Frontiers in Psychology (Apr 2015)
Spatio-temporal dynamics of word selection in speech production: Insights from electrocorticography
Abstract
Background Word selection refers to the act of choosing words as we speak. Different regions of the left lateral and medial prefrontal cortex (PFC), and left temporal cortex (LTC) are associated with word selection (Price, 2012) but their role and how they interact remain underspecified. Methods We recorded electrocorticography (ECoG) in 9 neurosurgical patients to examine where and when subregions of the PFC and the LTC were engaged in word selection (7 left, 2 right hemisphere cases). Patients performed picture naming wherein semantic context was manipulated to affect word selection difficulty: pictures of objects were presented within semantically-homogeneous or heterogeneous blocks and presented 4 times per block. This paradigm is known to elicit the classical semantic interference effect where performance is worse in homogeneous than in heterogeneous blocks from the 2nd presentation onward (Damian et al., 2001). Results Behavioral results: Subjects’ performance was worse in homogeneous vs. heterogeneous blocks, showing the usual semantic interference effect. There was a significant interaction between semantic context and presentation number for both reaction times (Wald Χ2 = 11.96, p = .008) and error rates (Wald Z = -2.00, p = .046). ECoG results: High-gamma (HG, 70 to 150 Hz) amplitude, indexing cortical activation magnitude, was sensitive to semantic interference at several cortical sites (i.e. greater in homogeneous vs. heterogenous blocks, Figure 1). Early effects were seen in HG activity starting before 250 ms post-stimulus at posterior inferior LTC and middle and superior frontal sites. Response-locked effects were seen in medial and left lateral PFC activity peaking before vocal onset and also in posterior superior and middle LTC activity peaking after vocal onset. Finally, across trials, HG amplitude co-varied between responsive left PFC and LTC sites. Single-case: One patient, whose seizure focus was in the posterior medial PFC, had poor performance (error rate > 40%) and his data was therefore analyzed separately. His semantic interference effect (321 ms) was more than 3 standard deviations larger than that of the other patients (mean = 43 ms, SD = 82 ms). Thus, when brain tissue in the posterior medial PFC is abnormal, interference caused by semantically-related alternatives is more difficult to overcome. Conclusions Our results suggest that the posterior inferior LTC is involved in word selection as semantic concepts become available. Posterior medial and left PFC regions may be involved in trial-by-trial top-down control over LTC to help overcome interference caused by semantically-related alternatives in word selection. The single-case result supports this hypothesis and suggests that the posterior medial PFC plays a causal role in resolving this interference in word selection. Lastly, the sensitivity to semantic interference of the post-vocal onset posterior LTC activity suggests the semantic interference effect does not only reflect word selection difficulty but is also present at post-selection stages such as verbal response monitoring. In sum, this study reveals a dynamic network of interacting brain regions that support word selection in language production.
Keywords
