Presentation

Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams

Assessing the role of implicit vs. explicit processes in L2 grammar task performance via reaction time-to-ERP correlations

Poster D1 in Poster Session D with Social Hour, Friday, October 7, 5:30 - 7:15 pm EDT, Millennium Hall

David Abugaber¹, Kara Morgan-Short²; ¹University of Michigan, ²University of Illinois - Chicago

Adult second language learning is a complex task involving both explicit and implicit processes (i.e., processes that do/do not involve conscious awareness). Studies on this topic using reaction time (RT) methods can measure overall performance but cannot ascertain whether RTs capture the same vs. different underlying processes between implicit and explicit learning (e.g., Leung & Williams, 2012). In turn, prior EEG research suggests separate event-related potentials (ERPs) associated with implicit and explicit processes (e.g., Batterink et al., 2014), but have not determined whether neural markers of explicit processing are directly tied to production of external responses or simply an epiphenomenon of conscious noticing that is only indirectly tied to task performance. Our study addresses this latter question by extending an established artificial language paradigm (Batterink et al., 2014) with an analysis of the relationship between RTs and ERP timing. English native speakers (N=48) performed a semi-artificial language task that covertly tests for learning of a hidden grammatical regularity by comparing responses to rule-adhering vs. rule-violating trials. Afterwards, participants’ awareness of the hidden regularity was assessed via a debriefing interview; 24 participants were coded as rule-aware and 24 as rule-unaware. Slower RTs and lower accuracies for rule-violating trials suggested grammar learning in both rule-aware and rule-unaware participants with no differences between the two, replicating prior findings using this paradigm (Batterink et al., 2014). We found a positive ERP in a late (800-1100ms) time window for rule-unaware learners but no significant ERPs for rule-aware learners (possibly due to trial-to-trial variability in conscious processes). We assessed the relationship between the ERP in rule-unaware learners and their RTs via four parallel analyses as in Sassenhagen et al. (2014). First, inspection of ERP images visualizing per-trial ERP latencies and RTs (Jung et al., 2001) suggested a relationship between the two. Secondly, per-participant RT quartile rank was significantly related to mean ERP latency per quartile (Marathe et al., 2013), F(3,48) = 21.81, p < .001, η2G = .53. Thirdly, we found a non-significantly positive mean per-participant correlation between individual trial RTs and ERP latencies as estimated via Woody filtering (Kutas et al. 1977), mean r = 0.10 (SD = 0.26, 95% CI = [-.03, 0.23]). Fourthly, we found non-significantly greater inter-trial phase coherence (Delorme et al., 2007) in response time-aligned (vs. stimulus-aligned) data, t(23) = 1.83, p = .080. In sum, our first and second analyses showed evidence for time-locking between RTs and ERPs. Our third and fourth analyses trended in the same direction but were not significant, which we attribute to the weakness of the observed ERPs in this artificial learning paradigm. Because we did not find a significant ERP for rule-aware learners, we were unable to investigate the relationship between conscious processes and task performance. We consider these results in light of previous findings that have tied late positive ERP components to explicit processing or knowledge (e.g., Batterink et al., 2014; Morgan-Short et al., 2022) and consider whether implicit processing is reflected in these late components.

Topic Areas: Multilingualism, Syntax