Functional specificity for high-level linguistic processing in the human brain

Abstract

Neuroscientists have debated for centuries whether some regions of the human brain are selectively engaged in specific high-level mental functions or whether, instead, cognition is implemented in multifunctional brain regions. For the critical case of language, conflicting answers arise from the neuropsychological literature, which features striking dissociations between deficits in linguistic and nonlinguistic abilities, vs. the neuroimaging literature, which has argued for overlap between activations for linguistic and nonlinguistic processes, including arithmetic, domain general abilities like cognitive control, and music. Here, we use functional MRI to define classic language regions functionally in each subject individually and then examine the response of these regions to the nonlinguistic functions most commonly argued to engage these regions: arithmetic, working memory, cognitive control, and music. We find little or no response in language regions to these nonlinguistic functions. These data support a clear distinction between language and other cognitive processes, resolving the prior conflict between the neuropsychological and neuroimaging literatures.

Fig. 1.

Sample trials for the localizer task and nonlinguistic tasks in experiments 1–6. In the math task, participants added smaller vs. larger addends. In the spatial/verbal WM tasks, participants remembered four vs. eight locations/digit names. In the multisource interference task (MSIT), participants saw triplets of digits and had to press a button corresponding to the identity of the nonrepeated digit. The hard condition involves two kinds of conflict: between the spatial position of the target response and the position of the response button and between the target response and the distracter elements. The verbal MSIT (vMSIT) task was a slight variant of the MSIT, where words were used instead of digits. In the Stroop task, participants overtly named the font color vs. noncolor adjectives. In the music task, participants listened to unfamiliar Western tonal pieces vs. the scrambled versions of these pieces (Materials and Methods has details).

Fig. 2.

Responses of the left hemisphere language ROIs to the language task (black, sentences; gray, nonwords; estimated using data from left-out runs) and the nonlinguistic tasks [shown in different colors, with the solid bar showing the response to the hard (experiments 1–6) /intact music (experiment 7) condition and the nonfilled bar showing the response to the easy/scrambled music condition]. Individual subjects’ ROIs were defined by intersecting the functional parcels (shown in blue contours) with each subject's thresholded (P < 0.001, uncorrected) activation map for the language localizer. The functional parcels were generated a priori based on a group-level representation of localizer activations in an independent set of subjects (29). Asterisks indicate significant effects (Bonferroni-corrected for the number of regions). The sentences > nonwords effect was highly significant (P < 0.0001) in each functional ROI. The only significant effect for the nonlinguistic tasks was the hard > easy verbal WM effect in the LMFG functional ROI (P < 0.005). Tables S1 and S2 have detailed statistics.

Fig. 3.

Sample individual activation maps (thresholded at P < 0.001, uncorrected) for the nonlinguistic tasks showing three representative subjects from each experiment. For experiments 1–6, the contrast is hard > easy; for experiment 7, the contrast is intact > scrambled.

Fig. 4.

Key results from the whole-brain overlap analysis with very liberally thresholded (P < 0.05, uncorrected) individual activation maps (Materials and Methods has details). Similar overlap regions—in the posterior part of the LIFG language ROI (whose outline is shown in gray above)—emerged in the analyses of language and verbal WM (green; present in 85% of the subjects with an average size of 76 voxels) and language and Stroop (orange; present in 71% of the subjects with an average size of 95 voxels). The left set of bars in each panel shows the responses of these overlap regions to the four conditions (estimated in left-out runs). The right set of bars in each panel shows the results of a complementary analysis where we examined voxels within the LIFG language ROI that respond to the language localizer contrast (at P < 0.001, uncorrected) but not to the nonlinguistic task contrast (at P < 0.1). Responses to the four conditions were again estimated using left-out runs. In each dataset, a substantial proportion of voxels (127/355 voxels, on average, in the language-Verbal Working Memory dataset and 163/410 voxels in the language-Stroop dataset) showed replicable selectivity for the language task, suggesting that a portion of the LIFG language ROI is selectively engaged in sentence understanding.