How do we extract phonemes, grammatical rules, and meaning from strings of words? What information is crucial for interpreting language?

Korean, as an agglutinative and morphologically rich language with syllable-timed prosody, offers unique insights into lexical processing. Previous research has demonstrated that the initial syllable of Korean words serves as a critical index for retrieving words from the mental lexicon.

Through a combination of behavioral experiments and EEG neuroimaging, I have investigated this phenomenon and discovered robust facilitative effects of syllable frequency on word recognition. The EEG data revealed that sub-lexical processing influences span from early perceptual stages through to later lexical access stages, providing a comprehensive view of how Korean speakers process linguistic information.

Currently, I am examining priming effects across different syllable types—specifically contrasting phonological versus orthographic representations—using EEG and computational network analysis. This research aims to determine whether established neural network models can accurately replicate the observed patterns in human neural data, bridging the gap between cognitive neuroscience and computational linguistics.

Can cognitive aging be prevented through neuromodulation?

As Korea rapidly transitions into an aging society, preventing cognitive decline has become one of the most pressing challenges of our time. With increasing life expectancy, understanding how to maintain cognitive vitality in older populations represents both a scientific frontier and a societal imperative.

My research focuses on enhancing various memory types—including working memory, short-term memory, and long-term memory—through targeted non-invasive brain stimulation using transcranial alternating current stimulation (tACS). By stimulating neural oscillations in the dorsolateral prefrontal cortex (dlPFC), a critical hub for executive function, I investigate how this stimulation can enhance cognitive performance across diverse memory tasks.

The aim of the study is to establish evidence-based protocols for cognitive enhancement that could ultimately benefit aging populations.

How can we resolve ambiguity in language processing?

The world is full of ambiguity, and we encounter it frequently in language use. Korean, with its rich morphological system, provides an ideal testing ground for understanding ambiguity resolution mechanisms. For example, the morphologically complex word "삶은" presents a classic ambiguity: it can be interpreted either as the noun "삶" (life) combined with the topic marker "은," or as the past participle form of the verb "삶다" (to boil).

Two competing theoretical frameworks offer different explanations for how such ambiguities are resolved. The affix stripping model (Taft & Forster, 1975) proposes that morphologically complex words are decomposed into their constituent parts during processing, with each component accessed separately from the mental lexicon. In contrast, the exposure-based immediate attachment mechanism (Nam et al., 2002) suggests that processing depends on the frequency of exposure to specific word forms, leading to direct lexical access without decomposition.

To determine which mechanism better accounts for Korean morphological processing, I am conducting an EEG experiment that contrasts root-and-affix-related conditions with root-only-related conditions. This design allows me to examine the neural signatures of morphological decomposition versus whole-word access.

The predicted outcomes are theoretically distinct: If the affix stripping model is correct, "삶은" should be decomposed into "삶" + "은," with separate neural activation for each component, resulting in comparable ambiguity effects across both experimental conditions. However, if the exposure-based mechanism is more appropriate, we should observe an interaction between word frequency and ambiguity resolution, with high-frequency ambiguous forms showing different processing patterns than low-frequency ones.