Seminars

Speaker Series: Individual Differences in Categorical Perception: Exploring
Links to Executive Functions and Language Experience

Categorical perception of speech (CPS) refers to the processing mechanism by which listeners perceive continuous speech signals as discrete phonetic categories, showing greater sensitivity to differences between categories than within categories. Previous studies have shown that more robust CPS—characterized by reduced sensitivity to within-category variation—reflects more established phonetic representations in both first and second language learners, as well as in individuals with dyslexia. 

More recently, however, research has reported a range of individual differences in how discretely listeners categorize speech sounds, with some individuals showing gradient perception by more effectively assessing redundant acoustic cues. 

This talk explores why some listeners exhibit gradient rather than categorical perception, and what functional advantages this perceptual style might offer. In particular, it focuses on how individual differences in executive functions—such as working memory, inhibitory control, and cognitive flexibility—as well as factors like age (children vs. adults) and language experience (first vs. second language), may be linked to the ability to make use of fine-grained acoustic information when making phonetic category decisions.

December 5, 2025

Speaker Series: Converging multi-modal evidences for the primary prevention of Alzheimer's dementia

I direct the Genome Intelligence Mining Lab, where we focus on uncovering the hidden codes and mechanisms embedded in human genetic information. Our overarching goal is to transform these insights into strategies for understanding complex diseases and improving human health.

In this presentation, I will highlight our efforts to develop predictive and preventive approaches for Alzheimer’s disease (AD). Specifically, we have pursued two complementary directions: (1) building flexible risk-prediction models that integrate multi-modal data, and (2) evaluating the strengths and limitations of each modality to optimize their effective use. We established a genetic risk prediction framework that incorporates common genetic variants (single nucleotide polymorphisms/variants) and identifies AD-associated loci. This genetic risk is dynamically updated by integrating additional evidence from other modalities—such as neuroimaging, plasma biomarkers, and cognitive measures—when available.

Validation using data from the Gwangju Alzheimer’s & Related Dementias (GARD) study showed that this multi-modal framework substantially improves prediction of AD, as measured by amyloid positivity and early cognitive decline, compared to single-modality strategies. These results illustrate how mining genetic information and combining it with multi-modal data can open new avenues for primary and secondary prevention of late-onset Alzheimer’s disease (LOAD).

November 18, 2025

Speaker Series: Speech Abnormality Patterns Throughout the Life Span

Speech production is a highly planned activity involving the complex interplay of multiple brain regions. For this reason, speech can serve as a powerful tool for the early screening and monitoring of brain and cognitive-related disorders.

This lecture will introduce unique speech and language signatures observed across various clinical conditions, including Autism Spectrum Disorder, psychosis (e.g., Schizophrenia), and neurodegenerative disorders (e.g., Alzheimer’s disease). This analysis utilizes scalable analytical tools capable of automatically assessing cognitive and social functions.

The research findings suggest that automated speech and language metrics provide non-invasive, objective, and sensitive measures that can be beneficially employed for the screening and monitoring of individuals with diverse clinical conditions.

October 17, 2025

Language Acquisition through the Lens of Data : Meta-analysis 

The fifth tutorial topic, Meta-analysis, is a research method that yields new information by analyzing the results of existing studies as data. It can also be applied to planning sample sizes through power analysis when designing an experiment.

Dr. Katie Von Holzen from the Technical University of Braunschweig, Germany, will first deliver a lecture on research that applies the meta-analysis methodology, followed by a hands-on session where participants can follow along. We hope this will be an excellent learning opportunity.

• Part 1: Talk (Meta-analysis and mispronunciations)

• Part 2: Hands-on follow-along (Practical session)

March 15, 2024 

Language Acquisition through the Lens of Data : WordSeg

The fourth tutorial topic, WordSeg, addresses the fundamental question of how infants first learn words by locating word boundaries within the continuous stream of speech input, much like finding structure within flowing water.

The session compares the performance of various algorithms in segmenting words from corpus text input. It also shares a deep analysis of whether word segmentation performance is superior when the input data is Child-Directed Speech (CDS) compared to Adult-Directed Speech (ADS), and the reasons behind any observed superiority.

• PART I: How caregivers' speech patterns enhance word segmentation: A computational modeling approach with a Korean corpus

  • Speaker : Jun Ho Chai, Research Professor, Institute for Data Science in Humanities, Chosun University

• PART II: A tutorial on Grapheme-to-Phoneme conversion & WordSeg package

  • Speaker : Seongmin Mun, Research Professor, Institute for the Humanities, Ajou University

December 15, 2023 

Language Acquisition through the Lens of Data : ELAN

The third tutorial topic, ELAN, is a program used as a tool for annotating and analyzing audio and video materials. Researchers can add annotations for various linguistic and non-linguistic elements to videos of infant speech or interaction, and perform detailed analysis by synchronizing these with the media file's timeline.

The session will feature a research presentation by Dr. Rana Abu-Zhaya from University College London, followed by a tutorial on how to use ELAN, presented in Korean by SuHan Kim from Chosun University.

• PART I: Unpacking the multimodal nature of early language environments: Evidence from naturalistic interactions with infants

• • • • Speaker: Dr. Rana Abu-Zhaya, Lecturer, University College London (Website)

  • PART II: ELAN Tutorial

    • • Speaker: Suhan Kim (Master's Candidate), Chosun University

      • Software Information: ELAN Website

June 9, 2023

Language Acquisition through the Lens of Data : CHILDES

The second tutorial focuses on CHILDES, a database featuring transcribed dialogues of infants and young children. Researchers can extract desired information using web commands or the CLAN software, or utilize the childes-db package in R/Python for more tailored analysis according to individual research goals.

The session will begin with a research presentation by Dr. Stephan Meylan from MIT, followed by a tutorial on how to use CHILDES, presented in Korean by Jinyoung Jo (PhD Candidate) from UCLA.

• PART I: Child-directed listening: How adults understand what young children say

  • Speaker : Dr. Stephan Meylan, MIT  (Website)

• PART II: A tutorial on how to use CHILDES (CHILDES Link)

  • Speaker: Jinyoung Jo (PhD Candidate), UCLA (Website)

May 19, 2023

Language Acquisition through the Lens of Data : WordBank

In this lecture, Professor Michael C. Frank of Stanford University will introduce large-scale data-driven approaches to the study of child language development worldwide.

While all children learn language, there are significant individual differences in the pace and manner of acquisition. Professor Frank has established large-scale, public databases, such as Wordbank and Peekbank, to predict this variation and to identify developmental patterns that are common across different languages and cultures.

The presentation will share the latest research findings on understanding the variability and universality of child language learning, focusing on predictive modeling research built upon these massive data resources.

April 12, 2023