2.1. Participants

Imaging data were downloaded from the OpenNeuro database accessed on 3 April 2022 (https://openneuro.org/datasets/ds001747). This dataset was chosen because of its well-designed data collection, e.g., categorization of three groups based on the L2 acquisition age, and strictly controlled for age, sex, and education (over 80% participants were full-time college students), which can minimize the influences of these factors on statistical analysis. A total of 102 participants aged between 18 and 29 years and had specific language background were recruited at Brigham Young University [17]. Recruitment focused on three types of populations, namely early bilinguals (EBs), late bilinguals (LBs), and monolingual controls (MCs). Individuals were invited to participate based on their survey responses, specifically whether they would fit into one of the three research groups. The EB group was composed of individuals proficient in English and Spanish with the experience of both languages that began before the age of ten. The LB group consisted of individuals proficient in English and Spanish and who started to learn the L2 (Spanish) after the age of fourteen. The MC group included individuals who only had English experience or had experiences with other languages (in addition to English) that were acquired at any age but was self-rated as “Novice” level [17]. The data exclusion criteria were as follows: left-handed, color-deficient or color-blind, suffered a traumatic brain injury, diagnosed with a psychiatric or neurological disorder, and pregnancy. All participants met the criteria for participating in the MRI scans. In addition, participants with evident head movement during scan were also excluded (see Section 2.3 for details). The demographic information is shown in Table 1. The participants’ English proficiency and Spanish proficiency tests were designed as follows: the study design called for all participants to complete an English proficiency test and any participant with experience with the Spanish language (from taking a semester in junior high to growing up with the language) to take a Spanish proficiency test. The tests were provided by Emmersion Learning, which administered the exams via an online platform, with participants completing them from a home computer. Both proficiency tests used elicited imitation (EI), in which the test taker hears an utterance in the target language and is prompted to repeat the utterance as accurately as possible. If the participant is completely unfamiliar with the language of the utterance they hear, each syllable of the utterance will count toward that limited capacity, reducing their ability to accurately repeat the utterance to only a few syllables. In this way, EI can reliably approximate a learner’s proficiency level by measuring the accuracy of the repetition of utterances of increasing length and complexity [17].

Demographic and clinical information of participants.

Abbreviations: EB, early bilingual; LB, late bilingual; n.s., non-significant. Note: values are mean ± SD. ^a ANOVA. ^b Chi-square test.