Giulia Calignano on Labvanced Research for Language Research on Memory with Song and Prosody

The semantic interference in 9- to 36-month old infants: An at-home eye-tracking study on infants' lexical abilities

Researcher: Giulia Calignano, University of Padua

In this interview, Giulia Calignano shares her experience in working with infants and eye tracking during a looking-while-listening task. Check out the abstract from her poster presentation at WILD 2022open in new window on page 127 and keep reading to learn about her motivation and take home messages from using Labvanced.

Tell us about your research background and your field of study.

I started working on infant cognition at University of Padova during my internship following my Master’s degree in Neuroscience. Afterwards, I continued collaborating there at the Babylab, where I had the opportunity to learn specific techniques for investigating cognition in babies ages 4 months to 3 years. We use eye tracking (and pupillometry) to capitalize on the physiology of the eye in order to investigate the interface between attention and perception. Those techniques are not invasive and the babies can look at stimuli as they want. We can collect a good amount of information from the free looking of babies.

Then, I earned a PhD in psycholinguistics (University of Trento) and investigated the interface between attention and language in infants and adults. Using eye tracking, I specifically examined the impact of spoken and written language in guiding attention.

Now, I am a post-doctoral researcher at the Department of Developmental and Social Psychology, University of Padua, still investigating the interplay between language and attention across the life span.

What motivated you to look into this area?

As a psychology student, I studied a lot about cognitive processes and behavior in adults, but the question has always been ‘when and how does everything start?’ and ‘how can you develop into such a complex individual?’

I developed a passion for infant cognition because it allows you to trace the origin of adult psychology, that can be seen in terms of development outcomes. The orienting of audiovisual attention was the main topic, then understanding that language in infants can guide and shape behavior and cognition before they can even speak has played a central role in my research interest. It is more difficult than studying adults! Observing infant behavior is like observing a more free, less constrained behavior and it open a window of plausible explanation about why and how humans actively build their knowledge from environmental stimulations.

What are you currently working on or have recently worked on that you'd like to discuss now with us?

We recently presented a poster at the Workshop of Infant Language Development (WILD 2022) about lexical comprehension and semantic interference in 9- to 36-month-old children. It is an ongoing project, so the results are preliminary but very intriguing. I and my colleagues, i.e. A. Micelli, S. Russo, E. Di Giorgio, N. Reoyo Serrano, S. E. Benavides Varela & E. Valenza, implemented a classical paradigm – the Looking-while-Listening task. The main idea was to replicate some classical results by using Labvanced as a new tool for eye tracking online.

Giulia Calignano at the WILD 2022 conference presenting her poster on language acquisition

Can you please describe the research design and how you set up the experiment?

In the Looking-while-Listening task that is a well-established lexical comprehension ability task, we present 2 images to the infant. The pictures can be 2 objects from the same category, like an apple and a banana, or 2 objects from different categories, like an apple and a toy. There is a period of time where the infant can freely look at both images, then a voice names one of the images. We measure eye movements and time spent on each picture to see if the label guides attention towards the target image. Usually it is expected to find a pattern of exploration while looking at both pictures for same amount of time (not including individual preference), then after the voice named a picture, the baby is expected to look at the named picture more than the other. Doing so we aim at measuring the vocabulary size in infants who cannot speak; basically the fundamental question behind are ‘when does the vocabulary start to develop?’ and importantly ‘how linguistic and semantic knowledge guides pre-verbal and verbal infants’ behavior?’

Tell us about your conclusion and the implications of that.

Our preliminary results match traditional in-person findings from the same classic paradigm. We found that semantic interference (having two pictures from the same category) introduced interference in the ability of the child to select the correct image. That is, objects from different categories are easier to identify because the differences are more evident at both the perceptual and the linguistic level. In addition, we are interested in tracing the developmental trajectory of semantic development that is, detecting when from 9- to 36-months of age, infants start to organize information into categories.

What are your next steps with this research?

Within this project, we intend to recruit more infants and increase our sample size. We will continue to collect data to be able to generalize to the broader population by also including as many languages as possible. We intend doing so also by taking advantage of the online eye-tracking to reach families that usually do not come in our lab.

Do you have a tip or trick for colleting good data that you could share?

The measurement error is always around the corner but we can deal with it. In my view, in addition to selecting a sound experimental paradigm tailored to a specific research question and developmental window, a safe way to collect useful data in general is to get informative datasets (equipped with data dictionary). That is, collecting sufficent information about the main sources of noise that interfere with the timecourse of a planned experiment in order to run sanity and robustness checks.

In what ways is conducting research online different from in-lab?

It is a new frontier with online research. We changed the setting but conducted the same experiment, so there were gains and losses.

You lose meeting with the family and having the in-person experience. It is important for the family to know more about research and for the experimenters to know more about the baby and the parents. There is less opportunity to know them as people.

We gained the possibility to reach more families and to make it easier for those families who live abroad in different areas. They are now able to collaborate and participate in research, by choosing the moment they prefer to participate. We also trained the parents to be experimenters.

We can also think about multi-lab studies to obtain more robust results. This opens new lines of research to shape with new degrees of freedom.

Do you see online research as the future of your field?

I see online research as a useful complementary tool. Research, in general, is such that every day there are new questions. Different questions require different tools. Online research still has limitations, but it is a new opportunity that will be useful for answering many research questions for instance about the specific impact of the environment in which a baby is immersed on language and attention development across context and cultures.

How did you choose Labvanced for your research?

I found Labvanced on GitHub. I really liked that it was an open tool. I could go through the code and logic of the algorithm of the software, and I was immediately able to build a classical experiment on the platform. I loved that it was easy to use on an interface level and also allows the possibility to use code inside the events. So I started spreading the word in my department.

What stands out to you about Labvanced?

I am passionate about data analysis, so I liked the way that you can download data and select the format and information that you need. It also includes the technical information on how the experiment worked, as well as the measures that you’re interested in. Finding the code on GitHub was easy and it was nice to see how to make plots and examine the measurement accuracy, so we could easily compare in-person data with the online data that we had collected.

What would you recommend to students hoping to begin research in your field?

Collaborate! Look for opportunities to work with other labs within your field and even researchers from other fields. Raise an open science mind and use errors as a learning opportunity. I suggest that you try to be a part of multi-lab studies as well.

Do you have a message to share with other Labvanced users?

Let’s collaborate more! We should share our knowledge about designs but also tips and tricks to collect the best data for example by also sharing paradigms to get more comparable studies.

Can you tell us about the WILD2022 conference you recently attending and your experience of sharing your results?

The Workshop on Infants Language Development (WILD) is a great occasion to discuss the most recent evidence about the underlying mechanisms of language development in infancy, across languages and countries. In particular, it attracs senior as well as early career researchers by offering a unique opportunity to let ideas circulate.