Individual difference analyses

Our results showed that inter-individual variation plays a large role in the elicitation of the negativity. There is growing evidence from the ERP literature that brain responses associated with language processing can vary qualitatively across individuals, and even the typical brainwave components like N400 and P600 have been substantially manifested to have qualitative individual differences. In fact, it is possible for ERP effects presented in grand mean waveforms to be interfered by systematic distortion of the signal during averaging process (Tanner et al., 2018).

For example, Nakano at el. (2010) demonstrated that qualitatively different brain

responses were shown between listeners with high or low working memory span when comprehending simple active sentences. They found in the contexts of animacy violations of grammatical sentences (e.g., The box is biting the mailman), N400 was elicited at verb in low working memory span subjects, yet a P600 was elicited in high working memory span ones. Similar report also indicated that unexpected words during sentence comprehension trigger a tradeoff between the two qualitatively distinct brain responses reflected in the N400 and P600 components of the ERP, and this tradeoff was constrained by individual differences in verbal working memory capacity (Kim et al., 2018). Such findings conclude that individual differences may be crucial to language comprehension.

Individual variation in patterns of language-related ERPs has not only been inspected from several aspects but also been unveiled qualitative individual differences from some sorts of behavioral measures. In particular, the variation in syntactic processing is associated with one’s proficiency in grammar and vocabulary, whereas the processing of conflicting sentence was related to individual differences in cognitive control. Still more studies asserted that individuals’ cognitive styles could delineate ones’

social information processing. With respect to the disambiguation, Nieuwland and Van Berkum (2008) conducted event-related brain potential (ERP) study to examine the interplay between semantic and referential aspects of anaphoric noun phrase resolution in Dutch stories, indicating that some participants elicited an unexpected LPC to three types of problematic anaphors (ambiguity, incoherence, and ambiguity/incoherence) while others did not. Their result suggested that large individual differences exist and people differ in comprehension processes. Lee and Federmeier (2012) examined how aging affect the processing of ambiguity resolution by comparing ERP responses to homographs and unambiguous words in context which had only syntactic but semantic- neutrally information (e.g. plausible sentence: “Ben tried the duck in the dish prepared by a famous

chef.” vs. implausible sentence: “Ben tried to duck in the dish prepared by a famous chef.”) The result revealed that a frontal negativity showed not only in the young adults but also those older adults who had high verbal fluency, indicating the older adults who have better cognitive abilities can display more young-like ERP patterns within the same task.

Depending on the studies, it is therefore reasonable to infer the absence of frontal negativity observed in our data might be modulated by individual differences in participants’ scores on neuropsychological tests since the individual waveforms indeed showed a half negative-going and other half positive-going ERPs at frontal sites. In order to examine the effects of individual differences within the target epoch, we looked into the ERPs based on the following individual difference measures.

Beginning with the inspection of verbal fluency, which is assessed in letter and category fluency tasks, the performance on these tasks is related to indicators of vocabulary size, lexical access speed, updating, and inhibition ability. It is well-accepted that individuals vary in their verbal fluency performance, and that this variability can have implications for language processing (Just & Carpenter, 1992; Cohen et al., 1999;

Weckerly et al., 2001; Federmeier et al., 2010; Henry & Crawford, 2004; Fitzpatrick et al., 2013). Past work asserted that the ambiguity effect patterns were more likely to correlate with the performance in the fluency tasks (Lee & Federmeier, 2011). Evidence from clinical studies has also demonstrated that damage to frontal brain areas is associated with poor performance in the fluency tasks (Schwartz & Baldo, 2001). Surprisingly, our study yielded different ERP results which conflicted with the behavioral findings. To illustrate, the frontal negativity was nearly no significance in the high score group of verbal fluency, whereas the negativity was far clear to observe in the high score group of another index— reading experience. According to cognitive neuroscience literature on reading experience, individuals’ overall meaning preferences can reflect their reading

experience across a wide range of timescales even from minutes to years (Rodd et al., 2016). Some research also suggest that lexical-semantic representations are not fixed as previously thought, but very malleable and dynamic. More specifically, those who are equipped with highly reading skills not only being good at learning new meanings for previously unambiguous words, but being adept at updating their representations of word meanings based on their current linguistic experience (Rodd et al., 2013). Such ability, likewise, manifests on the ambiguous words. Using a word-meaning priming paradigm, Rodd et al. (2016) conducted four experiments to demonstrate how previous and recent reading experience influence word meaning accessibility and found that the interpretation of ambiguous words was influenced by experience that recently encountered meanings become more readily available. These works, to certain extent, may provide an explanation that the effect only prominent in the high score group of reading experience.

By comparison, skilled language comprehension depends on the ability to disambiguate the precise meaning of individual words to build an accurate representation of the intended message. Once one has more chances to disambiguate information, it therefore qualifies themselves for better reading ability. Those who are good at reading might also show a greater sensitivity to the limited information context when processing ambiguous words, and might be more likely to possess advantage in the mechanism of meaning selection. The view proposed by Gernsbacher (1993) that less skilled readers are less effective in meaning suppression and tend to maintain activation of contextually-irrelevant meanings for a long may support the result as well. However, this result does not mean that verbal fluency task is not a good indicator. Prior study has pointed out that better verbal fluency performance is associated with greater amount of frontal negativity elicited by the older adults (Lee & Federmeier, 2011). Since our participants were all young adults, perhaps, the overall high verbal fluency scores across all participants was

relatively unable to show a notable difference under this indicator. We can merely speculate that the present group differences are not reflected on the indicators as previously. Perhaps, only if the score of verbal fluency is below a certain threshold, it will then exert an impact and interference on the ambiguity resolution.