Friday, 11 October 2013

Are abstract action words embodied? An fMRI investigation at the interface between language and motor cognition.


Katrin Sakreida et al. (2013)

This substantial fMRI study was very interesting to read. Recently my own thoughts about language are  leaning towards the embodied cognition camp as I am increasingly struggling with the idea that higher processing of language should be considered as 'pure thought'. I don't see any function, cognitive process of the brian to be independent of each other and working in isolation. What I see is an establishing idea that the brain is a network (a really complicated one) that includes a multitude of sensory facets that in themselves compose our conscious experiences. Is it a false dichotomy when one segregates higher level comprehension from basic low level discriminative processing? I think the embodied view of the brain provides a way of avoiding having to make such a distinction. Here, the comprehension of language is claimed to be really shaped by the body, the visuo-motor system for example.

This paper uses fMRI and a novel stimuli paradigm to explore the neural correlates of concreteness through to abstractness, along a continuum. Somewhere along the middle of the continuum, results replicated previous findings of sensorimotor activation in the pre central gyrus and the Supplementary motor area. At the concrete end, they found associations with the pars triangularis of the left IFG and the left inferior parietal cortex, suggesting the recruitment of a fronto-parietal sensori-motor network. Whereas the purely abstract stimuli elicited prominent activations in the left ant MTG, a classical region of the language system functionally associated with semantic categorisation, and recognition.

This dissociation between concrete and abstract functional processing suggests that concrete and 'cognitively graspable' concepts within language are so due to the recruitment of sensori-motor representations. Whereas the less concrete, abstract, concepts are themselves further removed and provide less of an association between the concept and potential sensori-motor activity and thus the representations are supported by the linguistic systems. 

Sakreida, K., Scorolli, C., Menz, M. M., Heim, S., Borghi, A. M., & Binkofski, F. (2013). Are abstract action words embodied? An fMRI investigation at the interface between language and motor cognition. Frontiers in human neuroscience7.

FIGURE3|Directcontrastsofconcretenessvs.abstractness. Differencesbetweenprocessingconcretenoun-verbcombinations(top panel)comparedtoabstractnoun-verbcombinations(bottompanel)and extractedcontrastvaluesforthepureabstract,thesummarizedmixed conditionsandthepureabstractconditionfromdefinedlocalmaxima. Note thatforvisualizationthestatisticalimageswerethresholdedat p < 0.001,uncorrected,withanextendedclustersizeof ≥45 contiguous voxels(360mm3), superimposedontheMNItemplateusingthe softwareMRIcronVersion12/2012( mricro/mricron/index.html). Thecontrastvalueswereextractedfromthe individualbetaimagesandaredepictedasgroupmeanwithstandard deviationofthemean.Asterisksindicatestatisticaldifferencesof post-hoc paired tests (p < 0.05, Bonferroni-correctedformultiple comparisons). 

No comments:

Post a Comment