Research Internship Report

(1)

Research Internship Report

Bilingualism, Mind & Brain Lab

University of California Riverside

Floor van den Berg S---

Research Master Language & Cognition, University of Groningen

Internal supervisor: Dr. Merel Keijzer External supervisor: Prof. Judith Kroll

(2)

1

1. Introduction

As part of my Research Master in Language & Cognition at the University of Groningen I was able to do an internship in Prof. Judith Kroll’s Bilingualism, Mind & Brain (BMB) lab at the University of California Riverside (UCR). This research internship is a compulsory component of the program, and its purpose is for students to gain research experience abroad. During these internships, students learn valuable research skills, which may be applied to future research. Importantly, the internship stimulates collaboration between research institutions abroad and the University of Groningen.

Ever since I started studying English Language & Culture in Groningen I have been interested in the psycholinguistic aspects of bilingualism and second language learning, so an internship in a laboratory focused on researching the bilingual experience seemed a logical choice. In addition, I wanted to familiarize myself more with neurocognitive methods such as recording electroencephalograms (EEG). I inquired with Dr. Merel Keijzer, Associate Professor of Applied Linguistics at the University of Groningen and my bachelor thesis supervisor, whether she knew any great internship options for bilingualism research. In March of 2018 she reached out to Prof. Kroll and Dr. Eleonora Rossi on my behalf. Prof. Kroll is the Principal Investigator of the BMB lab at UCR, and this lab has Dr. Rossi as one of its research associates. Dr. Keijzer had previously supervised another Research Master student, Isabel Eyer, who had completed an internship in Prof. Kroll’s lab the previous year under the supervision of Dr. Rossi. During her time there she worked on the Rosetta Stone project, which investigates the consequences of short-term language training. After several e-mail and Skype conversations, Prof. Kroll, Dr. Rossi, Dr. Keijzer and I decided that I was going to continue the work on the training study with EEG, which has been the main focus during my internship.

The internship took place from September 3 to December 21, 2018. It included an extension of 3 ECTS in the form of a tutorial on how to pre-process and analyze EEG data with BrainVision Analyzer. In total, I spent approximately 644 hours working in the lab, based on a full-time appointment for a duration of 16 weeks.

2. The Research Lab

The BMB lab has been at the University of California Riverside since the Summer of 2016. UCR is known for its diversity, both in the programs they offer and regarding its students and staff. A large portion of the population in California is comprised of bilinguals and heritage speakers, which makes UCR a great location for bilingualism research.

At the moment of writing this report, the lab consists of Principal Investigator Prof. Judith Kroll, three post-doctoral fellows (Natsuki Atagi, Eve Higby, and Megan Zirnstein), three graduate students (Christian Navarro-Torres, Emily Mech, and Andrea Takahesu-Tabori), several research associates (including Dr. Eleonora Rossi from the University of Florida), lab manager Dalia Garcia, and many undergraduate research assistants. This lab is part of the Department of Psychology and is associated with a number of other research labs both at UCR and across the United States. Furthermore, the lab collaborates with a number of research institutions in Europe,

(4)

3

Asia, Latin-America, and the US on the NSF-funded PIRE (Partnerships for International Research) project. The aim of this project is to develop an international network of bilingualism research with respect to language use and learning context by providing training opportunities to undergraduate and graduate students. Essentially, these students get the opportunity to conduct their own research abroad at one of the partner institutions.

The main themes that are reflected in the lab’s research all relate to the interplay between several aspects of bilingual language use and their cognitive and neural mechanisms. My

background in psychology and cognitive neuroscience is not strong, but my interests within the realm of bilingualism are broad. As a result, I was excited to be able to work in this laboratory. Prior to this internship I had read about the controversy surrounding advantages in executive functioning as a result of bilingualism, and I was intrigued by the concept. Thus, I was eager to learn more about it for the training study. Not only did I find the combination of language learning and broader theoretical concepts such as the bilingual advantage appealing, I also liked that this project has practical applications as well, not to mention the fact that it is a collaborative project taking place at two linguistically distinct locations.

3. Research Project – Rosetta Stone

During my internship I fully devoted my attention to the Rosetta Stone project. This project is a collaboration between Prof. Judith Kroll, Dr. Eleonora Rossi, and Prof. Christine Chiarello, as well as PIRE partner universities in the Netherlands (i.e., Radboud University Nijmegen and University of Groningen).

Research on bilingualism in the past two decades has generated the robust finding that the two languages in the bilingual mind are always activated in parallel, even when the use of only one language is required (for a review, see Kroll, Bobb & Hoshino, 2014). It is speculated that the regulation of two (or more) languages in the language system draws on language-independent cognitive and neural resources, and consequently shapes domain-general cognitive networks (Grant, Dennis, & Li, 2014). For that reason, it is believed that knowing multiple languages, and thus constantly needing to negotiate the activation of these languages, leads to certain advantages in executive functioning, including inhibitory, monitoring, and switching skills (Bak, 2016). However, relatively little is known about how and when these cognitive benefits arise, and, more specifically, how much language learning is required for benefits to emerge. This is where the Rosetta Stone project comes in. This project employs a mini-longitudinal language learning paradigm which examines how linguistic, cognitive, and neural signatures may change as a function of learning depending on the language (learning) context, and how cognition and language abilities prior to learning may predict language learning outcomes.

In the early stages of this project, native English students were tested on their language skills and cognitive abilities before and after ten days of learning a foreign language (in the case of this project, Dutch). The project was continued in Groningen by in addition employing an

event-related potential (ERP) paradigm to investigate the neural signatures of auditory lexical

(5)

4

Southern California, where English is the dominant language and where exposure to Dutch is negligible, to native English university students living in the Netherlands, who are immersed in Dutch but have no (prior) knowledge of it. Previous research suggests that immersion may result in dynamic changes in both the L1 and L2, creating conditions that give rise to enhanced

language regulation skills, which in turn may benefit novel language learning (e.g., Linck, Kroll, & Sunderman, 2009). However, the EEG component had yet to be carried out in the US in order to be able to compare language learning in immersed and non-immersed participants. Thus, I continued with the current design at UCR with the goal to compare neural responses to spoken Dutch words in immersed versus non-immersed participants both before and after receiving Dutch language training.

3.1 Theoretical background

During the past two decades or so, evidence has emerged that second language learning may take place quickly even in young adults with a late age of L2 acquisition. Although behavioral effects may not always be observed in the short term, consequences of language learning in young adults may be revealed by differences in online language processing before and after L2 exposure (e.g. McLaughlin, Osterhout & Kim, 2004; Osterhout et al., 2006).

An interesting point of discussion is whether changes related to cognition and language control can also be noticed in the early stages of the L2 learning process. Although controversy remains on the question whether the juggling of two or more languages in the mind leads to enhanced executive functioning (e.g., Paap & Greenberg, 2013; see Baum & Titone, 2014, for a critical discussion), there is an increasing body of literature providing evidence for changes in cognition and the brain as a function of bilingualism and second language learning (e.g., Mårtensson et al., 2012; Rossi, Cheng, Kroll, Diaz, & Newman, 2017). The main cohort of studies investigating the cognitive consequences of bilingualism has focused on already proficient second language learners or lifetime bilinguals. However, as mentioned before, the time course of these changes and the factors that predict them are not well-understood. Can we already see evidence of language learning after a few hours of exposure, and does this lead to an improvement of domain-general executive functioning? In addition, which individual and environmental factors may contribute to developing these benefits?

Perhaps the best approach to tackle these questions is by means of longitudinal studies. To the best of my knowledge, there are currently three published studies targeting the early stages of actual second language learning (going beyond basic vocabulary learning) and its neural or cognitive consequences by employing a longitudinal design. Mårtensson et al. (2012) investigated whether brief but intense second language learning resulted in changes in brain anatomy in language-related areas. Differences in cortical thickness in 14 subjects were measured before and after receiving a 3-month interpreter training. They found increased cortical thickness in several regions in the left hemisphere associated with foreign language acquisition and

increased gray matter volume in the hippocampus after language learning.

Sullivan, Janus, Moreno, Astheimer, and Bialystok (2014) studied neural changes in executive control tasks as a result of a 6-month second language training in Spanish. They

(6)

5

compared ERP responses of students enrolled in an introductory Spanish course with those of students attending an introductory Psychology course. Both before and after learning, all subjects performed in a non-verbal go/no-go task and a sentence judgment task while their EEG was recorded, and in addition completed a verbal fluency task. Although the two groups did not differ on behavioral measures for the ERP tasks, participants in the language learning group elicited a larger P3 amplitude on the go/no-go task than the control group after learning, signaling

enhanced attentional skills for the language learners. Furthermore, the group learning Spanish showed a smaller P600 amplitude for items in the sentence judgement task as compared to the control group, suggesting less effort in syntactic processing for the former group.

Lastly, Bak, Vega-Mendoza, and Sorace (2016) conducted a training study of which the design and results were crucial for the current project. They investigated changes in attention as a result of approximately 14 hours of learning in Scottish Gaelic learners and active and passive control groups. Participants completed three measures of attentional control before and after engaging in a course (in case of the foreign language learners and the active control group). Comparing behavioral data for both time points revealed that the language learning group significantly improved in attention switching as compared to the passive control group. Yet, the active control group was also better on this measure than the passive control group, but to a lesser extent than the language learning group.

In summary, there is some promising evidence that differences in language processing and cognition may be visible after short-term language training. In addition, it was suggested by preliminary data from the Rosetta Stone project that individual differences in executive

functioning and language abilities at baseline may contribute to second language development (Eyer, 2018). Importantly, our project employs a similar design to Bak et al.’s (2016) study, but targets online language learning for even less than 14 hours. The project aims to contribute to previous findings by examining the role of language learning context in addition to consequences for language and cognitive skills. It is expected that language learning networks may be tuned by immersion context, resulting in different behavioral and neural outcomes across immersed and non-immersed learners.

3.2 Methods and design

As stated before, the project employs a mini-longitudinal design with a pre- and post-test and ten days of language learning through an online language learning software (Rosetta Stone) in between. The pre- and post-tests in this subproject included behavioral cognitive and language tasks, and an EEG session, where resting-state EEG and brain waves in response to spoken Dutch words were recorded. The task battery in the pre- and post-test was mostly identical. The pre- and post-tests took approximately 2-2.5 hours to complete, including capping.

One of the research questions of the project targets the role of immersion to examine the impact of previous language exposure on language learning outcomes. Thus, a critical element in this study’s design is the location at which the experiments were conducted. For the sake of comprehensibility, this section will primarily discuss the portion of the study conducted at UCR during the Fall of 2018.

(7)

6

3.2.1 Participants

The experiment at UCR involved eight experimental participants (two males), who all took part in learning Dutch through Rosetta Stone. Previous (behavioral) data had been collected for both experimental and control groups at UCR, the University of Groningen, and Radboud University. Control groups participated in a pre- and post-test, but did not engage in the Dutch learning sessions. Details of previous data collection for this project can be read in Eyer (2018).

Additional data for both experimental and control subjects will be collected during the Winter and Spring of 2019.

The participants at UCR were between 18-25 years old (M = 19.8, SD = 2.3), were right-handed, had normal or corrected-to-normal vision, and had no history of neurological disorders or reading deficiencies. In addition, they were required to have no previous knowledge of Dutch or German. Importantly, all participants at both locations were native speakers of English. However, the participants’ language background varied: some were English monolinguals, and others spoke two or even more than two languages. This enabled us to potentially investigate the impact of previous language experience on new language learning.

Participants engaged in 12 largely consecutive sessions, consisting of the pre- and post-test and ten language training sessions with a maximum gap of two days in between the sessions (see Figure 1). Participants were recruited by means of flyers and via e-mail. Written informed consent was obtained for each participant during each session. Participants were reimbursed for their time ($10/hour in cash).

Figure 1. A schematic of the experimental procedure at UCR with non-immersed participants

3.2.2 Language measures

Language History Questionnaire. All subjects completed a language history questionnaire during

the pre-test, which identified how participants use their language(s), ages of acquisition for each of these languages, their socio-economic background, and any other personal characteristics such as gender and age.

Verbal Fluency Task. The second language task in the battery was the verbal fluency task. This

task is used to assess language production abilities, and its outcome reflects language fluency in a given language. The purpose of this task was to measure if fluency in the participant’s languages

(8)

7

changed as a result of foreign language learning, and if fluency in the first language at baseline predicted language learning outcomes.

In this task, participants were presented with a semantic category on a screen (for example, vegetables or animals). They had 30 seconds to name as many words within that category that they could think of in that language. All participants started with a verbal fluency task in English (their L1/dominant language), and subsequently completed the task in other known languages as well. No more than three languages were tested. Categories were

counterbalanced across even and odd participants and across languages. Auditory responses were recorded, and coded for number of correct items named.

In the post-test, participants at UCR completed an additional verbal fluency task, in which they were required to name as many Dutch words (nouns, adjectives, verbs, and short phrases) as they could think of in one minute, without being restricted to any semantic category. For this task, auditory responses were coded for the total number of attempts, accuracy, the number of items named correctly, and the number of cognates and non-cognates named.

Dutch Picture Naming Task. Participants at UCR completed a Dutch picture naming task in both

the pre- and post-test to compare the outcome with baseline performance on Dutch vocabulary knowledge. In this task, participants were exposed to 102 pictures included in the Rosetta Stone learning curriculum. The pictures depicted nouns that were (about to be) learned during the Dutch language training. Participants were instructed to try to name each picture out loud to the best of their ability. Because this task was also administered in the pre-test, even before Dutch learning had taken place, they were told that they could try to guess or remain silent if they did not know the word in Dutch. Reaction times (RTs) were recorded by a microphone connected to a response box. Auditory responses were collected, and the data was coded for RTs, overall accuracy

(correct/incorrect response), and accuracy for cognates and non-cognates separately.

Exit Questionnaire. During the post-test, participants filled out a brief exit questionnaire, which

included questions about their self-reported language learning aptitude, their experience with learning through Rosetta Stone, and their motivation to participate in the study and to learn Dutch.

3.2.3 Cognitive measures

In addition to the language tasks, participants completed a battery of cognitive tasks, each targeting one of the subcomponents of Miyake and Friedman’s model of executive functioning (i.e., shifting, updating, and inhibitory control; Miyake & Friedman, 2012).

AX-Continuous Performance Task. Participants completed the AX-Continuous Performance Task

(AX-CPT; Cohen, Barch, Carter, & Servan-Schreiber, 1999) first in the cognitive battery. The purpose of this test is to measure proactive goal maintenance and reactive inhibitory control, and it has been shown to be sensitive to changes in control ability as a result of second language

(9)

8

learning or bilingualism (Morales, Gómez-Ariza, & Bajo, 2013). In this task, participants were instructed to respond as quickly as possible to sequences of five letters that were presented at the center of the screen. The sequence consists of a red letter cue, followed by three white filler letters, and finally a red letter probe. Importantly, participants had to pay attention to the relationship between the probe and the cue. For every cue and filler letter, participants were instructed to press a button corresponding to “No”. Only if the probe was an “X” preceded by the cue letter “A”, participants had to respond with a second button representing “Yes”. In other cases, they had to respond “No” to the probe. Overall, there were four conditions: AX, which required a “Yes” response; AY, for which proactive control was needed; BX, which required reactive control; and BY, the control condition. The AX trials comprised 70% of the 100 trials, and the other conditions were distributed evenly across the other 30%. This distribution serves to make responding to the AY and BX trials more challenging. Consequently, this bias consistently results in greater error and RT costs for AY trials in comparison to AX trials in young adults. As a result, looking at error rates and RTs for correct trials serves as a good indication of individual differences in the engagement of inhibitory control. A schematic of an AX trial in the task appears in Figure 2.

Dot Counting Task. Next, participants completed the Dot Counting Task, which targets verbal

working memory1. Higher working memory has been previously correlated with better second language learning abilities (Juffs & Harrington, 2011). In this task, participants were shown a screen containing blue circles, green circles, and blue squares. They were told to count the blue circles on the screen out loud, one at a time, and to remember the final total. After an increasing number of screens across the experiment (ranging from two to seven screens), participants were prompted to recall the totals of the previous trials in the right order. Responses were scored during the task by the experimenter on a scoring sheet. Auditory responses were recorded for reference. The data was coded for the number of trials the participant could recall correctly (maximum score: 27).

N-Back Task. In our cognitive battery we used the 1-back and 2-back subsets of the N-back task,

representing two difficulty levels. The N-back task targets working memory, but more

1_{According to Miyake and Friedman (2012), working memory largely overlaps with updating.}

A

W

J

L

X

NO NO NO NO YES

(10)

9

specifically, flexible updating capabilities. In this task, a series of squares appeared at 15 different locations on the screen. Participants were instructed to press a response key representing “Yes” when the next square in the series appeared in the same location as the previous square (1-back) or two squares before it (2-back), and “No” when they appeared in a different location. In the 1-back version, participants also saw a number in between each square, which they were required to name out loud before proceeding to the next trial. The data was coded for accuracy, resulting in scores for hit rate and false alarm rate.

Auditory Digit Span Task. The final task in the cognitive battery was the auditory digit span task,

which measures auditory working memory. Participants were instructed to type in the numbers they heard in the correct order. The number of stimuli kept increasing after a couple of trials (range: 3-8). Participants did not complete practice trials for this task. The data was coded for average accuracy per digit condition and average RTs for these conditions.

3.2.4 EEG measures

Resting-state EEG. A robust finding in the neurobiological literature is that power in certain brain

wave frequencies reflects general cognitive abilities, and accurately predicts performance on a variety of cognitive tasks. Previous research has indicated that especially power distributions in the right hemisphere low-beta and gamma frequencies can predict the outcome of second

language learning (Prat, Yamasaki, Kluender, & Stocco, 2016). The specific method that is used to assess this is quantitative electroencephalography (qEEG), which “involves converting

electrophysiological recordings from the time domain to the frequency domain” (Prat et al., p. 45) and subsequently determining the power distribution across frequencies over a number of scalp locations (see Gudmundsson, Runarsson, Sigurdsson, Eiriksdottir, & Johnsen, 2007, for a discussion). Following Prat et al.’s finding, we collected approximately five minutes of eyes-closed resting-state EEG to investigate if differences in power distribution could predict new learning, motivation, and language learning aptitude in our participants, and if power

distributions changed as a function of language learning.

Semantic Categorization Task. In addition to resting-state EEG, participants performed in a

semantic categorization task while their EEG was recorded.

During this task, participants listened to recordings of a total of 160 individually presented spoken Dutch words. The stimuli were recorded by a near-native speaker of Dutch. Half of these words (n = 80) were items that the participants studied during the Rosetta Stone training, and half were items that are not present in the course and were thus (presumably)

unknown to the participant. Both the studied and not studied words contained 50% Dutch-English cognates (e.g., banaan/banana), and 50% non-cognates. Stimuli across these four conditions were balanced for phonological length and frequency per million.

Because we were interested in the extent of semantic integration of newly learned Dutch words, one goal of this categorization task was to ensure semantic processing of the auditory

(11)

10

stimuli. To establish this, participants had to press one of two keys in response to 70% of the stimuli to categorize the target word according to one of two criteria:

Criterion 1: Is it bigger or smaller than a shoe box? Criterion 2: Is it a natural object or a man-made artifact?

Figure 3. Visual cues in the semantic categorization task.

In order to cue the participant as to which of these two criteria they needed to respond to, after hearing the target word, a simple drawing representing the categorization criteria was presented the middle of the screen (see Figure 3). The drawing remained on screen until the participant responded2_{. Importantly, the categorization task was applied only to 70% of the target items, with}

the aim of discouraging anticipation and preparatory motor movements. In each block, the number of items belonging to each semantic category across conditions were controlled for.

The experiment consisted of four blocks. Experiment versions were counterbalanced for stimuli lists and response allocation. In total, there were sixteen versions of the script (e.g., PRE_A_V1, POST_B_V2, etc.) The task took 15-20 minutes to complete.

ERP recording. The electroencephalogram (EEG) for both the ERP task and resting state was

continuously recorded using an 64-channel actiCHamp amplifier (Brain Products, Germany) without hardware filters. The scalp EEG was recorded with 32 active sintered Ag/AgCl

electrodes mounted in an elastic cap (Brain Products actiCAP, Germany) at International 10/20 system sites, and included the right and left mastoids as online reference electrodes. During the recording of resting state EEG, all electrodes were referenced to the left mastoid, and for the ERP task all signals were referenced to the right mastoid. In addition, four electrooculography (EOG) electrodes, placed at the outer canthi and above and below the left eye, recorded horizontal and vertical eye movements, respectively. Data were collected using a DC recording and were digitized at a sampling rate of 1000 Hz. Impedances of the electrodes were kept below 20 kΩ. The EEG data analysis is described in a separate paper (see Preliminary Results and Discussion).

2_{In the original experiment, there was a maximum response window of 4000 ms. However, due to an error in the}

(12)

11

3.2.5 Short-term foreign language learning with Rosetta Stone

Participants learned Dutch during ten sessions using the Rosetta Stone software. Dutch was chosen for the present project’s language learning intervention, because it is a language that participants in Southern California are not familiar with due to lack of exposure. Consequently, those participants start the study with similar knowledge of Dutch. We compare these participants to native English speakers immersed in the Netherlands, where Dutch is the majority language. However, especially in higher education in the Netherlands, English is spoken in class, and knowledge of Dutch is not always required. Consequently, these speakers can keep speaking their first language, but are also exposed to Dutch statistical regularities outside of class.

All experimental participants completed Unit 1 (‘Language Basics’) and Unit 2

(‘Greetings and Introductions’) of Level 1 of the Rosetta Stone curriculum in a laboratory setting on a university laptop. Each unit consists of four ‘lessons’, and each lesson contains a Core Lesson (which introduces new lexical items and grammar), a Pronunciation or Speaking activity, a Grammar lesson, and a Vocabulary activity (see Appendices A-D). In addition, at days five and ten of the training, participants completed a Review block where material of the previous four days was repeated. At the end of each review block, participants completed the Milestone

activity, a brief exercise summarizing the Unit. Participants at UCR completed the curriculum in approximately 5.5-7.5 hours.

All activities consist of listening and/or speaking exercises of increasing difficulty, during which learners listen to native Dutch speech, repeat words or sentences, and freely produce Dutch speech, without receiving any translations or explicit instruction on orthographical, phonological, or grammar rules. Consequently, all learning occurs implicitly, as learners have to derive the meaning of words and sentences through spoken Dutch and accompanying images or the written equivalent in Dutch.

3.3 Preliminary results and discussion

3.3.1 Behavioral results

Unfortunately, discussing the behavioral results of this project is beyond the scope of this

internship report. Preparations have been made to analyze the data, and as soon as more data has been collected, we can resume the behavioral data analysis.

3.3.2 EEG data

In addition to my internship, I followed a tutorial on the analysis of ERP data for 3 ECTS. As part of my assessment for this tutorial I will write an additional report, which will describe the

analysis of that data in more detail. The preliminary results and discussion of the ERP data are therefore included in the tutorial report.

(13)

12

4. Additional Research Activities

Besides conducting research, I participated in a number of additional activities related to doing research. These are worked out in detail below.

4.1 Managing Research Assistants

As mentioned before, undergraduate students are able to work in Prof. Kroll’s lab either

voluntarily or for credits. Because I was anticipating an intense testing schedule, I recruited three undergraduate research assistants. At the beginning of the quarter, I briefly interviewed them, and asked about their research interests and their motivation to work in the lab. I was fully

responsible for managing these RAs. This meant that I trained them in behavioral methods, scheduled their hours, and organized weekly meetings where we discussed testing, and

occasionally, discussed papers related to the project. All in all, it was educational and valuable to teach these students aspects of how to conduct research. Because they worked so closely with me on the project, I was able to track their progress, and give constructive feedback when they needed more guidance. I had to learn to how to respectfully address certain issues, and to explain complex theory in everyday language. These are all competences that will be beneficial to me in the future.

4.2 PSYC 233: Research Methods in Cognitive Psychology – MATLAB

During my internship I had the opportunity to audit an introductory class in MATLAB programming taught by Prof. David Rosenbaum. Prof. Rosenbaum requires that his students present their own work and address which issues they faced and how they overcame them. Essentially, we would teach each other our MATLAB programming skills. I found it valuable to learn from each other’s mistakes and accomplishments. Auditing this class allowed me to gain more knowledge in methodological approaches in the field of psychology. I acquired only basic knowledge of MATLAB programming, but I sincerely believe the class provided me with a good base to take home and apply to my own research. I can now write simple programs and make appealing graphs. The practical nature of Prof. Rosenbaum’s class was refreshing, considering the few courses on research methods I was able to take during my master’s in Groningen. Although the course was at times challenging, Prof. Rosenbaum always kept encouraging us, which motivated me to work hard.

4.3 Meetings

4.3.1 Meetings with supervisors

Prof. Kroll and I held bi-weekly meetings in her office or via Zoom. During these meetings we discussed my progress on the Rosetta Stone project, and we talked about future directions for both myself and the project in general. We also addressed possible ideas for my master thesis. Although Prof. Kroll was my first supervisor on paper, I also held frequent meetings with Dr. Rossi to discuss practical matters relating to the project, and I also coordinated with other lab members and research assistants on future directions for the project. Near the end of my

(14)

13

internship, I also received practical training in EEG data analysis from Dr. Rossi, where she taught me how to use BrainVision Analyzer for pre-processing and analysis.

4.3.2 Writing groups

The writing groups were bi-weekly meetings where a small group of lab members got together for two hours to read each other’s work and provide feedback on it. For my first writing group I was encouraged by Prof. Kroll to get started on a review article which highlights current

theoretical and methodological issues in research related to short-term learning and brain

plasticity. I continued to work on this paper throughout my internship, and Prof. Kroll, Christian, and Emily provided me with useful feedback on my drafts. Prof. Kroll and I approached Dr. Rossi and Dr. Keijzer if they wanted to jump on board and publish this article with us. I will continue to work on the manuscript when I get back and collaborate with Prof. Kroll, Dr. Rossi, and Dr. Keijzer. This part of the internship encouraged me to improve my academic writing skills.

4.3.3 Lab meetings

The lab held weekly meetings on Wednesday afternoons from 2-3:30 pm. During these meetings we discussed recent papers, listened to each other’s practice talks for conferences or job

applications, and discussed general lab matters. Being encouraged to critically evaluate

presentations taught me more about what to include in a conference presentation, or what the best way to present your results on a poster is. In addition, I learned more about what everyone in the lab was working on, and I was able to discuss theoretical issues and results.

Sometimes we had an invited speaker at our meetings, with whom we exchanged research interests. One time I was able to give a short elevator talk about some preliminary results of the Rosetta Stone project in front of the whole lab and Prof. Patricia Kuhl. It was a great experience to be able to present to everyone in the lab what I had been doing, and not to mention, in front of such a great name in the field of cognitive psychology. Prof. Kuhl seemed really interested in my talk and asked a couple of questions, which gave me a great confidence boost.

4.3.4 Cognitive Brown Bag and Bilingualism Matters talks

The Department of Psychology held weekly meetings from 12:10-1:30 pm where researchers from the Cognitive Psychology department or invited speakers presented their research or held discussions about (their) academic careers. I attended all talks that were related to language and bilingualism. These included a talk by Prof. Lise Abrams on proper name representation and processing, and Prof. Kuhl on her academic career path. Prof. Kuhl also gave a longer talk in the afternoon, which was a part of the Bilingualism Matters talks organized by among others Prof. Kroll. During that talk, she elaborated on her research, which includes early language and

bilingual development in young children. These Bilingualism Matters talks were very interesting and inspiring. Attending the Cognitive Brown Bag and Bilingualism Matters talks enabled me to keep abreast of the latest developments in psycholinguistics and cognitive neuroscience.

(15)

14

5. Reflection on the Internship

Before commencing my internship, I could not have imagined the amount of new practical and theoretical knowledge I would gain during my stay in the lab. I believe that actively participating in a research group has provided me with a good insight into the day-to-day life of an academic. Working in this lab made me realize how being successful in academia requires passion,

perseverance, the willingness to work hard, and collaboration.

Overall, this internship allowed me to improve the academic skills I obtained during my bachelor and the first year of my masters in a more practical setting by practicing academic writing, oral presentation skills, and critical thinking. For example, I am in the process of writing a review article, in which I have to critically review previous research and suggest future

directions in an eloquent way. I also practiced critical thinking by providing feedback on drafts and presentations provided by fellow lab members. My (oral) presentation skills were needed for preparing my elevator talk during one of the lab meetings and by constructing plots of ERP data for a conference presentation.

Besides revisiting academic competences I had prior to this internship, I also improved on more general skills applicable to fields outside academia. One of my goals during this internship was to become more independent, but also to be able to work together in a team. When I arrived in the lab in early September, the lab was not yet very active. During this time, I became aware that I was thriving when working independently and simultaneously receiving guidance from the sidelines. In the end, I got the best of both worlds: I was able to work on my own portion of the project largely by myself, but had to communicate large decisions and practical matters with Prof. Kroll, Dr. Rossi, and Isabel. This level of independence allowed me to develop myself more on a personal level and to step out of my comfort zone. Because I was the main person

responsible for the execution of the Rosetta Stone project at UCR, I developed a high sense of responsibility, independency, and autonomy. My tasks included recruiting, scheduling, and testing participants, managing research assistants to assist me with data collection and coding, and pre-processing and analyzing the data. Fitting these tasks into my weekly schedule was certainly challenging, considering the number of studies running in the lab at that time, and the high number of experimental sessions in the project’s design. Nevertheless, I felt that trying to solve this puzzle highly contributed to my organizational skills.

Furthermore, my primary goal of this internship, becoming familiar with EEG

methodology and data analysis, has certainly been reached. By attending practical EEG training sessions led by Megan and eventually being encouraged to explain the methodology to her research assistants, I quickly learned how to address and cap participants, work with BrainVision Recorder, and clean the electrodes and cap. In addition, I now also know the procedure for pre-processing and analyzing EEG/ERP data with BrainVision Analyzer. Acquiring these skills gave me the confidence to possibly design and conduct an ERP experiment for my master thesis. Although originally my internship work plan included working with eye-tracking equipment as well, in the end I am glad that I was able to focus on one particular method, which enabled me to become quite proficient in one method rather than to become average in two.

(16)

15

Certainly during the beginning of the internship, I felt it was difficult to grab theoretical concepts in (cognitive) psychology due to my lack of training in this field, which made it challenging to keep up with discussions sometimes. However, I also feel that by working in this interdisciplinary lab, I gained more knowledge of theoretical concepts and issues in both

linguistics and cognitive psychology and neuroscience by being able to listen to discussions of these concepts during meetings I attended. Hopefully I will be able to contribute more to these discussions myself in the near future.

Another valuable competence, which I did not anticipate of acquiring, was the ability to teach research skills to others. Not only was I able to teach my fellow classmates in the

MATLAB class, I also taught my research assistants more about how to conduct bilingualism research and how to code. I feel that it would have been more difficult to acquire these skills regarding managing research assistants in Groningen, because labs in Groningen are structured differently. Thus, I am grateful that I got the opportunity to acquire it at UCR. Another instance in which I got to share my knowledge was when I explained my research to my roommates and their friends. Because their majors were completely different from mine, I was able to practice explaining difficult concepts in everyday language.

All in all, this internship has provided me with a couple of exciting future opportunities. For instance, I will continue to be involved with the execution of the project in Groningen, and will likely collaborate with the other researchers on writing a scientific article once more participants have been tested. In addition, working on this project has also inspired me to continue with the project’s main ideas for my master thesis. Furthermore, I feel that meeting people working in the field of research that I am passionate about will prove to be a valuable addition to my network. Lastly, although I was not able to attend a conference during my stay in California, abstracts including my data have been submitted to other conferences, which is very exciting. During this internship, I learned a lot about myself and my capabilities. Prior to going on this internship I had already considered applying for a PhD program, and this internship has highly contributed to my decision whether to pursue this. Even if I turn out not to apply to any positions, I believe that the competences I acquired here will benefit me in any career.

During my internship I was supervised by Dr. Keijzer over bi-weekly Skype meetings, during which we mainly discussed my activities in the lab. I liked the idea of being able to connect with the home front on a regular basis, and I feel that Dr. Keijzer provided me with plenty of space to be independent, while also making clear that she would be available in case I needed help. I also enjoyed the meetings with Prof. Kroll and Dr. Rossi, considering that their passion for research was highly contagious and that their ideas were very inspiring. Although I arrived during a busy time for both women, which resulted in me having to rely on myself a lot, it felt good to know that they would be always there when I would need them. I want to express my gratitude towards all my supervisors for being involved, encouraging, and inspiring. I also want to sincerely thank Mariamme, Vicky, and my research assistants for helping me with data collection, and thanks to all the lab members for adopting me in their lab, encouraging me, and making me feel at home in California. This internship was an experience I will never forget.

(17)

16

References

Bak, T. H. (2016). Cooking pasta in La Paz: Bilingualism, bias, and the replication crisis.

Linguistic Approaches to Bilingualism, 6(5), 699–717.

Bak. T. H., Long, M. R., Vega-Mendoza, M., & Sorace, A. (2016). Novelty, challenge, and practice: The impact of intensive language learning on attentional functions. PLoS ONE,

11(4), 1-12.

Baum, S., & Titone, D. (2014). Moving toward a neuroplasticity view of bilingualism, executive control, and aging. Applied Psycholinguistics, 35 857–894.

doi:10.1017/S0142716414000174

Cohen, J. D., Barch, D. M., Carter, C., & Servan-Schreiber D. (1999). Context-processing deficits in schizophrenia: converging evidence from three theoretically motivated cognitive tasks. J. Abnorm. Psychol., 108, 120–133. doi:10.1037/0021-843x.108.1.120 Eyer, I. S. (2018). The neural processes and cognitive effects of short intensive second

language learning (Unpublished master’s thesis). University of Groningen, the

Netherlands.

Grant, A., Dennis, N. A., & Li, P. (2014). Cognitive control, cognitive reserve, and memory in the aging bilingual brain. Frontiers in Psychology, 5(DEC), 1–10. https://doi.org/10.3389/fpsyg.2014.01401

Gudmundsson, S., Runarsson, T. P., Sigurdsson, S., Eiriksdottir, G., & Johnsen, K. (2007). Reliability of quantitative EEG features. Clinical Neurophysiology, 118(10), 2162–2171. Juffs, A., Harrington, M. (2011). Aspects of working memory in L2 learning. Language

Teaching, 44, 137-166.

Kroll, J. F., Bobb, S. C., & Hoshino, N. (2014). Two languages in mind: Bilingualism as a tool to investigate language, cognition, and the brain. Current directions in psychological

science, 23(3), 159-163.

Linck, J. A., Kroll, J. F., & Sunderman, G. (2009). Losing access to the native language while immersed in a second language: evidence for the role of inhibition in second-language learning. Psychological science, 20(12), 1507-15.

Mårtensson, J., Eriksson, J., Bodammer, N. C., Lindgren, M., Johansson, M., Nyberg, L., & Lövdén, M. (2012). Growth of Language-Related Brain Areas after Foreign

Language Learning. NeuroImage, 63(1), 240-244.

McLaughlin, J., Osterhout, L., & Kim, A. (2004). Neural correlates of second-language word learning: minimal instruction produces rapid change. Nature Neuroscience, 7(7), 703– 704. doi: 10.1038/nn1264

Miyake, A., & Friedman, N. P. (2012). The Nature and Organization of Individual Differences in Executive Functions : Four General Conclusions. Current Directions in Psychological Science, 21(1), 8–14.

Morales, J., Gómez-Ariza, C. J., & Bajo, M. T. (2013). Dual mechanisms of cognitive control in bilinguals and monolinguals. Journal of Cognitive Psychology, 25(5),

(18)

17

Osterhout, L., Poliakov, A., Inoue, K., McLaughlin, J., Valentine, G., Pitkanen, I., . . .

Hirschensohn, J. (2008). Second-language learning and changes in the brain. Journal of

Neurolinguistics, 21(6), 509–521. doi:

10.1016/j.jneuroling.2008.01.001.Second-language

Prat, C. S., Yamasaki, B. L., Kluender, R. A., & Stocco, A. (2016). Resting-state qEEG predicts rate of second language learning in adults. Brain and Language, 157-158, 44–50. doi: 10.1016/j.bandl.2016.04.007

Rossi, E., Cheng, H., Kroll, J. F., Diaz, M. T., & Newman, S. D. (2017). Changes in white-matter connectivity in late second language learners: Evidence from diffusion tensor imaging. Frontiers in Psychology, 8, 2040. doi:10.3389/fpsyg.2017.02040

(19)

18

Appendices

Appendix A. Example of a Rosetta Stone Core Lesson activity

Appendix B. Example of a Rosetta Stone Pronunciation activity (left) and a

(20)

19

Research Internship Report