Research,

Current research:

Lexical-semantic network growth in children from different socio-economic backgrounds

GitHub Repo

A mother’s educational background and socio-economic status (SES) have been shown to have a significant influence on the child’s early vocabulary development. For example, children coming from families of higher SES generally have larger vocabulary sizes than children from families of lower SES at the same age during early development. Although vocabulary size is a key indicator of early language development, the vocabularies of children from different backgrounds may also differ in other dimensions, such as vocabulary diversity and structural complexity. However, most current studies on this topic focus on vocabulary size and we have very limited knowledge about how the socio-economic background affects vocabulary development in other dimensions. To fill the research gap, the main goal of this project is to characterize the vocabulary structure in children of different SES by analyzing the lexical-semantic network (a structure representing relations between words using pre-trained language models. Additionally, the lexical-semantic network of maternal child-directed speech (CDS) was also examined to look for correlations between the mother’s and the child’s vocabulary.

The data used in this project contains more than 5.5 million words from 13 corpora in the Child Language Data Exchange System (CHILDES). An efficient pipeline was developed to clean and analyze large language datasets using Python. For quality control, mean length of utterance by morphemes (MLU-m), a common way to measure a child’s linguistic productivity, was used as an indicator of data quality.

This project demonstrates the possibility of studying early vocabulary development with semantic network analysis using a combined source of data. In addition to child language development, the data analysis framework developed in this project can also be used in other studies such as adult language acquisition and bilingualism.

This work is close to completion and is expected to be published in 2023.

Experience-dependent maturation of the reward circuitry

Stresses experienced in early life increase the susceptibility to a wide range of psycho-pathological conditions in adulthood, such as addiction and depression. These conditions are likely driven by impairments in reward-motivated cognitive functions. For example, addiction exhibits compulsive reward-seeking behaviors while depression displays a generalized lack of motivation. These behaviors are regulated by the reward circuitry in the nucleus accumbens (NAc) of the basal forebrain, which integrates inputs from different brain areas, coordinating different cognitive functions. Currently, little is known about how the reward circuitry develops, nor how early life experience modulates the development of the reward circuitry. Besides, it is unclear how early life stress is mechanistically related to the impaired reward-motivated cognition in adulthood. Therefore, this project aims to (1) study the role of early life stress in the development of NAc circuitry; and (2) study the involvement of the change in NAc circuitry due to early life stress in the impaired reward-motivated behaviors in adulthood.

This project uses the mouse brain as a model to study the neural correlates of reward-motivated cognitive functions. Environmental stress is induced in young pups. Cell-type specific electrophysiological recordings are performed to study the effects of early life experience on synaptic connectivity in the reward circuitry in the NAc. Behaviors in adulthood are analyzed to look for links between the synaptic changes in the NAc and the impairments of reward-motivated cognitive functions.

This work is close to completion and is expected to be published in 2023.