Interpretation-mediated changes in neural activity during language comprehension.
Abstract:
Using functional magnetic resonance imaging (fMRI), we
identified cortical regions mediating interpretive processes
that take place during language comprehension. We
manipulated participants’ interpretation of texts by asking
them to focus on action-, space-, or time-related features
while listening to identical short stories. We identify
several cortical regions where activity varied significantly
in response to this attention manipulation, even though the
content being processed was exactly the same. Activity in
the posterior and anterior sections of the left inferior
frontal gyrus (IFG), which are thought to have different
sensitivities to high-level language processing, was
modulated by the listeners’ attentional focus, but in ways
that were quite different. The posterior left IFG (Pars
Opercularis) showed different activity levels for the three
conditions. However, a population coding analysis
demonstrated similar distributions of activity across
conditions. This suggests that while the gain of the
response in the Pars Opercularis was modulated, its core
organization was relatively invariant across the
experimental conditions. In the anterior left IFG (Pars
Triangularis), the analysis of population codes revealed
different activity patterns between conditions: there was
little similarity between activity during time-attention and
action- and space-attention, however there were similar
activity patterns while attending to spatial and action
information. In addition, both the left superior temporal
gyrus and sulcus showed greater activity in the space and
action attention conditions when contrasted with time
attention. We discuss these findings in light of work on the
role of left IFG in processing semantic information in
language, and in light of theories suggesting that temporal
information in language is processed in the brain using
similar mechanisms as spatial information. Our findings
suggest that a substantial source of variance in neural
activity during language comprehension emerges from the
internally-driven, information-seeking preferences of
listeners rather than the syntactic or semantic properties
of a text.