Evolutionary Trajectories of Cognitive Abilities and of their Putative Neuroanatomical and Allometric Correlates: Testing Novel Hypotheses of Cognitive Evolution and Cognitive Integration with Phylogenetic Comparative Methods

Abstract

The study of primate intelligence, and specifically of general intelligence, has progressed rapidly in the last two decades, however several issues remain unexplored. While neuroanatomical volume measures (NVMs) such as brain size, neocortex size, and absolute or relative size of other brain regions have been frequently framed as substrates for general intelligence, such claims are largely based on simple correlative analyses. Furthermore, while factor analytical techniques have identified a general factor among cognitive abilities when using datasets of species’ mean performances, there have been no examinations of whether the common factor is predictably more strongly present in some primate lineages than in others, and whether relations among brain regions are similarly stronger in some primate lineages than in others. Here, such issues in the comparative literature on primate intelligence are addressed in a new set of studies that present two main novel contributions to the scientific understanding of primate intelligence: First, the evolutionary patterns of the history of changes in general intelligence in primates is examined and compared to those behind the history of changes in brain size and the size of brain regions most commonly used in comparative cognition studies. Studying evolutionary patterns behind a trait permits examining how conserved it is across evolutionary time, how fast it has evolved, and the degree to which it has evolved in a particular direction (i.e., if natural selection regimes have been consistent). Phylogenetic comparative methods employed on datasets of primate species reveal that general intelligence has evolved at a faster pace than NVMs and it has evolved more consistently towards a selection optimum. In contrast to the overall emphasis given in the literature to brain size and neocortex ratio as substrates for intelligence, the NVM with results that most strongly approached the patterns identified for general intelligence is residual cerebellar size (relative to body size). Secondly, a hypothesis is advanced that species higher on general intelligence exhibit a stronger manifold (i.e., higher factor loadings) as general intelligence has previously been empirically associated with ecological generalism, rather than with specialism. As such, cognitive specialization and independence among abilities should be a hallmark of species that have not evolved strong general intelligence. The Continuous Parameter Estimation Model (CPEM) is used in a dataset of cognitive abilities in primate species, and largely confirms the hypothesis forwarded. However, when the same analytical approach is replicated using data on sizes of brain regions, it is found that brain size fails to predict or coevolve with factor loadings of brain regions. Similarly, telencephalon size (a brain structure that holds several of the regions theoretically proposed to serve as substrates for intelligence) fails to predict the strength of factor loadings of telencephalic regions. This set of studies supports the notion that the comparability between the evolution of general intelligence and the evolution of volumes of neuroanatomical structures is more limited than previously thought. Alternative substrates for general intelligence are discussed.