Frontiers in Psychology (Apr 2015)
Neural correlates of successful and unsuccessful syntactic processing in primary progressive aphasia
Abstract
Introduction Syntax is one of the defining features of human language, yet the functional neuroanatomy of syntactic processing is not well understood. Lesion studies have not revealed robust relationships between damage to specific regions and syntactic deficits (Caplan et al., 2007), while functional imaging studies have faced a major challenge in that attempts to manipulate syntactic processing invariably also vary demands on processing resources such as working memory, attention, and executive function (Bahlmann et al., 2007). In this study, we took a novel approach to identifying brain regions involved in syntactic processing: we compared the neural correlates of syntactic processing in patients with primary progressive aphasia whose syntactic processing was either relatively spared or relatively impaired. We reasoned that activity in regions involved in task-related processes would be correlated trial-by-trial with reaction time (a proxy for extent of processing) in all patients, whereas regions involved in syntactic processing would be correlated with reaction time only in patients with spared syntax. Methods Fifty-one patients with primary progressive aphasia were scanned with fMRI as they performed a two-alternative forced-choice syntactic processing task involving auditory sentence-picture matching with sentences varying in syntactic complexity (7 conditions varying in complexity; Siemens 3 Tesla Trio scanner; TR = 2000 ms; 470 volumes; whole brain coverage) (Wilson et al., 2010). In each individual patient, we identified regions modulated by the task by correlating BOLD signal with reaction time, as a proxy for trial-by-trial demands on syntactic and task-related processes (Taylor et al., 2014). We performed group analyses in the worse-performing and better-performing halves of the patient group. To identify significant differences as a function of degree of syntactic impairment, we correlated individual maps with overall accuracy across subjects. Results All patients attempted to perform the task, and showed similar patterns of reaction times (longer for more syntactically complex sentences). However accuracy was highly variable, ranging from 52.4% correct (chance) to 91.7% correct (near-normal). In the worse-performing half of the patients, with relatively impaired syntax, the regions modulated by the task were the anterior insula bilaterally, the supplementary motor cortex bilaterally, and left dorsal premotor cortex. In the better-performing half of the patients, with relatively spared syntax, these same regions were modulated, but additional regions were also modulated: the left inferior frontal junction, left posterior superior temporal sulcus, and left intraparietal sulcus. The correlation with accuracy revealed that modulation of the left inferior frontal, left posterior temporal, and left inferior parietal activations were significantly correlated with overall accuracy on the syntactic task (p < 0.05, corrected for multiple comparisons). Discussion Our findings suggest that some of the regions modulated by a syntactic processing task reflect task-related functions such as working memory, attention, and executive function, specifically the anterior insula bilaterally, the supplementary motor cortex bilaterally, and left dorsal premotor cortex. In contrast, other regions were modulated only in individuals with relatively intact syntactic processing, namely the left inferior frontal junction, left posterior superior temporal sulcus, and left intraparietal sulcus, suggesting that these regions are important for syntactic processing.
Keywords
