Introduction and context

When asked to give an account of his creative process, Albert Einstein reported that language played virtually no role in his thought process. ‘The psychical entities which seem to serve as elements in thought ... are, in my case, of visual and some of muscular type’ [1]. When Einstein’s brain was examined postmortem, it was noted that among other irregularities, the inferior parietal lobe was significantly larger than normal [2].

The posterior parietal cortex (PPC) plays a central role in multisensory integration [3] and environmental-spatial cognition [4]. Many recent human imaging (and also some monkey neurophysiology) studies demonstrate that the PPC additionally supports various forms of high-order non-spatial cognition that are not necessarily directly related to physical space itself. For example, while the superior parietal lobule (SPL) tends to process spatial information in a conventional way, the inferior parietal lobule (IPL) is often credited with non-spatial cognition. Debate exists on the comparative anatomy of the primate PPC (Figure 1). The IPL is said to be evolutionarily new and uniquely expanded in humans and the monkey PPC is said to correspond to the human SPL [5,6], but there is another view that the monkey PPC possesses functional homologues of both regions [7,8]. If the latter view is correct, neuroscientists should be able to identify precursory mechanisms of non-spatial information processing in non-human primates. Indeed, when monkeys were trained to acquire tool use - a high-order behavior they rarely exhibit naturally - PPC expansion was observed at both microscopic [9] and macroscopic [10] scales.

Figure 1.

Comparison of substructures (unshaded) of monkey and human posterior parietal cortex (PPC)

(a) monkey PPC; (b) human left PPC; (c) human right PPC. Results of meta-analysis (based on the references listed in Figure 2) of cortical areas responsible for non-spatial cognition are superimposed. Data points in monkeys are projected onto the left hemisphere (because no particular laterality has been claimed), whereas those in human subjects are illustrated independently for each hemisphere (being biased toward the left hemisphere). Large and dashed symbols indicate areas estimated from reviews and original papers, respectively, in which coordinates were not clearly specified. Arabic numerals indicate Brodmann areas (40: supramarginal gyrus; 39: angular gyrus). The subdivisions of the PPC were referred from the original sources. AIP, anterior intraparietal area; CIP, caudal intraparietal area; CS, central sulcus; IPS, intraparietal sulcus; LIP, lateral intraparietal area; MIP, medial intraparietal area; PCS, post central sulcus; POS, parieto-occipital sulcus; SF, sylvian fissure; STS, superior temporal sulcus; VIP, ventral intraparietal area.

The following section reviews various recently reported forms of non-spatial representation in the PPC. They are ordered roughly in terms of the level of abstraction of the ‘objects’ and pseudo-spatial relations represented (Figure 2). The last section considers potential mechanisms that could have enabled the hominid PPC to establish such representations and hypothesizes a general principle to unify these newer forms of representation with the PPC’s original functions.

Figure 2.

Hypothetical mechanisms of non-spatial representations processed by posterior parietal cortex as a function of abstractness

Left, hypothesized mechanism for increasing abstract levels of representations (see ‘Future directions’ section for details). Right, structure of various levels of abstract spaces (references for respective levels are listed in the far-right column). 3D, three-dimensional; 4D, four-dimensional; AG, angular gyrus; AIP, anterior intraparietal area; CIP, caudal intraparietal area; IPL, inferior parietal lobule; IPS, intraparietal sulcus; LIP, lateral intraparietal area; PPC, posterior parietal cortex; SII, secondary somatosensory cortex; SMG, supramarginal gyrus; SPL, superior parietal lobule; TPJ, temporo-parietal junction.

Major recent advances

Future directions

As outlined above and illustrated in Figure 2 (right), various kinds of non-spatial cognition can be grouped and ordered based on the levels of abstraction of the ‘objects’ and ideally defined spaces represented. Assumed coordinate systems for such ‘spaces’ are summarized here.

The pseudo-spatial nature of the high-order cognition supported by the PPC may derive from the essential characteristics of the objects represented but alternatively may derive from the nature of the PPC’s pre-existing information-processing mechanisms, namely as a hub for multisensory integration and representing physical environmental space. The meta-analysis in Figure 1 illustrates that the PPC areas responsible for these novel forms of cognition are not necessarily clearly segregated, either in monkeys or humans, but does suggest a trend of gradual expansion toward the IPL as the level of abstraction proceeds. Thus, it seems that the PPC gradually incorporated high-order cognition as it expanded during hominid evolution while preserving its original principles of operation.

What is the explanation for this expansion of function? Here is a hypothesis involving mechanisms for ‘selecting’ and ‘switching’ between objects among different represented spaces: (a) in classical (physical) space, spatial attention toward concrete objects was typically expressed as eye movement; (b) when such attention needed to be sustained or when the attending content needed to be memorized, invisible ‘time’ was ‘visualized’ in the mind’s eye, becoming a new virtual dimension in the PPC’s existing suite of spatial coordinate systems; (c) once the PPC could visualize an invisible virtual entity, a similar objectification process could be extended further, enabling intentional perspective switching (Figure 2, left).

Acquiring representations of social space might have accelerated this process. Through objectification processes [53] and the development of ‘virtual eyes’ [48,73], flexible and mutually integrated representations of the bodily self, the analogous selves of others, and tools as equivalents of body parts (and vice versa) might have served as a bridge between concrete physical and abstract conceptual spaces. As the PPC expanded in both physical volume and range of function [9,10] (Figure 2, dashed circle), a positive feedback process (arrow) could have been established to achieve further human-specific forms of non-spatial conceptual cognition.

Thus, crucial components of human intelligence derive their character from the precursorial spatial cognition process of the PPC. Indeed, language is rife with spatial metaphors for abstract thoughts. Non-human primates possess the same precursorial parietal mechanisms as human ancestors; through training, their cognitive capacities can be artificially extended in the hominid/human direction. If correct, this hypothesis will be valuable in guiding the design of future experimental paradigms, in both animal and human studies, to elucidate the principles of parietal information processing.