasi et al. 2012). For reasons explained above (see Results), we have chosen to EM-examine the diameter of intracortical axons in the parietal cortex. Our EM measurements were in the order of the previous LM estimates of intracortical axon diameters, suggesting, as previously proposed (Tomasi et al. 2012), that “the structure of the cortical modules “columns” or “cortical output units” is invariant across areas because they perform similar computations,” while the computations performed by long projections, and hence their structure, vary across areas.

Comparison with Previous Studies

The cortico-cortical projections from area PEc resemble those documented in case 3 of Schmahamann and Pandya (2009). They described projections to areas 4 and 6 as well as a projection to the bottom of the cingulate sulcus in their sections 81–89. Unfortunately, our material did not extend further and therefore, we could not analyze the more rostral sites of termination. As mentioned in Results section, our site of termination near the bottom of the cingulate sulcus has distinct cytoarchitectonic features from the more dorsal part of area 6, which are most evident in layers 3 and 5, particularly in the pattern of myelination and it appears to correspond to a subdivision of area 24.

Our injection sites in area 4 resemble the more lateral injection of case 1 of Kultas-Ilinski et al. (2003) and case 27 of Schmahamann and Pandya (2009), but are somewhat smaller and more medial than the latter. The trajectory and location of the descending projection are similar to that documented in the second study. In a study by Kultas-Ilinski et al. (2003), the projection was studied within the thalamic nuclei themselves and in ours at the entrance into the thalamus, to measure axonal trunks rather than collaterals. The 2 studies agree that the projection to the VL overwhelmingly consists of “small” axons, which in their case, distributed small terminal boutons in the order of 0.21–0.6 µm2 occasionally larger. In addition, they found large axons projecting to other thalamic nuclei. These axons might have contributed to our previous sample of the projection to the VL (Tomasi et al. 2012), which consisted of slightly thicker axons. Differences in the diameter of axons directed to different thalamic nuclei might be a general feature of cortical projections, across systems and species (e.g. Rockland 1994; Rouiller and Welker 2000; Cappe et al. 2007). The 2 types of axons might originate from different layers, with thick axons mainly from layer 5 and thin ones from layer 6 as discussed by Kultas-Ilinski et al. (2003). The projection from area 4 included large axons directed to the cerebral peduncle.

Our conclusion that the projection from area 4 to the thalamus, to the caudate, and to the cerebral peduncle consists mainly or exclusively of different axonal systems seems consistent with the literature, which reports a higher degree of collateralization for descending projections in the brain stem (Lemon 2008). The projection to the thalamus originates mainly from layer 6, while that to the cerebral peduncle and to the striatum is mainly from layer 5, presumably, in primates, from different neuronal populations. This conclusion is tentative since our work cannot exclude the possibility that, as in the rat, a small population of small or medium size axons may send collaterals to multiple targets, including the striatum, thalamus and to the subthalamic nucleus (Kita and Kita 2012).

Our injection in area 9 is more medially located than that of case 31 of Schmahamann and Pandya (2009), but the axonal trajectories are similar, that is, to the dorsal aspect of the caudate, to the thalamic nucleus MD, and to the zona incerta where we sampled and measured axons.

The finding that axons from area 4 projecting to the cerebral peduncle conduct much faster than those to the caudate is compatible with the electrophysiological evidence that axons to the pyramidal tract are faster than those to another compartment of the basal ganglia, the putamen (Bauswein et al. 1989). However, the distribution of axons diameters in the cerebral peduncle (Fig. 8) did not provide the clear bimodal distribution of fast and slow axons, expected from the classical electrophysiological work on the pyramidal tract. Furthermore, only 23 of 185 axons measured in 2 experiments had diameters equal or greater than those compatible with velocities of 25 m/s, which is the cut off proposed by Humphrey and Corrie (1978) between the 2 populations. This is far from the 45% found by the same authors electrophysiologically with antidromic activation. Humphrey and Corrie (1978) attributed the discordance between their electrophysiological findings and the anatomical data available at the time to a strong electrophysiological sampling bias in favor of large cell bodies. The correction they proposed does indeed provide a better match between their electrophysiological data and our estimates of conduction velocities computed from diameters (Supplementary Fig. 1), although our sample still misses axons that could conduct at >35 m/s. Perhaps thicker axons originate from portions of area 4 corresponding to the distal extremity representation than in the proximal limb/trunk area we injected. Alternatively, the axons we measured might have shrunken, due to the fixation and histological processing. The correction for shrinkage we applied (Materials and Methods) returned slightly higher conduction velocities (Supplementary Fig. 2), but the results are still far from the electrophysiological data. The electrophysiological bias remains the most likely explanation of the discrepancy of data, although in a previous paper (Tomasi et al. 2012) we found that conduction velocities and latencies predicted from the anatomical data (uncorrected for shrinkage) of interhemispheric connections of peristriate areas perfectly matched with those found in an electrophysiological study of antidromic responses.

Consequences and Determinants of Axon Diameters

Confirming and extending the previous results (Caminiti et al. 2009; Tomasi et al. 2012), we found that axons of different caliber originate from different areas. The comparison of projections issued from area 4 and 9, chosen because they represent the extremes in the spectrum of outgoing-axon diameters, showed thinner axons from area 9 to the thalamus and internal capsule, in addition to the corpus callosum (Caminiti et al. 2009; Tomasi et al. 2012). Interestingly, this appears not to apply to the projections to the caudate, which receives similarly thin axons from the 2 areas nor to the intra-areal axons.

It remained to be clarified if an area also sends different diameter axons to different targets. The finding that parietal cortex projects thinner cortico-cortical axons to area 24 than that to areas 4 and 6, and the analysis of comparable projections in the 2 animals injected in the motor cortex area 4 indicates unequivocally that they do. Moreover, it appears that axons with different diameters originate from different neuronal populations and proceed in parallel to the different targets.

The relations between origin and termination of cortical axons and their diameters highlight what appear to be new aspects of cortical organization. Together they suggest that the overall picture which should arise from systematic studies of this kind is one of an extraordinary complexity of axonal pathways operating at different conduction speed and generating different conduction delays between brain sites. The complexity is exaggerated by the finding that each projection consists of axons with different diameters, which therefore cause a spectrum of activation delays at their targets. In the evolution of primates, the diameter of axons increases less than the distance between sites caused by brain enlargement and the spectrum of delays increased as well with probable consequences for brain dynamics (Caminiti et al. 2009). An additional question is whether feed-forward and feed-back projections between 2 areas use similar spectra of axon diameters and conduction velocities as this appears to be the case for the V1–V2 projections in the monkey (Girard et al. 2001).

Which could be the consequences of the different conduction properties of cortical axons? One might be the existence of a hierarchical organization of cortical areas in terms of processing speed, the primacy going to the premotor, motor, and somatosensory systems (Tomasi et al. 2012). Unfortunately, at least 3 crucial pieces of information are missing. One is that the axon diameters are roughly proportional to the size of the parent somata (Tomasi et al. 2012), which might have different activation threshold and this could compensate for the delays generated by the axons. The second is that different diameter axons might contact different types of neurons. For example, it was suggested that inhibitory influences between the hemispheres might be carried by faster axons (Makarov et al. 2008), although the slower interhemispheric input may be overwhelmingly excitatory (Makarov et al. 2008; Wunderle et al. 2012). The third unknown concerns the conduction and synaptic properties of the intracortical, terminal portion of the axons and the dynamics of the neuronal compartments they impinge. This includes the neuronal integration time, discussed by Budd and Kisvarday (2012).

A different set of questions, which will be only briefly mentioned here, relates to the mechanisms that control axon diameter. In development, large axon diameters progressively differentiate from an initially uniform or almost uniform distribution of small axons (Berbel and Innocenti 1988; LaMantia and Rakic 1990). What causes this differentiation? Area-of-origin axonal signatures appear to be a new aspect of cortical regionalization and an additional, possibly genetically determined feature of cortical areas. The search for area-specific genes has proven to be a difficult task, particularly in primates (Yamamori and Rockland 2006; Molnár and Clowry 2012), although promising perspectives emerge from complex layer- and area-specific gene clustering analyses (Bernard et al. 2012) and from the study of the differences in the transcriptional activity of certain genes (Konopka et al. 2012). Axonal differences related to the site of termination are probably the expression of the genotypic identity of cortical projection neurons (e.g. Molyneaux et al. 2007; Leone et al. 2008; Shim et al. 2012), but also of possible retrograde signals from the target. Beyond the likely genetic predisposition, neural activity certainly plays a role as shown by the catastrophic consequences of interfering with activity on the development of callosal projections (reviewed in Innocenti 2007; see also Wang et al. 2007).

The cellular mechanisms of axonal radial growth are incompletely known. Cytoskeletal components the neurofilaments, in particular their heavy subunit, but also the intermediate molecular weight subunit, are involved in axonal maturation (Figlewicz et al. 1988) and radial growth (Rao et al. 2003; Perrot et al. 2007 and references therein). It was long believed that all the axonal proteins are synthetized in the cell body, but now unequivocal evidence of axonal translation has accumulated (Koenig 1967; Giuditta et al. 1968; reviewed in Jung et al. 2012), although thus far evidence that the cytoskeletal proteins are also axonally synthetized is lacking. A key element in the regulation of axonal caliber is the oligodendrocyte as shown by the finding that axons fail to grow at sites where oligodendrocytes have been deleted by irradiation (Colello et al. 1994). Some of the molecular mechanisms of the oligodendrocytic action on the axon have recently been discovered (Lee et al. 2012). Oligodendrocyte proliferation is in turn regulated by a host of factors, including the already mentioned neurofilament proteins (Fressinaud et al. 2012). Obviously, the reciprocal interactions between axons, axonal cytoskeleton, oligodendrocytes, and neural activity remain elusive.

It would be important to know how the many complex cellular and molecular events that determine axon diameters and the conduction delays are coordinated and regulated over the whole brain, in development and possibly in adulthood. The counterpart of this question is what the consequences of deregulation might be, a possible cause for misconnectivity syndromes, including schizophrenia (Innocenti et al. 2003), resulting in psychiatric and/or neurological symptoms.

Supplementary Material

Supplementary material can be found at: http://www.cercor.oxfordjournals.org/.

Funding

This work was supported by the European Union contract # 029023, Paul BROCA II (GMI), and the Compagnia di San Paolo (A.V. and R.C.).

Notes

We are grateful to Ms A. Faccio for her excellent technical assistance with EM sectioning. Conflict of Interest. None declared.

References

Barazany
D
Basser
PJ
Assaf
Y
In vivo measurement of axon diameter distribution in the corpus callosum of rat brain
Brain
 , 
2009
, vol. 
132
 (pg. 
1210
-
1220
)
Bauswein
E
Fromm
C
Preuss
A
Corticostriatal cells in comparison with pyramidal tract neurons: contrasting properties in the behaving monkey
Brain Res
 , 
1989
, vol. 
493
 (pg. 
198
-
203
)
Belmalih
A
Borra
E
Contini
M
Gerbella
M
Rozzi
S
Luppino
G
A multiarchitectonic approach for the definition of functionally distinct areas and domains in the monkey frontal lobe
J Anat
 , 
2007
, vol. 
211
 (pg. 
199
-
211
)
Berbel
P
Innocenti
GM
The development of the corpus callosum in the cat. A light– and electron–microscopic study
J Comp Neurol
 , 
1988
, vol. 
276
 (pg. 
132
-
156
)
Bernard
A
Lubbers
LS
Keith
Q
Tanis
KQ
Luo
R
Podtelezhnikov
AA
Finney
EM
McWhorter
MME
Serikawa
K
Lemon
T
, et al.  . 
Transcriptional architecture of the primate neocortex
Neuron
 , 
2012
, vol. 
73
 (pg. 
1083
-
1099
)
Budd
JML
Kisvarday
ZF
Communication and wiring in the cortical connectome
Front Neuroanat
 , 
2012
, vol. 
6
 (pg. 
1
-
23
)
Caminiti
R
Chafee
MV
Battaglia-Mayer
A
Averbeck
B
Crowe
DA
Georgopoulos
AP
Understanding the parietal lobe syndrome from a neurophysiological and evolutionary perspective
Eur J Neurosci
 , 
2010
, vol. 
31
 (pg. 
2320
-
2340
)
Caminiti
R
Ghaziri
H
Galuske
R
Hof
PR
Innocenti
GM
Evolution amplified processing with temporally dispersed, slow neuronal connectivity in primates
Proc Natl Acad Sci USA
 , 
2009
, vol. 
106
 (pg. 
19551
-
19556
)
Cappe
C
Morel
A
Rouiller
EM
Thalamocortical and the dual pattern of corticothalamic projections of the posterior parietal cortex in macaque monkeys
Neuroscience
 , 
2007
, vol. 
146
 (pg. 
1371
-
1387
)
Colello
RJ
Pott
U
Schwab
ME
The role of oligodendrocytes and myelin on axon maturation in the developing rat retinofugal pathway
J Neurosci
 , 
1994
, vol. 
14
 (pg. 
2594
-
2605
)
Dyrby
TB
Søgaard
LV
Hall
MG
Ptito
M
Alexander
DC
Contrast and stability of the axon diameter index from microstructure imaging with diffusion MRI
Magn Reson Med
 , 
2012
 
. [Epub ahead of print]
Figlewicz
DA
Gremo
F
Innocenti
GM
Differential expression of neurofilament subunits in the developing corpus callosum
Dev Brain Res
 , 
1988
, vol. 
42
 (pg. 
181
-
189
)
Fressinaud
C
Berges
R
Eyer
J
Axon cytoskeleton proteins specifically modulate oligodendrocyte growth and differentiation in vitro
Neurochem Int
 , 
2012
, vol. 
60
 (pg. 
78
-
90
)
Girard
P
Hupé
JM
Bullier
J
Feedforward and feedback connections between areas V1 and V2 of the monkey have similar rapid conduction velocities
J Neurophysiol
 , 
2001
, vol. 
85
 (pg. 
1328
-
1331
)
Giuditta
A
Dettbarn
WD
Brzin
M
Protein synthesis in the isolated giant axon of the squid
Proc Natl Acad Sci USA
 , 
1968
, vol. 
59
 (pg. 
1284
-
1287
)
Humphrey
Dr
Corrie
WS
Properties of pyramidal tract neuron system within a functionally defined subregion of primate motor cortex
J Neurophysiol
 , 
1978
, vol. 
41
 (pg. 
216
-
243
)
Innocenti
GM
Von Hofsten
C
Rosander
K
Subcortical regulation of cortical development. Some effects of early cortical deprivation
Progr Brain Res
 , 
2007
, vol. 
Vol. 164
 (pg. 
23
-
37
)
Innocenti
GM
Ansermet
F
Parnas
J
Schizophrenia, development and corpus callosum
Mol Psychiatry
 , 
2003
, vol. 
8
 (pg. 
261
-
274
)
Jung
H
Yoon
BC
Holt
CE
Axonal mRNA localization and local protein synthesis in nervous system assembly, maintenance and repair
Nat Rev Neurosci
 , 
2012
, vol. 
13
 (pg. 
308
-
324
)
Kita
T
Kita
H
The subthalamic nucleus is one of multiple innervation sites for long-range corticofugal axons: a single-axon tracing study in the rat
J Neurosci
 , 
2012
, vol. 
32
 (pg. 
5990
-
5999
)
Koenig
E
Synthetic mechanisms in the axon. IV. In vitro incorporation of [3H]precursors into axonal protein and RNA
J Neurochem
 , 
1967
, vol. 
14
 (pg. 
437
-
446
)
Konopka
G
Friedrich
T
Davis-Turak
J
Winden
K
Oldham
MC
Gao
F
Chen
L
Wang
GZ
Luo
R
Preuss
TM
, et al.  . 
Human-specific transcriptional networks in the brain
Neuron
 , 
2012
, vol. 
75
 (pg. 
601
-
617
)
Kultas-Ilinski
K
Sivan-Loukianova
E
Ilinski
IA
Reevaluation of the primary motor cortex connections with the thalamus in primates
J Comp Neurol
 , 
2003
, vol. 
457
 (pg. 
133
-
158
)
LaMantia
AS
Rakic
P
Cytological and quantitative characteristics of four cerebral commissures in the rhesus monkey
J Comp Neurol
 , 
1990
, vol. 
291
 (pg. 
520
-
537
)
Lee
Y
Morrison
BM
Li
Y
Lengacher
S
Farah
MH
Hoffman
PN
Liu
Y
Tsingalia
A
Jin
L
Zhang
PW
, et al.  . 
Oligodendroglia metabolically support axons and contribute to neurodegeneration
Nature
 , 
2012
, vol. 
487
 (pg. 
443
-
448
)
Lemon
RN
Descending pathways in motor control
Annu Rev Neurosci
 , 
2008
, vol. 
31
 (pg. 
195
-
218
)
Leone
DP
Srinivasan
K
Chen
B
Alcamo
E
McConnell
SK
The determination of projection neuron identity in the developing cerebral cortex
Curr Opin Neurobiol
 , 
2008
, vol. 
18
 (pg. 
28
-
35
)
Makarov
VA
Schmidt
K
Castellanos
NP
Aguado
L
Innocenti
GM
Stimulus–dependent interactions between the visual areas 17 and 18 of the ferret
Cereb Cortex
 , 
2008
, vol. 
18
 (pg. 
1951
-
1960
)
Molnár
Z
Clowry
G
Hofman
MA
Falk
D
Cerebral cortical development in rodents and primates.
Progr Brain Res
 , 
2012
, vol. 
Vol. 195
 (pg. 
45
-
70
)
Molyneaux
BJ
Arlotta
P
Menezes
JR
Macklis
JD
Neuronal subtype specification in the cerebral cortex
Nat Rev Neurosci
 , 
2007
, vol. 
8
 (pg. 
427
-
437
)
Panzeri
S
Brunel
N
Logothetis
NK
Kayser
C
Sensory neural codes using multiplexed temporal scales
Trends Neurosci
 , 
2010
, vol. 
33
 (pg. 
111
-
120
)
Paxinos
G
Huang
X-F
Toga
AW
The rhesus monkey brain
 , 
2000
London
Academic Press
Perrot
R
Lonchampt
P
Peterson
AC
Eyer
J
Axonal neurofilaments control multiple fiber properties but do not influence structure or spacing of nodes of Ranvier
J Neurosci
 , 
2007
, vol. 
27
 (pg. 
9573
-
9584
)
Picard
N
Strick
PL
Motor areas of the medial wall: a review of their location and functional activation
Cerebr Cortex
 , 
1996
, vol. 
6
 (pg. 
342
-
353
)
Rao
MV
Campbell
J
Yuan
A
Kumar
A
Gotow
T
Uchiyama
Y
Nixon
RA
The neurofilament middle molecular mass subunit carboxyl-terminal tail domains is essential for the radial growth and cytoskeletal architecture of axons but not for regulating neurofilament transport rate
J Cell Biol
 , 
2003
, vol. 
163
 (pg. 
1021
-
1031
)
Roberts
JA
Robinson
PA
Modeling distributed axonal delays in mean-field brain dynamics
Phys Rev E Stat Nonlin Soft Matter Phys
 , 
2008
, vol. 
78
 (pg. 
5
-
8
)
Rockland
KS
Further evidence for two types of corticopulvinar neurons
Neuroreport
 , 
1994
, vol. 
5
 (pg. 
1865
-
1868
)
Rouiller
EM
Welker
E
A comparative analysis of the morphology of corticothalamic projections in mammals
Brain Res Bull
 , 
2000
, vol. 
53
 (pg. 
727
-
734
)
Roxin
A
Brunel
N
Hansel
D
Role of delays in shaping spatiotemporal dynamics of neuronal activity in large networks
Phys Rev Lett
 , 
2005
, vol. 
94
 pg. 
238103
 
Schmahamann
JD
Pandya
DN
Fiber pathways of the brain
 , 
2009
New York
Oxford University Press
Shim
S
Kwan
KY
Li
M
Lefebvre
V
Sestan
N
Cis-regulatory control of corticospinal system development and evolution
Nature
 , 
2012
, vol. 
486
 (pg. 
74
-
79
)
Tettoni
L
Georghita-Bachler
F
Bressoud
R
Welker
E
Innocenti
GM
Constant and variable aspects of axonal phenotype in cerebral cortex
Cereb Cortex
 , 
1998
, vol. 
8
 (pg. 
543
-
552
)
Tomasi
S
Caminiti
R
Innocenti
GM
Areal differences in diameter and length of corticofugal projections
Cereb Cortex
 , 
2012
, vol. 
22
 (pg. 
1463
-
1472
)
Wang
N
Warren
S
May
PJ
The macaque midbrain reticular formation sends side-specific feedback to the superior colliculus
Exp Brain Res
 , 
2010
, vol. 
201
 (pg. 
701
-
717
)
Wang
CL
Zhang
L
Zhou
Y
Zhou
J
Yang
XJ
Duan
SM
Xiong
ZQ
Ding
YQ
Activity-dependent development of callosal projections in the somatosensory cortex
J Neurosci
 , 
2007
, vol. 
27
 (pg. 
11334
-
11342
)
Wunderle
T
Eriksson
D
Schmidt
KE
Multiplicative mechanism of lateral interactions revealed by controlling interhemispheric input
Cereb Cortex
 , 
2012
, vol. 
23
 (pg. 
900
-
912
)
Yamamori
T
Rockland
KS
Neocortical areas, layers, connections, and gene expression
Neurosci Res
 , 
2006
, vol. 
55
 (pg. 
11
-
27
)