3.1 Topology, thickness and myeloarchitecture in night monkey cerebral cortex
Nine night monkeys were scanned using 3T MRI and a NHP 24-channel multi-array RF coil (Autio et al. 2020), the T1w and T2w images were processed using the NHP version of the HCP pipelines and averaged to generate the NightMonkeyRIKEN-KU9 dataset (Supp. Figure 1). The total volume of the night monkey brain was 18.2 ± 1.2 cm3, the volume of cortex (per hemisphere) was 3.73 ± 0.30 cm3, and the total surface area of the cortical midthickess (per hemisphere) was 20.3 ± 1.3 cm2 (Fig. 1A). The cortical pial surface (Fig. 1A) shows a distinct lateral fissure that extends to the dorsoposterior part of the brain, a superior temporal sulcus, a relatively short cingulate sulcus, a central dimple (but no central sulcus) and arcuate and principal dimples in dorsal and ventral prefrontal cortex. These cortical features were consistently observed across all night monkeys. The midthickness surface (Fig. 1B), very inflated surface (Fig. 1C) and flatmap (Fig. 1D) also facilitated visualization of large proportions of cerebral cortex buried inside the sulci (e.g. parietal cortex within lateral fissure, medial occipital lobe), albeit with more distorted vertex areas.
The cortical thickness maps shown in Fig. 2A indicate that cerebral cortex is relatively thick in much of prefrontal and lateral parietal cortex, and in both superior and inferior temporal gyri. It is thin in early sensory areas, including occipital cortex, auditory cortex and somatosensory cortex. The average cortical thickness is 1.91 ± 0.04 mm (N = 9), and the lower 5th percentile of cortical thickness in the group average was 1.27 mm. Thus, our image resolution (0.25 x 0.25 x 0.5) was well within the criterion of containing at least two voxels within the thinnest parts of the cortex (Glasser et al. 2016; Autio et al. 2021).
The T1w/T2w myelin maps shown in Fig. 2B show relatively heavy myelination in the primary motor (M1) and somatosensory areas (S1) close to the central dimple, primary auditory (A1) and surrounding auditory cortex, early visual areas including primary visual cortex (V1), the middle temporal complex (MT+), retrosplenial cortex (RSC), and the dorsomedial (DM) visual area (Fig. 2B). T1w/T2w myelin contrast was moderate in the ventroposterior parietal (VPP) area and frontal eye field (FEF) and relatively low in association areas (e.g. prefrontal, medial parietal, insular, and lateral temporal cortices). These trends in T1w/T2w myelin contrast are consistent with other primate species (Glasser et al., 2014; Van Essen et al. 2019; Autio et al. 2020; Hayashi et al. 2021).
The bottom row of Fig. 2 compares T1w/T2w myelin contrast (Fig. 2D) and histological myelin staining density (Sereno et al. 2015) from a tangentially cut flattened cortex (Fig. 2E). T1w/T2w exhibits dense myelination in MT + complex, DM, and VPP, surrounded by a more lightly myelinated lateral parietal cortex. In particular, the MT + complex (analogous to Sereno and colleagues area MT and MSTd; Fig. 2E) exhibits a sharp transition in histological myelin density relative to the surrounding cortex. The lateral bank of the posterior lateral fissure is lightly myelinated and is surrounded by the higher myelin areas such as MT + complex, DM, and VPP (Fig. 2D). Overall, the T1w/T2w myelin contrast (Fig. 2D) and histological myelin density (Fig. 2E) have similar spatial distributions. However, a quantitative validation is hampered by different distortions in the two flat-maps (Fig. 2D and E).
3.2 Interspecies comparison of parieto-temporal cortex
Cortical T1w/T2w myelin contrast, thickness, and their gradients were used to evaluate areal boundaries in macaque, night and marmoset monkeys scanned and pre-processed using the harmonized HCP-NHP methodology (Autio et al. 2020; Hayashi et al. 2021; Ose et al. 2022). In each species, the posterior temporal cortex contained a very heavily myelinated region (Fig. 3A-F, pink border) surrounded by robust gradient-ridges (Fig. 3G-I). This highly myelinated inland likely corresponds to the MT + complex, which includes middle temporal areas MT and MST (Tootell et al. 1985; Desimone and Ungerleider 1986; Large et al. 2016). The medium-to-high T1w/T2w myelin contrast from the posterior bank of lateral fissure to the top of superior temporal gyrus was defined as the auditory cortex in each species (Fig. 3), which adjoins a moderately myelinated retroinsular area located rostromedially (Lewis and Van Essen 2000a). This myelinated area surrounded by strong T1w/T2w myelin gradients likely includes primary auditory cortex (A1) and its surrounding regions such as rostral field (R), caudomedial field (CM), and caudolateral (CL) in night (Imig et al. 1977; Morel and Kaas 1992), macaque (Hackett et al. 1998), and marmoset (de la Mothe et al. 2006) monkeys. Dorso-medial to the MT + complex and auditory cortex, there is an island of relatively low T1w/T2w myelin values (Fig. 3A-F, cyan border) mostly surrounded by robust gradient-ridges (Fig. 3G-I) in each species. In night monkeys, these transitions are supported by histological myelin stain density which also exhibits an island of sparse myelination surrounded by sharp myelin density transitions to the densely myelinated cortex (Fig. 2F). This sparsely myelinated region may correspond to Brodmann area 7 (BA7) complex, which in the macaque includes areas 7a, 7b and 7op and in humans likely even more areas (Yokoyama et al. 2021). We calculated the surface area, average thickness, and cortical volume of these three parieto-temporal parcels for each species along with primary visual cortex (V1), and tested interspecies difference using two-way analysis of variance (ANOVA) with species (macaque, night monkey, marmoset) and cortical parcel (MT + complex, auditory cortex, BA7, V1; see Section 2.3.3). All variables of relative surface area, average thickness, and relative cortical volume showed significant interaction effect between species and cortical parcel (F6, 476 = 2716, 77, 1218, respectively. p < 0.001), indicating that patterns of species effects are different among cortical parcels.
The estimated surface area (per hemisphere) of the MT + complex was 89.8, 47.9, and 12.5 mm2 in macaque, night monkey, and marmoset, respectively, in reasonable agreement with previous reports (Table 1). Relative to total cortical surface area, MT + complex was substantially larger in night monkeys (47.9 ± 2.3 mm2 / 20.3 ± 1.3 cm2 = 2.4%) in comparison to macaque (89.8 ± 12.8 mm2 / 98.9 ± 14.7 cm2 = 0.9%) and marmoset monkeys (12.5 ± 1.5 mm2 / 10.5 ± 0.6 cm2 = 1.2%) (p < 0.001 t-test, Bonferroni corrected) (Fig. 4A). The average cortical thickness of the MT + complex was similar (≈ 2.0 mm) across the three species. Thus, the fractional volume of MT + complex compared to the total volume of cortex is significantly larger in night monkeys (2.6%) in comparison to macaque (0.8%) and marmoset (1.5%) monkeys (p < 0.001 t-test, Bonferroni corrected; see Section 2.3.3).
Table 1
Species comparisons of surface areas of MT + complex, V1, and auditory cortex.
Species | Cortical parcel of interest | Surface area (mm2) (N: number of hemispheres investigated) | Methods | Reference |
Macaque | MT + complex /MT (*) | 89.8 ± 12.8 (N = 64) | T1w/T2w myelin | Current study |
83.1 (N = 4) | Myelin staining (modified Heidenhain-Woelke method) | (Gattass and Gross 1981) |
68 (N = 1)* | Anterograde neuronal tracing from V1 (3H-proline) | (Weller and Kaas 1983) |
76 (N = 4)* | Myelin staining (Gallyas or Spielmeyer method) | (Ungerleider and Desimone 1986) |
39 (N = 3)* | Myelin staining (Gallyas method) | (Maunsell and van Essen 1987) |
73 (N = 10) | Cytochrome oxidase activity | (Sincich et al. 2003) |
78 (N = 6) | Myelin staining (Gallyas method) | (Large et al. 2016) |
V1 | 1156.9 ± 130.7 (N = 64) | T1w/T2w myelin | Current study |
1090 (N = 1) | Nissl & myelin staining (modified Weigert method) | (Van Essen and Maunsell 1980) |
823 (N = 2) | Myelin staining (modified Heidenhain-Woelke method) & recording | (Gattass et al. 1981) |
955 (N = 1) | Electrical recording | (Weller and Kaas 1983) |
1195 (N = 31) | Electrical recording | (Van Essen et al. 1984) |
1343 (N = 11) | Cytochrome oxidase activity | (Sincich et al. 2003) |
Auditory cortex | 57.7 ± 11.3 (N = 64) | T1w/T2w myelin | Current study |
88 (N = 10) | Cytochrome oxidase activity | (Sincich et al. 2003) |
Night monkey | MT + complex /MT (*) | 47.9 ± 2.3 (N = 18) | T1w/T2w myelin | Current study |
37 (N = 14)* | Cytochrome oxidase activity | (Tootell et al. 1985) |
V1 | 381.3 ± 32.2 (N = 18) | T1w/T2w myelin | Current study |
286 | Electrical recording | (Myerson et al. 1977) |
400 (N = 14) | Cytochrome oxidase activity | (Tootell et al. 1985) |
Auditory cortex | 51.2 ± 2.9 (N = 18) | T1w/T2w myelin | Current study |
48 (N = 22) | Nissl & electrical recording | (Imig et al. 1977) |
Marmoset | MT + complex | 16.1 ± 1.7 (N = 40) | T1w/T2w myelin | Current study |
14 (N = 6) | Myelin staining (modified Heidenhain-Woelke method) | (Pessoa et al. 1992) |
V1 | 215.2 ± 13.2 (N = 40) | T1w/T2w myelin | Current study |
182 (N = 6) | Myelin staining (modified Heidenhain-Woelke method) | (Pessoa et al. 1992) |
194 (N = 5) | Nissl staining | (Missler et al. 1993) |
205 (N = 4) | Electrical recording | (Fritsches and Rosa 1996) |
Auditory cortex | 12.2 ± 1.3 (N = 40) | T1w/T2w myelin | Current study |
8–12 (N = 5) | Electrical recording | (Aitkin et al. 1986) |
Note that the definition of each region varies across studies. *Studies specifically focusing on area MT only. Area estimates using histology may also be underestimated due to brain shrinkage, however, some of the studies compensated for that: 12% (Ungerleider and Desimone 1986), 16% (Van Essen and Maunsell 1980; Van Essen et al. 1984; Maunsell and van Essen 1987; Pessoa et al. 1992), 20–35% (Imig et al., 1977), unspecified (Fritsches and Rosa 1996). |
Auditory cortex showed similar trends with the MT + complex. The estimated surface area (per hemisphere) of the auditory cortex was 57.7, 51.2 and 16.1 mm2 in macaque, night monkey, and marmoset, respectively, which are also in good agreement with previous reports (Table 1). The relative surface area of auditory cortex was larger in night monkey (51.2 ± 2.9 mm2 / 20.3 ± 1.3 cm2 = 2.5%) than macaque (57.7 ± 11.3 mm2 / 98.9 ± 14.7cm2 = 0.6%) and marmoset monkeys (16.1 ± 1.7 mm2 / 10.5 ± 0.6 cm2 = 1.5%) (p < 0.01 t-test, Bonferroni corrected) (Fig. 4B). The average cortical thickness of the auditory cortex was comparable in night and macaque monkeys (≈ 2.0 mm), however, it was slightly thinner in marmosets (≈ 1.8 mm). The fractional volume of auditory cortex relative to the total volume of the cortex was significantly larger in night monkeys (2.7%) in comparison to macaque (0.5%) and marmoset (1.6%) monkeys (p < 0.001 t-test, Bonferroni corrected).
In contrast, BA7 showed a contrasting pattern of interspecies difference. The relative surface area of BA7 was substantially smaller in night monkeys (45.8 ± 2.6 mm2 / 20.3 ± 1.3 cm2 = 2.3%) in comparison to macaque (320 ± 54 mm2 / 98.9 ± 14.7 cm2 = 3.2%), but larger than in marmoset monkeys (20.6 ± 2.0 mm2 / 10.5 ± 0.6 cm2 = 2.0%) ( p < 0.01 t-test, Bonferroni corrected) (Fig. 4C). The average cortical thickness of the BA7 was 2.4, 2.2, and 2.1 mm in macaque, night and marmoset monkeys, respectively. Accordingly, the volume of the BA7 relative to the total volume of the cortex was smaller in night monkeys (2.6%) in comparison to macaque (3.3%) (p < 0.001 t-test, Bonferroni corrected).
The difference pattern of interspecies effects may reflect the difference between sensory area (MT + complex and auditory cortex) and association area (BA7). To control for this, we also compared V1, a heavily myelinated visual area in the occipital cortex (Supp. Figure 2). The estimated surface area (per hemisphere) of V1 was approximately 1160, 380, and 220 mm2 in macaque, night monkey, and marmoset, respectively. Although the V1 boundaries were less clear than those of the MT + complex and auditory cortex, our estimates are comparable to previous reports (Table 1). The relative surface area of V1 was significantly larger in night monkeys (381 ± 32 mm2 / 20.3 ± 1.3 cm2 = 18.8%) in comparison to macaques (1160 ± 181 mm2 / 98.9 ± 14.7 cm2 = 11.7%), but smaller than in marmosets (215 ± 13 mm2 / 10.5 ± 0.6 cm2 = 20.4%) (p < 0.001 t-test, Bonferroni corrected) (Fig. 4D). The average cortical thickness of V1 was significantly thinner in night monkeys (1.5 mm) and marmosets (1.4 mm) compared to macaques (2.0 mm). Accordingly, the volume of V1 relative to the total volume of cortex was 10.3, 15.3, and 18.3% in macaques, night and marmoset monkeys, respectively. Thus, unlike the MT + complex and auditory cortex, the relative surface area/cortical volume of V1 was distinct from all three of the other areas: smallest in macaques and largest in marmosets. These results suggest that the expansion of the MT + complex and auditory cortex in night monkeys is specific to those regions, and not the result of general expansion of sensory/visual areas.