Jeffery et al. ignore half of navigation, namely navigation in egocentric frameworks. Many of the mechanisms of this type of navigation – as important as allocentric navigation, as walking along a cliff demonstrates – have been investigated in primates, but are conspicuously absent from the present review. Yet, from an evolutionary and genetic standpoint, we humans are obviously much closer to apes and monkeys than to rodents. Monkeys typically live and climb in trees and, not surprisingly, possess a three-dimensional (3D) space representation in their posterior parietal cortex (PPC) (Andersen & Buneo Reference Andersen and Buneo2002). The PPC is dramatically expanded in primates compared to rodents, in which primary cortices V1 and S1 lie very close, leaving little room for the extrastriate or posterior parietal cortex (Whitlock et al. Reference Whitlock, Sutherland, Witter, Moser and Moser2008). Single-cell studies in monkeys have revealed neurons selective for optic flow components in several parietal regions: middle superior temporal dorsal (MSTd: Lagae et al. Reference Lagae, Maes, Raiguel, Xiao and Orban1994; Tanaka et al. Reference Tanaka, Hikosaka, Saito, Yukie, Fukada and Iwai1986), ventral intraparietal (VIP: Schaafsma & Duysens Reference Schaafsma and Duysens1996), and medial PPC (Squatrito et al. Reference Squatrito, Raffi, Maioli and Battaglia-Mayer2001). Human imaging studies suggest that additional regions, such as V6, might be involved (Wall & Smith Reference Wall and Smith2008). DeAngelis, Angelaki, and coworkers have compared visual and vestibular selectivities of neurons in MSTd, VIP, and visual posterior sylvian (VPS) for rotation and translation in three dimensions (see Chen et al. Reference Chen, DeAngelis and Angelaki2010; Reference Chen, DeAngelis and Angelaki2011a; Reference Chen, DeAngelis and Angelaki2011b; Takahashi et al. Reference Takahashi, Gu, May, Newlands, DeAngelis and Angelaki2007). Whereas MSTd and VIP neurons can support the perception of heading, neurons in VIP and VPS support those of self-rotation. In these areas, all directions in space are represented with no clustering of preferences close to the horizontal plane, as would be expected if the vertical direction were poorly represented. This is unsurprising as in primates equipped with binocular vision the z-axis (depth) is unique, not the vertical axis, which unlike depth is represented on the retina. Indeed, particularly in VIP and to a lesser degree in MSTd, the depth axis was under-represented among preferred directions. Preferred directions and tuning widths of MSTd neurons, however, were non-uniformly distributed along the horizontal meridian, explaining the decrease in behavioral sensitivity to self-motion direction with increasing eccentricity in humans and monkeys (Gu et al. Reference Gu, Fetsch, Adeyemo, DeAngelis and Angelaki2010).

In their quest to demonstrate that humans mainly use a horizontal reference frame for navigation, Jeffery et al. quote psychophysical studies of 3D slope perception and the misperception of vertical slopes they reveal. Contrary to the authors' claims, this misperception is part of a larger problem, depth reconstruction (see above), which is complicated by the dependence of distance scaling on the derivation order of disparity. Indeed, human subjects have no precise representation of metric 3D structure, whether using stereo or monocular cues (Todd & Norman Reference Todd and Norman2003). In no way can these misperceptions be taken as evidence that neuronal mechanisms for estimating 3D slopes do not exist. In fact, such neurons have been documented in caudal intraparietal (CIP), part of the PPC (Taira et al. Reference Taira, Tsutsui, Jiang, Yara and Sakata2000), and also in the inferotemporal cortex (Liu et al. Reference Liu, Vogels and Orban2004) – thus, in both the dorsal and ventral visual systems. Similarly, selectivity for second-order disparity gradients has also been documented (see Orban Reference Orban2011 for review). Some of the cortical areas in which these higher-order neurons have been recorded may not even exist in rodents, again underscoring the importance of using adequate models for the human brain. Together, these results indicate that the primate PPC includes representations in three dimensions of subject motion and of the layout of the environment, two critical pieces of information for navigating through 3D space.

The starting point of the target article is the importance of the cognitive map in the hippocampus for spatial navigation. Even if grid cells have recently been identified in monkey entorhinal cortex (Killian et al. Reference Killian, Jutras and Buffalo2012), the hippocampal system of primates bears only a vague resemblance to that of rodents. Relatively small numbers of human hippocampal neurons are place cells, and these occur in several brain regions (Jacobs et al. Reference Jacobs, Kahana, Ekstrom, Mollison and Fried2010). This is not surprising, as the hippocampus receives major projections from the ventral stream (Kravitz et al. Reference Kravitz, Saleem, Baker, Ungerleider and Mishkin2013), which in primates processes information about objects at least as much as scenes, and even information concerning conspecifics. Indeed, considerable evidence indicates that the human hippocampal formation processes not only spatial, but also non-spatial and even non-visual information (Liang et al. Reference Liang, Wagner and Preston2013). This species difference may be even further underestimated, as the human brain evolved to accommodate bipedalism, and thus navigation in open spaces, but such navigation has so far been little investigated for technical reasons. On the other hand, navigation supposes a transformation of coordinate information from spatial maps into body-centered representations, a function likely to be subserved by the PPC, which is connected to the hippocampal formation via the retrospenial cortex in rodents (Whitlock et al. Reference Whitlock, Sutherland, Witter, Moser and Moser2008). A similar route links the primate precuneus and adjacent dorsal superior parietal lobule with the hippocampal system (Kravitz et al. Reference Kravitz, Saleem, Baker and Mishkin2011). The involvement of these medio-dorsal PPC regions in the control and planning of navigation is supported by their activation during observation of climbing and locomotion in humans (Abdollahi et al. Reference Abdollahi, Jastorff and Orban2012). Hence, I propose a much more balanced view, whereby spatial navigation is both allocentric and egocentric and is controlled by both the hippocampal formation (fed by the qualitative scene analysis in parahippocampal gyrus where gravity is not incorporated) and the metric space representation in the PPC (in which gravity is incorporated). These two streams of information are integrated in the dorso-medial PPC devoted to planning locomotor body movements. A recent imaging study indicates that this scheme, clearly different from the bicoded scheme of Jeffery et al., applies to humans (Indovina et al. Reference Indovina, Maffei, Pauwels, Macaluso, Orban and Lacquaniti2013). To what extent it may also apply to rodents is unclear but, intriguingly, one of the subregions of the rodent PPC (the anteromedial visual area; Torrealba & Valdes Reference Torrealba and Valdes2008) may be a precursor of primate V6 and supply optic flow information.

In conclusion, including nonhuman primates among the animal models is essential for understanding the human brain, even for navigation and spatial memory, avoiding errors induced by jumping from rodents to humans as these authors have done.