On several occasions in his systematic comparison, Vaesen refers to neuroscientific data to make his argument. In some instances, he underlines the absence of certain functional regions in the brains of non-human primates that are relevant for tool manipulation and production (Orban et al. Reference Orban, Claeys, Nelissen, Smans, Sunaert, Todd, Wardak, Durand and Vanduffel2006). In other sections the author points to the human recruitment of the same specialized cortical regions during different tasks of complex motor behavior such as speech or tool manipulation (Higuchi et al. Reference Higuchi, Chaminade, Imamizua and Kawato2009; Stout et al. Reference Stout, Toth, Schick and Chaminade2008). The latter neuroscientific findings are taken as evidence for a common origin of tool use and language.
The use of neuroimaging and neurobehavioral research to speculate on evolutionary theories of cognition maybe tempting, but for the moment the available information is limited. I suggest that there are three major observations that can be employed in the discussion of neurocognitive evolution in humans: (1) neural overlap, (2) co-lateralization, and (3) cross-functional (effective) connectivity.
Cross-task neural overlap, or neurofunctional overlap, refers to the observation that a single brain region is recruited by different cognitive tasks. In neuropsychological studies, neurofunctional overlap is hinted at by the frequent co-occurrence of cognitive deficits, such as aphasia with apraxia or finger agnosia with acalculia; but lesion research provides only limited spatial resolution. Neuroimaging sparked a much more detailed investigation of the brain's functional organization, including clear cross-task activation in brain regions that can be measured at the 2–4 mm scale. Increased spatial resolution also allowed for a more detailed description of the match of cross-task neural overlap (overlap correspondence), although I'm unaware of systematic studies using this approach. The detection of overlapping neurocognitive circuits in specific cortical locations has been interpreted in terms of a functional and even evolutionary link, for example, between spatial and numerical processing (Hubbard et al. Reference Hubbard, Piazza, Pinel and Dehaene2005; Walsh Reference Walsh2003) or between language and tool use (Arbib Reference Arbib2005; Higuchi et al. Reference Higuchi, Chaminade, Imamizua and Kawato2009).
The question remains whether overlapping neural responses reflect activation of the same or different neuronal populations. Separate neuronal populations may be interleaved in the same cortical area on a spatial scale below the resolution of conventional fMRI, in which case the corresponding neural circuits may function independently, yet show co-morbidity when this region gets damaged or disrupted. Despite this caveat, we may assume that the neural network activated by a given cognitive function is not randomly distributed over the cortex, but that it engages regions that are of strategic relevance for that function given its connections with other regions. Co-activation of the same region by different cognitive tasks therefore at least suggests a strategic similarity in the recruitment of a specific cortical area with its particular connections, that may or may not share neuronal resources. The shared neural localization of certain domain-general skills, such as hierarchical processing, also enticed scholars to theorize on the specificity and chronology of cognitive evolution (Arbib Reference Arbib2005). For the time being, these valuable hypotheses remain to be tested.
Co-lateralization is defined here as the covariance in the side and degree of hemispheric preference of two cognitive functions. Although many functions are asymmetrically represented in the brain, similarity in hemispheric preference as such is generally not considered to reflect a specific functional link, and there exists remarkably little research on the strength of lateralization within, let alone across, cognitively induced neural activation patterns (Pinel & Dehaene Reference Pinel and Dehaene2010). Significant correlations in the degree of asymmetric activation on sites of neural overlap would strengthen claims of biological association between cognitive functions.
A potentially very interesting source of information may be found in people with atypical language lateralization, such as in some extreme left-handers or in patients who suffered early brain damage. In these individuals it is possible to investigate how the atypical language dominance impacts on other lateralized cognitive abilities (Kroliczak et al. Reference Kroliczak, Piper and Frey2011). Recently, we compared a group of atypical language-dominant volunteers with a matched group showing typical language dominance on a tool-pantomiming paradigm while undergoing fMRI. In the group with atypical right language dominance, all individuals also demonstrated atypical right-hemispheric preference for praxis. Activation patterns for the language and praxis tasks revealed neural overlap in five cortical regions that showed highly correlated lateralization indices within and across tasks (Vingerhoets et al., in press).
So far, my arguments focus on the characteristics and interactions of the neural responses induced by different cognitive functions. Similarities in location and co-asymmetry should be supplemented by behavioral evidence of a link between cognitive traits. If two cognitive functions share an evolutionary origin, it is plausible to assume that they exhibit a functional bond over and above a common reliance on central resources such as attention and working memory. If, for example, tool use and linguistic tasks activate Broca's area (neural overlap) because they both require hierarchical structuring (underlying cognitive process), then we might expect behavioral interference between tool use and language tasks that manipulate hierarchical complexity.
Statistical dependencies in performance or neural activity only suggest a functional relation between cognitive traits or neural units, they do not entail causal information. Over the last years, several methods have been devised to investigate effective (causal) brain connectivity (Rubinov & Sporns Reference Rubinov and Sporns2010). In view of evolutionary queries, directional effects are of importance, as they may hint at the temporal order of cognitive evolvement. Similarities in the directional interactions of networks of related cognitive functions and causal effects of cross-task interference may help elucidate the chronological sequence of neurocognitive evolution, such as the link between gestures and speech to explain the evolution of language.
I conclude that neuroscientific research on the cerebral organization of cognitive function in modern humans may contribute in unraveling the evolutionary trace of unique abilities such as tool use, language, and numerical cognition. Available methods for this endeavor include (1) detailed analysis of the neural overlap of activity patterns elicited by allied cognitive functions, (2) investigation of the correlation of co-lateralization in direction and degree across cognitive abilities that have an asymmetric hemispheric representation, and (3) comparison of the causal interactions in the neural networks of related cognitive functions and their cross-functional interference.
On several occasions in his systematic comparison, Vaesen refers to neuroscientific data to make his argument. In some instances, he underlines the absence of certain functional regions in the brains of non-human primates that are relevant for tool manipulation and production (Orban et al. Reference Orban, Claeys, Nelissen, Smans, Sunaert, Todd, Wardak, Durand and Vanduffel2006). In other sections the author points to the human recruitment of the same specialized cortical regions during different tasks of complex motor behavior such as speech or tool manipulation (Higuchi et al. Reference Higuchi, Chaminade, Imamizua and Kawato2009; Stout et al. Reference Stout, Toth, Schick and Chaminade2008). The latter neuroscientific findings are taken as evidence for a common origin of tool use and language.
The use of neuroimaging and neurobehavioral research to speculate on evolutionary theories of cognition maybe tempting, but for the moment the available information is limited. I suggest that there are three major observations that can be employed in the discussion of neurocognitive evolution in humans: (1) neural overlap, (2) co-lateralization, and (3) cross-functional (effective) connectivity.
Cross-task neural overlap, or neurofunctional overlap, refers to the observation that a single brain region is recruited by different cognitive tasks. In neuropsychological studies, neurofunctional overlap is hinted at by the frequent co-occurrence of cognitive deficits, such as aphasia with apraxia or finger agnosia with acalculia; but lesion research provides only limited spatial resolution. Neuroimaging sparked a much more detailed investigation of the brain's functional organization, including clear cross-task activation in brain regions that can be measured at the 2–4 mm scale. Increased spatial resolution also allowed for a more detailed description of the match of cross-task neural overlap (overlap correspondence), although I'm unaware of systematic studies using this approach. The detection of overlapping neurocognitive circuits in specific cortical locations has been interpreted in terms of a functional and even evolutionary link, for example, between spatial and numerical processing (Hubbard et al. Reference Hubbard, Piazza, Pinel and Dehaene2005; Walsh Reference Walsh2003) or between language and tool use (Arbib Reference Arbib2005; Higuchi et al. Reference Higuchi, Chaminade, Imamizua and Kawato2009).
The question remains whether overlapping neural responses reflect activation of the same or different neuronal populations. Separate neuronal populations may be interleaved in the same cortical area on a spatial scale below the resolution of conventional fMRI, in which case the corresponding neural circuits may function independently, yet show co-morbidity when this region gets damaged or disrupted. Despite this caveat, we may assume that the neural network activated by a given cognitive function is not randomly distributed over the cortex, but that it engages regions that are of strategic relevance for that function given its connections with other regions. Co-activation of the same region by different cognitive tasks therefore at least suggests a strategic similarity in the recruitment of a specific cortical area with its particular connections, that may or may not share neuronal resources. The shared neural localization of certain domain-general skills, such as hierarchical processing, also enticed scholars to theorize on the specificity and chronology of cognitive evolution (Arbib Reference Arbib2005). For the time being, these valuable hypotheses remain to be tested.
Co-lateralization is defined here as the covariance in the side and degree of hemispheric preference of two cognitive functions. Although many functions are asymmetrically represented in the brain, similarity in hemispheric preference as such is generally not considered to reflect a specific functional link, and there exists remarkably little research on the strength of lateralization within, let alone across, cognitively induced neural activation patterns (Pinel & Dehaene Reference Pinel and Dehaene2010). Significant correlations in the degree of asymmetric activation on sites of neural overlap would strengthen claims of biological association between cognitive functions.
A potentially very interesting source of information may be found in people with atypical language lateralization, such as in some extreme left-handers or in patients who suffered early brain damage. In these individuals it is possible to investigate how the atypical language dominance impacts on other lateralized cognitive abilities (Kroliczak et al. Reference Kroliczak, Piper and Frey2011). Recently, we compared a group of atypical language-dominant volunteers with a matched group showing typical language dominance on a tool-pantomiming paradigm while undergoing fMRI. In the group with atypical right language dominance, all individuals also demonstrated atypical right-hemispheric preference for praxis. Activation patterns for the language and praxis tasks revealed neural overlap in five cortical regions that showed highly correlated lateralization indices within and across tasks (Vingerhoets et al., in press).
So far, my arguments focus on the characteristics and interactions of the neural responses induced by different cognitive functions. Similarities in location and co-asymmetry should be supplemented by behavioral evidence of a link between cognitive traits. If two cognitive functions share an evolutionary origin, it is plausible to assume that they exhibit a functional bond over and above a common reliance on central resources such as attention and working memory. If, for example, tool use and linguistic tasks activate Broca's area (neural overlap) because they both require hierarchical structuring (underlying cognitive process), then we might expect behavioral interference between tool use and language tasks that manipulate hierarchical complexity.
Statistical dependencies in performance or neural activity only suggest a functional relation between cognitive traits or neural units, they do not entail causal information. Over the last years, several methods have been devised to investigate effective (causal) brain connectivity (Rubinov & Sporns Reference Rubinov and Sporns2010). In view of evolutionary queries, directional effects are of importance, as they may hint at the temporal order of cognitive evolvement. Similarities in the directional interactions of networks of related cognitive functions and causal effects of cross-task interference may help elucidate the chronological sequence of neurocognitive evolution, such as the link between gestures and speech to explain the evolution of language.
I conclude that neuroscientific research on the cerebral organization of cognitive function in modern humans may contribute in unraveling the evolutionary trace of unique abilities such as tool use, language, and numerical cognition. Available methods for this endeavor include (1) detailed analysis of the neural overlap of activity patterns elicited by allied cognitive functions, (2) investigation of the correlation of co-lateralization in direction and degree across cognitive abilities that have an asymmetric hemispheric representation, and (3) comparison of the causal interactions in the neural networks of related cognitive functions and their cross-functional interference.