by Daniel Messinger and Alyssa Viggiano
(Jointly authored; written from Daniel’s perspective)

Autism is a common and sometimes debilitating neurodevelopmental disorder whose symptoms typically emerge in the first three years of life. By focusing on three areas of autism research, I hope to illustrate autism’s attraction to the developmentalist. The three areas are (1) babysibs investigations, as well as research using objective measurements of autism relevant behavior (2) in the clinic as well as (3) in the classroom. First, though, a bit of personal history.

For the first 15 years of my career, my research focused on understanding the dynamics of development in what we frequently term “typically developing” infants. I was particularly interested in nonverbal communication—the interactive development of smiling, referential gaze, and gesture. Clinical syndromes—especially those like autism spectrum disorder (ASD) that involved exacting and occasionally confusing nomenclatures and technical diagnostic procedures—did not interest me. But Peter Mundy, a mentor and colleague, fatefully suggested to me that my expertise in nonverbal communication might be particularly suited to babysibs research. This was a crucial step in my own development.

Babysibs Research

Babysibs are the younger siblings of children with confirmed ASD. As detailed by Wilson (2023) in a recent BabyBlog, babysibs are at elevated likelihood for ASD simply because they have an older brother or sister with ASD. About one in five babysibs will themselves have an ASD outcome, and those without an ASD outcome can display subtle signs of the disorder (Charman et al., 2016; McDonald et al., 2019; Messinger et al., 2013; Messinger et al., 2015; Ozonoff et al., 2011). Fundamentally, babysibs research using its characteristic prospective design (following the infants at elevated likelihood over time) is tantalizing because it allows one to prospectively study the development of ASD. Salient examples involve sex differences, the concept of risk in the emergence of ASD, and ASD genetics.

A developmental perspective in babysibs research was helpful in understanding ASD sex differences (Messinger, et al., 2015). The clinical literature had not been clear whether girls with ASD had more or less severe symptoms than boys with ASD. The developmental design of babysibs research showed us that there were indeed sex differences in ASD symptoms between one and three years, but (arguably) they had nothing to do with ASD. As you might imagine, elevated likelihood infants with eventual ASD had the highest level of symptoms in our longitudinal design, followed by elevated likelihood infants without eventual ASD, followed by low likelihood infants. Strikingly, at least for me, boys had higher levels of restricted and repetitive behaviors—a key ASD symptom—than girls (Messinger, et al., 2015). And this sex difference held for all infants — whether or not they were at elevated risk for ASD, and whether or not they were eventually identified as having ASD. So, sex differences in ASD symptoms were simply a reflection of a larger normative difference between boys and girls (like girls’ higher levels of verbal abilities, which was also evident in this research). The clinical literature had suggested that sex differences in ASD reflected sex-specific manifestations of the disorder, which remains a distinct possibility and robust area of research (Burrows et al., 2022). Nevertheless, a parsimonious developmental perspective suggested that sex differences in ASD were simply a reflection of sex differences in infants more generally.

At a more personal level, the clinical import of ASD risk in babysibs research was an epistemological shocker that made me think about what is known, and not known, about development. Consider an infant at elevated likelihood for ASD, maybe a 6-month-old in the face-to-face/still face (FFSF) protocol. Let’s say this specific infant will be diagnosed with ASD two and half years in the future, at age 3. Did I believe an immutable causal mechanism was at play in the infant at 6 months? That is, did I believe ASD was latent in the infant, waiting to show itself? What about a more dynamic story? Would a series of developmental processes and cascades occur over the next two years of the infant’s life that would shape the emergence of their ASD? This latter approach was a more comfortable perspective for me. But did that perspective imply that the infant’s ASD could be avoided simply by altering their environment? Or if ASD could not be avoided, could the infant nevertheless develop functional skills that would improve their outcomes? Babysibs research made me confront the concrete implications of my theoretical commitments.

Empirical research certainly shed light on these developmental puzzles, but it did not make them less puzzling. ASD has recognizable behavioral characteristics in the third year of life but predicting those behaviors has proved challenging. Here’s an example. At six months, infants with eventual ASD show a less pronounced reduction in smiling from the play to the reunion episodes of the FFSF (Lambert-Brown et al., 2015). They show a lower ratio of gazing to the eyes (relative to the mouth) in eye-tracking tasks (Constantino et al., 2017; Jones et al., 2013). By 12 months, infants with later ASD show lower levels of initiating joint attention including joint attention that integrates smiling, referential looking, and coordination of multiple motor modalities (Adamson et al., 2019; Gangi et al., 2014; Iverson, 2021). Deficits in these triadic or referential communication abilities are characteristics of ASD itself. These are tantalizing results and reflect decades of productive research. But what predicts the deficits themselves?

One clue might be inheritance and genetic mutations. ASD research is interdisciplinary science and provided me an education—for example, in genetics–I would otherwise lack. ASD is highly heritable, yet genetic variants that confer high risk for ASD (highly penetrant variants) account for only a small percentage of ASD cases. Likewise, the combination of behavioral and genetic findings in babysibs research suggests that transmission of common negative variants also confers risk for ASD, but we do not know what common genetic variants are involved (D’Abate et al., 2019). It is understandable that an older child and their younger sibling, both with ASD, would have genetic variants linked to ASD. Paradoxically, however, the older and younger sibling may carry different genetic variants linked to ASD, suggesting that ASD risk may be associated with a susceptibility to such mutations rather than the presence of any specific one. Ultimately, ASD genetics can inform but not fully explain the emergence of ASD symptoms. However, some intellectual purchase on the emergence of ASD symptoms—such as a paucity of gazes to a partner’s face—can be gained by a foray into the world of ASD diagnosis and characterization.

Objectively measuring ASD-relevant behavior in the clinic

ASD is defined behaviorally, in that expert clinicians diagnose the disorder based on children’s behavior. We asked whether we could objectively quantify behaviors indexing autism symptoms during the gold standard clinical ASD assessment termed the ADOS (Autism Diagnostic Observation Schedule—remember my earlier trepidation about technical nomenclature and technical diagnostic procedures). Signal analysis of audio indicated that child vocalizations of higher frequency were associated with higher levels of ASD restricted and repetitive behavior symptoms (Moffitt et al., 2022). Furthermore, using computer vision tools, we found that lower levels of gazing at the parent’s face and lower levels of social smiling at the parent were associated with higher levels of ASD social-affective symptoms (Ahn et al., 2023). Well and good. Objective methods validated clinical characterizations of lower levels of social gaze at the parent in young children with ASD.

But wait. Adolph and West (2022) argue that gaze at the caregiver’s face is rare during naturalistic play with toys in both ASD and neurotypical kids (Yurkovic-Harding et al., 2022) and consequently that gazing at the caregiver’s face it is unlikely to be a mechanism for the emergence of the social deficits that characterize autism. However, our ADOS results suggest that gazing at the parent’s face—at least during playful but challenging interaction with a stranger—speaks to a different facet of everyday experience (Ahn, et al., 2023). In the ADOS, the child is confronted with challenges and surprises in which visually referencing the parent may be a key source of distraction from frustration and affective reassurance. Lower levels of gazing (and smiling) at the parent in this situation suggest that children with ASD are less likely to avail themselves of socioemotional information (and less likely to participate in socioemotional communication) crucial to confronting ambiguous situations. These results suggest (at least to me) a link between decreased social attention and ASD-linked social deficits. But we still lack a detailed understanding of how, for example, relative inattention to the eyes during social interaction contributes to ASD-linked deficits in referentiality and emotion expression.

ASD goes to preschool

Objective characterization of ASD cannot be limited to assessment settings like the ADOS. By definition, ASD involves persistent social communication deficits. But in practice we lack a quantitative sense of what those social communication deficits look like and the degree to which they are persistent or consistent across contexts.  Given this, we tried to grab the bull by the horns by objectively measuring the physical movement and vocalizations of 3- to 5-year-old children with ASD and their neurotypical peers in inclusion preschools at weekly to monthly intervals over the course of the school year. On the technical side, we paired a real time location system with machine learning identification of children’s vocalization from lightweight audio recorders (Banarjee et al., 2023; Fasano et al., 2023; Fasano et al., 2021).

Patterns of homophily characterized the movement and social contact of children with ASD and their neurotypical (non-ASD) peers. That is, children tended to approach peers who were similar with respect to ASD status (ASD-ASD and neurotypical-neurotypical) more quickly than they approached less similar children (e.g., ASD-neurotypical). Likewise, although both children with ASD and neurotypical children spent more time in social contact with like children, ASD-ASD pairs spent less time in social contact than neurotypical-neurotypical pairs (Banarjee, et al., 2023).

Reminiscent of our investigation of babysib sex differences, what we learned about children’s vocalizations while in social contact reflected general communication patterns as well as ASD-specific deficits. First, the general pattern. For all kids in a preschool inclusion classroom, the number of vocalizations one child directed to another (e.g., José directed to David) on one observation (when we made objective measurements in the classroom) predicted the number of reciprocated vocalizations (e.g., David directed to José) on the next observation (typically 2 – 4 weeks away). This pattern of dyadic communication was true for children with ASD and for their neurotypical peers.

We think of the classroom as a network in which one child is connected to another (e.g., José and David) by the quantity of speech they direct to one another, which we call co-talk. Children with ASD tended to be isolated in the co-talk classroom network. Both their co-talk bonds with neurotypical children and their bonds with other children with ASD were weaker than the co-talk bonds between neurotypical children.

It turns out that co-talk is a strong predictor of children’s end-of-year assessed language abilities, a potent reminder of the importance of social communication in peer groups (Fasano, et al., 2021; 2023). Overall, children with ASD were both spoken to less than and spoke less to their peers than neurotypical children. Moreover, the assessed language abilities of children with ASD were depressed. But the link between co-talk and language abilities — between interaction and indices of development—did not vary by ASD status (Fasano, et al., 2021). The results suggest the potential of peer vocal interaction to support the language abilities of all children in a classroom.

Conclusion

In brief, ASD research affords infancy researchers the opportunity to engage with developmental processes of serious clinical import, and to consider the role of group dynamics in children’s lives.  ASD research is fast-paced interdisciplinary science. In some of my non-ASD infancy research, a decade may pass before a finding is replicated, not replicated, or something in between. Not so in ASD research. In a year or so, the field often provides feedback on whether your latest result replicates. It is thus an important and exciting branch of infancy research.

References

Adamson LB, Bakeman R, Suma K, Robins DL (2019) An Expanded View of Joint Attention: Skill, Engagement, and Language in Typical Development and Autism. Child Development 90:e1-e18.https://doi.org/10.1111/cdev.12973.

Adolph KE, West KL (2022) Autism: The face value of eye contact. Curr Biol 32:R577-r580.10.1016/j.cub.2022.05.016. PMC9527854. NIHMS1838065.

Ahn YA, Moffitt JM, Tao Y, Custode S, Parlade M, Beaumont A, Cardona S, Hale M, Durocher J, Alessandri M, Shyu ML, Perry LK, Messinger DS (2023) Objective Measurement of Social Gaze and Smile Behaviors in Children with Suspected Autism Spectrum Disorder During Administration of the Autism Diagnostic Observation Schedule, 2nd Edition. J Autism Dev Disord.10.1007/s10803-023-05990-z.

Banarjee C, Tao Y, Fasano RM, Song C, Vitale L, Wang J, Shyu ML, Perry LK, Messinger DS (2023) Objective quantification of homophily in children with and without disabilities in naturalistic contexts. Sci Rep 13:903.10.1038/s41598-023-27819-6. PMC9845319.

Burrows CA et al. (2022) A Data-Driven Approach in an Unbiased Sample Reveals Equivalent Sex Ratio of Autism Spectrum Disorder-Associated Impairment in Early Childhood. Biol Psychiatry 92:654-662.10.1016/j.biopsych.2022.05.027. PMC10062179. NIHMS1877035.

Charman T et al. (2016) Non-ASD outcomes at 36 months in siblings at familial risk for autism spectrum disorder (ASD): A baby siblings research consortium (BSRC) study. Autism Res.10.1002/aur.1669.

Constantino JN, Kennon-McGill S, Weichselbaum C, Marrus N, Haider A, Glowinski AL, Gillespie S, Klaiman C, Klin A, Jones W (2017) Infant viewing of social scenes is under genetic control and is atypical in autism. Nature 547:340-344.10.1038/nature22999

http://www.nature.com/nature/journal/v547/n7663/abs/nature22999.html#supplementary-information.

Fasano RM, Perry LK, Zhang Y, Vitale L, Wang J, Song C, Messinger DS (2021) A granular perspective on inclusion: Objectively measured interactions of preschoolers with and without autism. Autism Research 14:1658-1669.10.1002/aur.2526.

Fasano RM, Mitsven SG, Custode SA, Sarker D, Bulotsky-Shearer RJ, Messinger DS, Perry LK (2023) Automated measures of vocal interactions and engagement in inclusive preschool classrooms. Autism Research n/a.https://doi.org/10.1002/aur.2980.

Gangi DN, Ibanez LV, Messinger DS (2014) Joint Attention Initiation With and Without Positive Affect: Risk Group Differences and Associations with ASD Symptoms. J Autism Dev Disord 44:1414-1424.10.1007/s10803-013-2002-9. Pmc4024338. Nihms544778.

Iverson JM (2021) Developmental Variability and Developmental Cascades: Lessons From Motor and Language Development in Infancy. Current Directions in Psychological Science:0963721421993822.10.1177/0963721421993822.

Jones W, Klin A (2013) Attention to eyes is present but in decline in 2-6-month-old infants later diagnosed with autism. Nature 504:427-431.10.1038/nature12715. Pmc4035120. Nihms527415.

Lambert-Brown BL, McDonald NM, Mattson WI, Martin KB, Ibañez LV, Stone WL, Messinger DS (2015) Positive emotional engagement and autism risk. Developmental Psychology 51:848-855.10.1037/a0039182.

McDonald NM, Senturk D, Scheffler A, Brian JA, Carver LJ, Charman T, Chawarska K, Curtin S, Hertz-Piccioto I, Jones EJH, Klin A, Landa R, Messinger DS, Ozonoff S, Stone WL, Tager-Flusberg H, Webb SJ, Young G, Zwaigenbaum L, Jeste SS (2019) Developmental Trajectories of Infants With Multiplex Family Risk for Autism: A Baby Siblings Research Consortium Study. JAMA Neurology.10.1001/jamaneurol.2019.3341.

Messinger D, Young GS, Ozonoff S, Dobkins K, Carter A, Zwaigenbaum L, Landa RJ, Charman T, Stone WL, Constantino JN, Hutman T, Carver LJ, Bryson S, Iverson JM, Strauss MS, Rogers SJ, Sigman M (2013) Beyond autism: A baby siblings research consortium study of high-risk children at three years of age. Journal of the American Academy of Child and Adolescent Psychiatry 52:300-308 e301.10.1016/j.jaac.2012.12.011. PMC3625370. Nihms431543.

Messinger DS, Young GS, Webb SJ, Ozonoff S, Bryson SE, Carter A, Carver L, Charman T, Chawarska K, Curtin S, Dobkins K, Hertz-Picciotto I, Hutman T, Iverson JM, Landa R, Nelson CA, Stone WL, Tager-Flusberg H, Zwaigenbaum L (2015) Early sex differences are not autism-specific: A Baby Siblings Research Consortium (BSRC) study. Mol Autism 6:32.10.1186/s13229-015-0027-y. Pmc4455973.

Moffitt JM, Ahn YA, Custode S, Tao Y, Mathew E, Parlade M, Hale M, Durocher J, Alessandri M, Perry LK, Messinger DS (2022) Objective measurement of vocalizations in the assessment of autism spectrum disorder symptoms in preschool age children. Autism Research, 15( 9), 1665– 1674 https://doiorg/101002/aur2731

Ozonoff S, Young GS, Carter A, Messinger D, Yirmiya N, Zwaigenbaum L, Bryson S, Carver LJ, Constantino JN, Dobkins K, Hutman T, Iverson JM, Landa R, Rogers SJ, Sigman M, Stone WL (2011) Recurrence risk for autism spectrum disorders: a Baby Siblings Research Consortium study. Pediatrics 128:15

Wilson R (2023) How quantitative evaluation of early infant movement may give us insight into autism. ICISS Babyblog. https://infantstudies.org/how-quantitative-evaluation-of-early-infant-movement-may-give-us-insight-into-autism/

Yurkovic-Harding J, Lisandrelli G, Shaffer RC, Dominick KC, Pedapati EV, Erickson CA, Yu C, Kennedy DP (2022) Children with ASD establish joint attention during free-flowing toy play without face looks. Current Biology.https://doi.org/10.1016/j.cub.2022.04.044.

About the Author

Developmental electroencephalography (EEG) is a constantly evolving field with researchers actively exploring cutting-edge approaches to studying brain development. One novel approach involves looking at live, naturalistic interactions by measuring changes in an infant’s EEG as a response to discrete ‘events’ or conditions. Another approach looks at general patterns in a free-flowing interaction rather than discrete events, and finds relations between infants’ EEG and ongoing environmental stimuli such as speech. A third approach looks at statistical ‘profiles’ or groups that emerge from studying infants’ ‘resting’ EEG, which captures background brain activity related to brain maturation and organization. All three of these approaches were showcased in a recent ICIS Webinar.

The first approach, discussed by Dr. Lindsay C. Bowman from the University of California Davis, consists of event-related analyses of live, naturalistic interactions between the infant and a social partner. The basic experimental design for this type of research involves an infant wearing an EEG cap while interacting with a social partner, coding for ‘events of interest’ (EOI) such as an infant’s gaze shift indicating a shift in their attention, and measuring the corresponding changes in the infants’ EEG. To do this, researchers must create experiences where a target EOI will occur, and then analyze the resulting EEG output in relation to the EOI and proposed hypotheses. This method allows researchers to examine an infant’s neural responses in an ecologically valid way by providing live events as stimuli, rather than the classic computer-based paradigms that can only simulate real-life events of interest.

Despite its enhanced ecological validity, there are challenges to this approach including equipment setup, study design, and analysis parameters. For example, because this method requires infant-caregiver interactions, the classical approach of having the infant sit on the caregivers’ lap is not appropriate. Possible solutions involve using carefully sourced chairs (e.g., a “sit-me-up” chair for young infants to provide security and prevent EEG cap compression). There must also be multiple camera angles to capture both the behavior from the infant and also the interacting social partner (e.g., caregiver). Researchers can use a hidden earpiece to communicate with the caregiver while allowing for a more naturalistic interaction as opposed to the experimenter being in the room and possibly causing disruptions to the interactions.

Following data collection, researchers use an analysis technique called ‘event-related EEG spectral power analysis’ to compare the data captured during a target event of interest (EOI) in relation to the data taken during a ‘baseline’, or non-target event. Good baselines occur close in time to the EOI, preferably directly before it, and are similar to the EOI. Researchers must determine what their baseline will be, and also how to select an ‘epoch’ or time frame of EEG to analyze for their EOIs. For example, if researchers want to study infant brain activity supporting infants’ attention to an object to which their caregiver is also attending (a type of activity called ‘joint attention’), then researchers must decide when to extract the EEG data to best capture the brain activity associated with this EOI— should the epoch begin before the infant shifts their gaze to the object? After? During the shift? Researchers must consider these issues when analyzing EEG data, and often will use previous studies to inform their analysis decisions.

A second cutting-edge approach to infant EEG research was discussed by Dr. Sam Wass from the University of East London, who also studies EEG during free-flowing naturalistic social exchanges. Dr. Wass’ approach—unlike Dr. Bowman’s—focuses on non-event-locked analysis, meaning that instead of finding specific events to analyze (such as an infant gaze shift), researchers examine the entirety of a free-flowing interaction, and look for patterns of connection between the neural activity of the two interacting partners. Like Dr. Bowman’s approach, this approach enables examining infant brain activity supporting live, naturalistic interactions, which Dr. Wass noted improves significantly upon past non-naturalistic techniques. For example, previous, non-naturalistic studies examining infants’ brain responses to gaze aversion, gaze following, and emotional expressions have limited real-world application because these studies flash static images of faces on a computer screen out of nowhere and for brief time periods, unlike the continuous reality people actually experience. Human social interaction is just that—interactive—and so examining infant brain activity only during passive-non-interactive computer-based paradigms seriously limits our understanding of how the human brain actually works to support activity in every-day life.

Similar to Dr. Bowman, Dr. Wass also discusses the challenges to studying EEG in a free-flowing and naturalistic environment. Chiefly, EEG data is susceptible to ‘artifact’ or ‘noise’ generated by movements of the body, face, and eyes, which mask researchers’ ability to analyze actual brain signals captured in the EEG. One possible solution to the artifact problem would be using software that allows for tracking minute movements based on previous data. These software programs, such as OpenFace and MediaPipe, allow researchers to get micro-levels of how participants move their heads and can quantify their facial expressions so that these influences on the EEG signal can be filtered out. Another solution involves side-stepping artifacts entirely by finding specific relations between infant brain activity and the free-flowing naturalistic stimuli in the environment, such as speech patterns during caregiver-infant interactions. Researchers can use statistical modeling that identifies an environmental stimulus (e.g., caregiver speech) and evaluates whether and how the infant brain activity changes as a function of the environmental stimulus pattern. This same modeling approach also allows researchers to examine how the infant brain activity may be related to the brain activity of their social partner during an interaction. These kinds of methods are referred to as environment-brain and brain-brain ‘entrainment’. Entrainment methods can address several problems related to EEG artifacts.

A third approach to infant EEG research was discussed by Dr. Lara Pierce from York University. Unlike Dr. Bowman and Dr. Wass, Dr. Pierce discussed how to examine infant brain activity from ‘resting’ EEG, which is the EEG recorded when infants are not doing any sort of interaction or task in particular. Of course, the brain is always ‘active’, even when we sleep, and so there is always brain activity to measure. The measurement of ‘resting EEG’ is meant to capture a picture of the activity in the ‘idling’ brain, which captures ‘background’ activity related to general brain maturation and organization.

Dr. Pierce discussed a new statistical technique for extracting information from resting brain activity called ‘Latent Profile Analysis’. This technique separates participants into ‘profiles,’ which consist of groups of participants that show similar responses across multiple measures (e.g., of the brain, demographics, cognition). This approach helps researchers to identify unique patterns of EEG activity that exist within our diverse human population, providing insight into which measures or ‘variables’ predict a given profile’s development; for instance, whether those who exhibit that profile are likely to share later developmental outcomes. Specifically, latent profile analysis uses the ‘scores’ of an individual on a given measure—called an ‘indicator variable’—to calculate the probability that the individual belongs to a particular profile. Researchers use theory to select a number of indicator variables and then run a sequence of models to obtain the “best fit” for different profiles. This method is broadly applicable to all sorts of investigations, and has the advantage of being able to consider the combined influence of multiple different variables altogether (such as gender, cultural background, multiple aspects of the EEG signal), rather than having to separate these variables. Thus, latent profile analysis broadens our understanding of the complex and interactive influences distinct variables have on infant development.

In sum, the webinar showcased the value of using EEG in understanding infant development, and highlighted different approaches: both event-locked and entrainment analyses of EEG collected during live, naturalistic social interactions, and latent profile analysis of infants’ ‘resting’ EEG. All three approaches are on the cutting edge of developmental EEG research, and open up many new avenues for future work.

About the Author

by David Tompkins

When labs and universities shut down for the COVID-19 pandemic, researchers flocked to remote methods to keep their research moving. For many infant researchers, this meant employing unmoderated looking time experiments through platforms like Gorilla or Lookit. These looking time studies are conducted on the participant’s device and involve the participant watching a prepared sequence of stimuli. The experimenter later receives a webcam recording and then codes where on the screen the child appeared to be looking. Although these methods played a critically important role in allowing research to continue during the shutdown and may have the potential to help researchers reach new participant pools, they may also introduce yet-to-be-considered new challenges and new sources of variation in participants’ performance on our tasks.

There are many salient reasons to expect some challenges or increased variability in remote testing. Each participant views our stimuli in spaces with different distractions and on devices with different settings and dimensions. It is reasonable to expect that participants in homes with barking dogs or crying siblings might pay a different amount of attention to the screen than their peers in quiet (although unfamiliar) laboratory spaces.

An overlooked source of differences can be found on the participant’s screen itself. Although researchers have considered and attempted to account for differences in the size of the monitors (and the accompanying differences in size of the stimuli), volume, brightness, and so on, much less attention has been given to whether what appears on the participants’ screen is precisely what we intend to appear. In part, this is because for most online testing, only the infants’ webcam is recorded. On most platforms, it is impossible to view what is actually on the participant’s monitor. When researchers do monitor what is on the screen—for example, during moderated remote sessions—they use what is seen on their own screen as a guide, not a recording of what actually appeared on the participant’s screen. When we analyze looking time data, we assume that we know what was actually visible on the participant’s screen. However, because most platforms don’t record the participant’s screen, we are relying on our systems working as intended.

When we converted some of our in-person studies to an online format at the start of the pandemic, we used Gorilla, and wrote a custom script to record not only the webcam but also capture what was on the participants’ screen.* We were very surprised by the variation in what the participants saw on their screen. 

We collected videos from 97 children 12-48 months of age in 2020-2021.** These videos were collected on the Gorilla platform with a custom script that used the RecordRTC JavaScript package. Our task was approximately six minutes in length and combined trials with left/right stimuli and trials with top/bottom-left/bottom-right stimuli. 

Of the videos of the 97 sessions, only 19 displayed the stimuli entirely as intended throughout the session. An additional 40 displayed the stimuli as intended but had a visible mouse cursor during some or all of the trials. The remaining 38 videos had further deviations during the session. These deviations varied from a momentary distraction that affected a single trial to a deviation that was present throughout the session. As many infant researchers use Lookit for similar studies, we attempted to recreate these issues in the Lookit platform from the participant perspective.

We saw the following issues in the videos collected on Gorilla, arranged by frequency. We note below whether we were able to recreate these same issues on Lookit:

• Mouse cursor visible on screen (72 of 97 videos). This was usually a fairly minor distractor, left by the parent somewhere on the screen. Sometimes however, the child seemed to take control of the mouse, moving it about over the stimuli. This was the sole issue in 40 of 97 videos, and is an issue that could have been prevented through programming. Lookit offers the capability to hide the cursor and researchers who use that feature should not encounter this issue. 
• Missing Stimuli (9 of 97 videos). Several participants were missing stimuli (either a full trial or only a single side) for specific trials. This appeared to be a loading issue – as it followed a delay in loading a new trial in each case. It’s possible this was an issue with the platform or with participant’s internet connection.  
• Right click somewhere on the screen (7 of 97 videos). Occasionally the child managed to right click somewhere on the screen, bringing up the options menu and obscuring part of the screen. We asked caregivers to close their eyes during the task, so they did not always see if their child took the task in an unexpected direction. Unlike the mouse cursor, this is not easily hidden, and we were able to recreate the issue on Lookit. 
• Volume changing mid-task (6 of 97 videos). Some participants found the volume up and down buttons and either silenced or maximized the task volume. We were able to recreate this on Lookit.

Image 1 – (Top-Left) A mouse cursor is visible in the lower center of the screen. (Top-Right) The two stimuli images failed to load for this participant, leaving a mostly blank screen. (Bottom-Left) The right-click options menu is visible in the lower center of the screen. (Bottom-Right) The volume adjustment and media player window is visible in the upper left part of the screen.  

• “Now Sharing” pop-up visible (6 of 97 videos). Our method for capturing the participant screen created a pop-up at the bottom of the screen. We requested that participants click the “Hide” button and most obliged, but for 6 participants this pop-up was visible throughout the session, rendering the session unusable. This was specific to our screen capturing method.
• Opened either the Windows search bar or the Mac applications dock (5 of 97 videos). This happened when the child hit a single button on Windows or dragged the mouse down on Mac computers. We were able to recreate this on Lookit, though it may vary by operating system/version.
• Plugin-related image overlays (5 of 97 videos). A few participants appeared to have a plugin – for example, the Pinterest plugin – active. These plugins placed colorful icons on top of our stimuli. We were able to recreate this on Lookit.
• Navigated away from the page (5 of 97 videos). This happened in different ways, including what appeared to be hitting ‘Alt+Tab’ on Windows. In our study, navigating away from the page did not pause the task. Within Lookit, however, actions that broke the full screen view also paused the task. This meant that actions like opening a new tab or printing the page would pause the task but hitting Alt+Tab would not.

Image 2 – (Top-Left) A pop-up that indicates screen sharing is visible in the lower center of the screen. (Top-Right) The Windows start menu is visible, obscuring the left side of the screen. (Bottom-Left) The Mac applications dock is visible along the bottom of the screen. (Bottom-Right) Icons are visible in the upper-right and upper-left of the screen. We believe these are generated by the Pinterest Plugin. 

• Revealing the icon to exit a full screen view (4 of 97 videos). We required participants to maximize their view of the task. By dragging their mouse pointer to the top of the screen, some children revealed a large X icon that could exit the full screen view. We were able to recreate this on Lookit.
• We also saw a number of unique and interesting issues:
• One child attempted to print the experiment (after first checking their security settings)
• One child opened the developer console and briefly inspected the website code
• One child opened their system information, then attempted to activate ‘Sticky Keys’
• One session had involved an email notification appearing mid-session.
• One session was completed in ‘dark mode’ (presumably via some plugin).
• A mysterious black line flashed above the right side of one participant’s stimuli throughout the task (We think this may have been a giant text entry cursor). 
• Two sessions were completed on computers with persistent warnings overlayed on their screen – possibly indicating that their version of Windows had not been activated.
• One caregiver shared their email tab instead of the task. We believe the participant completed the task but are not sure precisely what they saw.

Image 3 – (Top-Left) A menu for printing the page covers the experiment entirely. (Top-Right) The developer console is open on the right side of the window, offsetting the stimuli. (Bottom-Left) The background color of our experiment has been changed from white to black, presumably by a plugin. (Bottom-Right) A system information dialogue covers the center of the screen. We have censored potentially identifiable information in these images.

While these errors paint a fun picture of young children’s haphazard and presumably unintentional keyboard tapping, they also challenge our assumptions about the validity of our looking time analyses. When a child looks to the right side of the screen, they might be looking to the stimuli on that side, or they might be attending to the mouse, the icons overlayed by a plugin, or an unexpected pop-up. 

We found some form of distracting screen element in the majority of participants, but there was a high level of variability for what proportion of the session that the distractor was present. Some distractors (such as the presence of a mouse pointer) tended to persist across all trials of a participant, while others (such as missing stimuli or an email notification) affected only a few trials. They also varied in terms of severity. Ultimately a mouse pointer left at the bottom of the screen is not a terribly distracting item (especially when compared to other distractions in the house) but missing stimuli, opening a new tab, or viewing an email browser instead of the stimuli are obviously exclusion-worthy events. 

There are good and clear ways to mitigate these issues (programmatically hiding the mouse cursor as done on Lookit, for example), but they will be challenging to overcome entirely. Many of the issues we saw were simply caused by the presence of a small child in front of a computer while their caregiver’s eyes were closed. The most common non-mouse related issue was missing stimuli – a serious issue without easy mitigation. So, while asking caregivers to close other applications, disable plugins, and move their desktop keyboards are good recommendations to improve data quality, they might be insufficient. Ultimately, many of these issues are part of the trade-offs involved in conducting unmoderated looking time studies. The challenge is how can researchers have more confidence about what is on the screen during testing. One solution is obviously the one we chose; modify the program used to collect data to include screen recording as well as webcam recording. This may not be possible for all users, however. To be clear, we do not believe that these issues are serious enough to consider abandoning all unsupervised remote testing. Rather, it is important that researchers understand this source of error in their data and consider what steps they can take to reduce the occurrence of unintended deviations in what children see during such procedures. 

Thanks to Dr. Marianella Casasola, Dr. Lisa Oakes, Dr. Vanessa LoBue, Annika Voss, Mary Simpson, and the Cornell Play and Learning Lab research assistants who gave input on this draft or were involved in collecting this data. The project that this post is based on was supported by NSF Award BCS-1823489, and we greatly appreciate their support in our research.

Footnotes:

* We had incorporated screen recording primarily as a way to ensure the timing of our stimuli was accurate and to facilitate aligning our coding of the webcam recording with the start of the trial. 

** A total of 120 participants participated, but we excluded the remaining participants either for demographic reasons (e.g., age) or because they lacked webcam recordings. We did not review their screen footage.

About the Author

by Joshua L. Schneider and Kelsey L. West

The theoretical framework of developmental cascades has taken a firm foothold among infancy researchers. Indeed, it was the theme of this year’s ICIS meeting with not one, but two presidential addresses (Lisa Oakes and Catherine Tamis-LeMonda) opening with mountain stream metaphors to describe cascades. The mountain stream metaphor offers a simple but powerful idea about how distinct domains build upon one another. The rocks, dirt, and foliage that comprise the mountain landscape guide the water as it flows downstream. Simultaneously, the movement of the water carves out and shapes the landscape (which naturally, feeds back into the movement itself). And, droplets of water can travel very different routes down the mountain but often arrive at similar destinations; one droplet may move gently along the edges of the stream while another travels down its rough turbulent middle.

As trainees, we have “grown up” in research labs that apply a cascades framework: as graduate students and post-docs working with Jana Iverson at the University of Pittsburgh and Karen Adolph and Cathie Tamis-LeMonda at New York University. Needless to say, we spend a lot of time thinking about how developmental events cast wide ripples (to keep with our water-related metaphors) and affect seemingly separate events. The cascades approach has become engrained into the way we do our science, mentor and teach our students, and even how we think about the connections between our professional and personal lives.

Cascades in research. Initiation into a cascades-focused lab was exciting, but also a bit daunting. Surely to study multiple domains of development, one needs expertise in multiple domains (and as new grad students, we didn’t have expertise in any domain yet)! During weekly meetings, Jana Iverson reminded us what Esther Thelen always told her—“Everything matters!” Studying everything all at once seemed insurmountable (and indeed, no single study can address every factor shaping infant learning). But our trepidation subsided as we discovered how a cascades perspective could capture the complex inter-relations among the behaviors we observed in infants and their caregivers. Specifically, we have documented how new motor skills, like learning to sit, crawl, cruise, or walk, prompt changes in language development (West & Iverson, 2017; West et al., 2019), social interactions (West & Iverson, 2021; West et al., 2022), input from caregivers (Schneider & Iverson, 2022; West et al., under review), and even the physical contexts of play (e.g., how infants and caregivers position their bodies relative to one another during play time; Chen et al., under review; Schneider et al., in prep; Schneider et al., 2022). In addition, the cascades framework allowed us to examine the biological and contextual factors that set two infants on very different paths to similar development endpoints (like the water droplet racing down the mountainside or an infant learning to walk; Schneider & Iverson, under review). To this end, we documented cascades in infants with neurotypical and neurodivergent developmental pathways; particularly in infants who later develop autism (Calabretta et al., in revision; West et al., under review). In the course of each study, we’ve discovered something new by looking beyond our own research silos. By the way, it turns out that you don’t need to be an expert in many domains, you just have to know an expert—our mentors and collaborators filled in the gaps in our knowledge.

Cascades in mentoring. We have been lucky to mentor an incredibly talented group of undergraduates in our labs. Through a cascades perspective, we’ve come to appreciate how the skills students develop in the lab build on themselves over time and ripple into other aspects of their lives. Take for example the “2-minute spiel”, an exercise we picked up from Karen Adolph. Students read a research article and then, sometimes with slight groans and nervous faces, deliver an elevator pitch summarizing the paper in a way that anyone could understand—they highlight only the essential information and leave out the minutia. With practice, students hone soft skills that extend beyond the lab. They learn to digest primary-source research articles, which strengthens their ability to evaluate claims they encounter in life (an especially useful skill in an era of “alternative facts”). Students also begin to see the forest, not just the trees, of individual studies. As a result, their own writing becomes more streamlined and focused-in on a broader argument. Finally, students learn to communicate about science in an accessible way, a skill that greatly benefits them during high-stakes interviews. And because students can clearly articulate their science to many people—not just experts in their niche—they elicit feedback from a broad audience with diverse expertise, adding breadth and maturity to their ideas (no cascade hypothesis is complete without a feedback loop, right?)

Cascades in everyday life. The idea that seemingly separate domains can interrelate also shows up in our attitudes about work-life balance. Sometimes in academia, young scientists can feel pressure to suffer for their art: to work long hours on little sleep, give up their weekends and evenings, and spend little time resting, socializing, or pursing hobbies. But like the water and the mountain landscape, personal and professional domains are reciprocally influential. Investing in non-research activities (like art, music, reading fiction, or hiking) has downstream benefits for our science. For instance, our shared backgrounds in art and design influence the way we visualize and present our data such that it is well-composed and clear. And in our writing, as in our artwork, we’ve learned to put the rough broad strokes down on paper first, and then fill in with the fine lines and detail later. In addition, we’ve both learned the hard way that sacrificing sleep just isn’t worth the consequences (some cascades deliver consequences). And finally, socializing is perhaps the most essential non-academic investment of our time. Cultivating a strong social support system (filled with good people, good food, and of course, cake and champagne) was a life-saver in our doctoral and post-doctoral journeys, and surely kept us going when the going got tough.

Seeing the world through “cascades-colored glasses” has shaped our outlook in many ways. We consider developmental cascades when we do our science, apply its principles when we mentor our students, and attempt to take its heed as we learn to balance work and life. And finally, as we all learned while traveling to and from the ICIS 2022 meeting in Ottawa, even planes can developmentally cascade—from delay to delay, that is.

References

1. West, K. L., & Iverson, J. M. (2017). Language learning is hands-on: Exploring links between infants’ object manipulation and verbal input. Cognitive Development, 43, 190–200. https://doi.org/10.1016/j.cogdev.2017.05.004
2. West, K. L., Leezenbaum, N. B., Northrup, J. B., & Iverson, J. M. (2019). The relation between walking and language in infant siblings of children with autism spectrum disorder. Child Development, 90(3), e356–e372. https://doi.org/10.1111/cdev.12980
3. West, K. L., & Iverson, J. M. (2021). Communication changes when infants begin to walk. Developmental Science, 24(5), e13102. https://doi.org/10.1111/desc.13102
4. West, K. L., Fletcher, K. K., Adolph, K. E., & Tamis-LeMonda, C. S. (2022). Mothers talk about infants’ actions: How verbs correspond to infants’ real-time behavior. Developmental Psychology, 58(3), 405–416. https://doi.org/10.1037/dev0001285
5. Schneider, J. L., & Iverson, J. M. (2022). Cascades in action: How the transition to walking shapes caregiver communication during everyday interactions. Developmental Psychology, 58(1), 1–16. https://doi.org/10.1037/dev0001280
6. West, K. L., Saleh, A. N., Adolph, K. E., & Tamis-LeMonda, C. S. (under review). “Go, go, go!” Mothers’ verbs align with infants’ locomotion. Developmental Science.
7. Chen, Q., Schneider, J. L., West., K. L., & Iverson, J. M. (under review). Locomotion shapes infant-adult proximity during everyday play. Infancy.
8. Schneider, J. L., Marty, O. C., & Iverson, J. M. (in prep). Everyday environments for infant locomotion: Cataloging the natural affordances for infant cruising.
9. Schneider, J. L., Roemer, E. J., Northrup, J. B., & Iverson, J. M. (2022). Dynamics of the dyad: How mothers and infants co-construct interaction spaces during object play. Developmental Science, 00, e13281. https://doi.org/10.1111/desc.13281
10. Schneider, J. L. & Iverson, J. M. (under review). Equifinality in infancy: The many paths to walking. Developmental Psychobiology.
11. Calabretta, B. T., Schneider, J. L., & Iverson, J. M. (in revision). Bidding on the go: Links between walking, social actions, and caregiver responses in infant siblings of children with autism spectrum disorder. Autism Research.
12. West, K. L., Steward, S. E., Roemer, E. J., & Iverson, J. M. (under review). Walking boosts communication for neurotypical infants, but not infants later diagnosed with ASD. Journal of Autism and Developmental Disorders.

About the Authors

by Hallie Garrison

Why get involved in outreach

Developmental science has the potential to positively impact the lives of children and families by informing policymakers and practitioners. However, researchers, policymakers, and practitioners tend to be entrenched in entirely separate communities with a lack of communication and trust, leading to limited uptake of new scientific evidence in policy and the criticism that researchers are out of touch with issues of practice. Bridging the gap between research and practice and research and the public will take large-scale cooperation between scientists, policymakers, and practitioners. Taking cost and time investment into account, is there anything that individual scientists and labs can offer to encourage engagement with research outside of academic circles?

In the Bergelson Lab at Duke University, we have spent the past few years focusing some of our energy beyond the walls of the ivory tower. For us, this takes two forms: writing about science for non-academic audiences and periodically visiting a local afterschool program to lead activities focused on communication science and sensory differences.

Much of our writing takes place on a blog, Babies and Language, where we try to make information about language learning interesting and accessible to a broad audience. We hope that the blog attracts parents, practitioners, and others who might be interested in language acquisition in typically-developing, blind, deaf, or hard-of-hearing babies and children. We also promote our blog using Twitter and TikTok (and our PI tweets, too, from a lab account where she makes nerdy jokes to connect(?)s with folks inside and outside academia about topics related to the lab’s research). Twitter offers an opportunity to reach teachers, therapists, advocates, and legislators. If networking efforts are focused locally, online relationships can also create fruitful arenas for recruitment and collaboration. And—as though you needed another excuse to spend time on Twitter—while scientists are mostly followed by other scientists, their following actually diversifies once they exceed 1000 followers to include non-academic audiences as well (Côté and Darling, 2018).

Tips and tricks of the trade

Committing to outreach goals is a big undertaking for any lab. There is always pressing research-related work, making it hard to find time for networking, relationship-building, and non-academic writing. What can labs do to prioritize interfacing with the public while maintaining their focus on the everyday work of science?

Delegate

Learning to communicate complex ideas to a general audience is a skill that is worth cultivating in trainees. Graduate students, post-docs, staff, and research assistants can help disseminate scientific information to the public while strengthening their communication skills and solidifying their own knowledge about their specialty area. Undergraduates are often particularly great at generating ideas for social media–in our lab, we asked undergraduates to translate our written blog posts into TikToks. To our great surprise, one research assistant’s TikTok even went semi-viral for instigating a comment debate about whether a small toy was a cow or a pig. 

Practicing

Communicating research to a layperson audience is useful to students and it can be a welcome creative break from scholarly work.

Expand Networks

Leaders in child care centers, pediatrician’s offices, or schools may be willing to engage in public-facing writing, including guest blog posts or newsletter articles about their work with children. When your guest writer shares a post to their network, you benefit from an expanded audience who you may not have otherwise reached. You might be able to return the favor with your own short piece of writing for a newsletter, blog, or by speaking at a community event or on a podcast. Community leaders can also introduce you to their colleagues, diversifying your contacts and providing opportunities for you to keep apprised of current issues in practice.

When looking to get involved in K-12 schools, it’s critical to remember that teachers and administrators are already overburdened with responsibilities. Look for opportunities where your team can get involved outside of precious classroom time, such as in afterschool programs, camps, and other enrichment programs. We also recommend networking with other groups who are conducting outreach in similar ways to you. For our lab, the University of Maryland’s Language Science Center provided valuable feedback on our activities before we ever set foot in a classroom, reducing the need for repeated piloting of our curriculum.

Sharpen the Skill

Increasing public understanding of science can help expand funding for research, but communicating with the public about science is notoriously difficult. Even journalists with experience in science communication can struggle with overstating a new finding or getting bogged down in details. Luckily, there are many resources to help researchers learn to communicate more effectively. Universities often have communications teams who may host workshops or webinars about writing for a public audience. There are also formal training programs for science communication (e.g. through the American Association for the Advancement of Science or COMPASS), and trainees especially should be encouraged to seek out these kinds of opportunities, as they are (unfortunately) not likely to receive non-academic writing instruction as a part of graduate school.

Potential impacts and how to evaluate

The National Science Foundation seeks to keep its grant recipients accountable to taxpayers through the broader impacts criterion. Scientists must justify how their work will benefit society, a requirement that can be fulfilled through the work itself or through adjacent, related projects that are delivered directly to the public. In designing broader impacts proposals, scientists must think about how they will carry out the project, from how they will reach the intended audience, to how they will evaluate the project’s success.

One issue with trying to evaluate outreach programs is that each effort is highly unique.  One scientist might have expertise in robotics and see the possibility of sharing knowledge with high school students, while another scientist might be more interested in and well-suited to serving on the board of a community organization. Even if we narrow the scope to only look at outreach with school-aged children, it is difficult to compare across goals and outcomes. Is a program’s goal to increase students’ knowledge about a certain topic? To encourage positive feelings about science among students?

However, there’s value in collecting data to assess the impact of outreach, even informally. In fact, there’s an entire center and website devoted to collecting informal summaries and evaluations of STEM learning experiences (Center for Advancement of Informal Science Education (CAISE); informalscience.org). While we may not be well-positioned to determine if meeting a scientist while in second grade impacts whether that student chooses to become a scientist, we may be able to ask whether they enjoy science more after more exposure. Similarly, we may be able to track the potential for science communication efforts to reach a broad audience by keeping tabs on simple metrics like page hits, IP address locations, and examining the most popular posts.

It takes a village

The list of responsibilities for academics is ever-growing. It’s reasonable for outward-facing tasks to be shuffled to the bottom of the list, behind all of the research, administrative, and teaching-related work. No one person can take on the task of making sure that their research reaches a non-academic audience, that it is well-communicated, and that it inspires budding young scientists (or whatever your specific goals may be). It is critical to bring other members of your research community on board to design and implement outreach projects–whether that’s graduate students, collaborators, staff members, or friends. You’re likely to be met with a great deal of enthusiasm when you begin to explain your goals to others. After all, who doesn’t want their work to make an impact beyond their office walls?

This work was supported by the National Science Foundation CAREER Grant (BCS-1844710) to Elika Bergelson.

About the Author

by Richard Aslin and David Lewkowicz

Automated devices for recording where you are looking are so common (even your smartphone can do it) that we forget how it is done, the potential pitfalls when applied to infants, and how to interpret the massive amount of data that such devices can spit out over a short testing session.  Here we provide some history, principles of operation, an overview of options, and evidence that eye trackers can answer unique questions that older methods cannot (see comprehensive reviews by Oakes, 2012 and Hessels & Hooge, 2019).

There are three basic ways to record eye movements:  electrooculography (EOG), coils embedded in a contact lens and placed in a magnetic field, and video images of the eye along with reflections from its corneal surface.  EOG was employed first because it was easy to place two small electrodes on the face near the outer edge of each eye and it provided highly precise recordings of horizontal eye position with excellent temporal resolution.  One of us published an EOG study of 1- and 2-month-olds as his second-year project in grad school (Aslin & Salapatek, 1975) and many other EOG studies have documented the development of horizontal, vertical, and binocular eye movements in young infants. Unfortunately, for many applications (e.g., determining where on a face the infant is looking), EOG is not suitable unless you can hold the infant’s head in a fixed position (and we know how much infants like that constraint). 

We don’t need to dwell on the contact lens method because infants will not tolerate (1) a large diameter contact lens in their eye, (2) an embedded coil of fine wires that extends off the contact lens to transmit electrical signals induced by a large magnetic field, and (3) the fine wires rubbing against their cornea every time they blink.  Anesthetic eye-drops are tolerated by aspiring PhD students but not by infants (or their parents).

The most commonly used eye-tracking method is based on an old principle called corneal reflection photography.  If you hold a small light source in front of the eye, it creates a reflection that is (approximately) in the center of the pupil.  Since that light source is distracting, you can make it invisible by using infrared light and still detect the image of the eye and the corneal reflection with film or a video camera that is sensitive to light in the infrared region of the spectrum.  This method was perfected by Salapatek & Kessen (1966) using infrared-sensitive film and then by Haith (1969) using newly introduced infrared-sensitive video cameras.  The position of the corneal reflection with respect to the center of the pupil does not vary linearly as the eye moves, but this can be mapped quite nicely by a calibration routine in which the infant is presented with a small target in several locations in the stimulus display.  The main advantage of the corneal reflection method is that (within limits) it is not affected by head movements because the position of the eye is not measured with respect to the head, but rather with respect to the fixed location in space where the light creating the corneal reflection is located.

Once a commercial market emerged for eye-trackers, the “home brew” systems were quickly replaced by standard off-the-shelf instruments.  Initially, these instruments were very pricey (~$22,500 when one of us purchased his first system in 1977) and had trouble detecting small eye-movements due to relatively low-resolution video sensors.  Moreover, the field-of-view of the camera had to be just slightly larger than the diameter of the pupil which meant that even small head movements shifted the eye out of the field-of-view of the camera.  To address this problem, subsequent commercial systems introduced motor-driven cameras that could compensate automatically for small head movements.  Both of us had such a system (Applied Science Laboratories 504) in our respective labs in the 1980s and 1990s, but then switched to a more robust system, the Eye-Link 1000 (https://www.sr-research.com/eyelink-1000-plus/).

Crucially, as video sensor technology improved and as prices declined in the late 1990s, several eye-trackers, most notably Tobii (https://www.tobiipro.com/applications/scientific-research/), employed very high-resolution video sensors, thereby eliminating the need to move the camera to compensate for small head movements because the field-of-view of the camera was much larger (~ 6 inches instead of 0.6 inches).  The high cost of eye-trackers remains, although a $229 version for gamers called the Tobii-5 (https://gaming.tobii.com/product/eye-tracker-5/) is an intriguing option.  Nevertheless, the primary limitation of corneal-reflection eye-trackers is that once the head moves outside the camera’s field-of-view, no data are recorded.  Most studies circumvented this problem by presenting stimuli on a fixed video display and positioning infants so that they were unlikely to look anywhere except at the screen.  This, of course, is not how infants engage their attention in the real world.

The solution to this fixed-screen constraint was to create head-mounted eye-trackers in which the camera was fixed to the head using a miniature camera on a “stalk” that was attached to a headband.  A second camera also mounted on the headband was pointed outward in the direction the participant was facing.  This provided a view of the scene in front of the participant, and calibration data mapped the position of the eye with respect to locations in the scene.  The natural progression of miniaturization of components in the 2000’s eventually led to a head-mounted eye-tracker from Positive Science (https://www.positivescience.com/) that is small enough and light enough to be tolerated by young infants.  This has opened up data collection in natural contexts as infants engage in everyday activities, including crawling, walking, and reaching for objects.

Eye tracking is tailor-made for investigating the role of selective attention in perception and learning in infancy. For example, one of us has used it to investigate the emergence of lipreading in infancy and its link to speech and language development (Lewkowicz & Tift-Hansen, 2012). In these studies, infants watch videos of talking faces while we collect eye gaze data. Using specific areas of interest (AOIs) such as the face, eyes, and mouth we then export gaze measures such as first fixation, latency to first fixation, number of fixations, duration of individual fixations, total fixation, etc. for each AOI.

The main advantage of data generated by an eye-tracker is that they provide novel insights into selective attention processes underlying perception, learning, and memory that traditional looking time methods do not provide. Aggregate measures of total looking time at specific aspects of stimuli do not provide the temporal resolution to assess rapid changes in attention to objects referred to by language or to determine with precision when in a sequence of visual events the infant looks at stimuli or decides to terminate fixation. 

A cautionary note is that there is a tendency to employ a favored measure even when its use is not always justified (like the proverbial hammer that is used even when no nail is in need of pounding).  Beware the tendency to expect that more detailed data collected from large samples of infants will automatically reveal insights about underlying cognitive processes.  Data-driven approaches are only as good as the hypotheses they are intended to test.  Always keep in mind that you must link the eye-movement measures you collect with a theory about how attention, perception, language or learning develops.  

———————

References

Aslin, R. N., & Salapatek, P. (1975). Saccadic localization of visual targets by the very young human infant. Perception & Psychophysics, 17, 293-302. 

Haith, M. M. (1969). Infrared television recording and measurement of ocular behavior in the human infant. American Psychologist, 24, 279.

Hessels, R. S., & Hooge, I. T. (2019). Eye tracking in developmental cognitive neuroscience–The good, the bad and the ugly. Developmental Cognitive Neuroscience, 40, 100710.

Lewkowicz, D. J., & Hansen-Tift, A. (2012). Infants deploy selective attention to the mouth of   a talking face when learning speech. Proceedings of the National Academy of Sciences, 109, 1431-1436.

Oakes, L. M. (2012). Advances in eye tracking in infancy research. Infancy, 17, 1–8.

Salapatek, P., & Kessen, W. (1966). Visual scanning of triangles by the human newborn. Journal of Experimental Child Psychology, 3, 155-167.

About the Authors

by Audun Dahl

When I started graduate school, I knew little about the everyday life of infants. I had only the vaguest ideas about the daily joys and woes of a 12-month-old—even one who lived a block from our infant research lab in Berkeley, CA. The laboratory experiments I read about in journal articles didn’t teach me about the lives of infants; they brought infants to researchers’ labs, not researchers to infants’ homes. With the encouragement of my advisor, this predicament led me to do naturalistic research: videotaping infants in their home environments. We wanted to study infants doing whatever they would have been doing if we weren’t there.

To see the need for naturalistic research, let’s imagine that we discovered a new species of sociable aliens on Mars. As scientists, we wanted to understand the new species, but we faced a dilemma. The aliens were perfectly willing to come along back to Earth and participate in laboratory research here. Yet, by bringing the Martians to our planet, we would learn nothing about how they developed under the unique conditions on Mars. How did they learn to walk on their red planet, where surface gravity is less than half of gravity on Earth? How did they learn to communicate with others in their Martian households?

The solution to this methodological dilemma seems clear: Do both. We would do laboratory research on Earth and field research on Mars. This is the approach animal researchers take when they combine field and laboratory research to understand earthly species. The laboratory studies would examine causal processes of perceptions, emotions, and actions using carefully designed experiments; the naturalistic observations would document everyday experiences and activities. Together, these two methodologies would build a science of how Martians developed.

But the need for naturalistic research has not been as widely recognized in psychological research on infants and children—nor adults, for that matter. By reading textbooks and taking classes on research methods, psychology students can come away believing that experiments, not naturalistic observations, are the “gold standard” of research. Laboratory experiments with random assignment control the environments of research participants, rule out confounds, and—it is implied—offer the safest road to knowledge about infants.

Several factors tempt us to skip naturalistic methods in research on human infants. First, it’s easy to think that we already know enough about what infants do and experience. After all, we, too, were infants once. Many of us have observed infants in our homes, either infant siblings, infant nieces and nephews, the infants of our friends, or our own infants. A second reason is practical. Naturalistic research is time-consuming and typically generates hundreds or thousands of hours of unstructured recordings. In some of our naturalistic studies, we are looking for behaviors—such as hitting others—that only happen once or twice per hour of normal interactions. We wouldn’t want to do naturalistic research unless we had to.

The problem with skipping naturalistic methods is that our intuitions about everyday life are often wrong, or at least idiosyncratic. When we use our own experiences to generate intuitions about everyday life, we are essentially doing unsystematic case studies. This is a sensible way to

start a research program, but usually not a good way to complete it. Case studies have a venerable history in developmental psychology—carried out by founding figures like Darwin, Freud, and Piaget—yet every psychology major has heard of their shortcomings. One infant differs from the next. One person’s recollection differs from another’s, and both may differ from what actually happened. I have seen the same claim about infants’ everyday experiences be described as obviously true by one reviewer and obviously false by another. Intuitions about everyday experiences are about as disparate as human life is diverse; insofar as researchers come from privileged backgrounds, their intuitions will likely reflect the biases and inequities associated with those backgrounds. To separate the right intuitions about everyday lives from the wrong ones, we need data. Naturalistic methods offer a way of getting those data.

Consider research on infant helping. In the past two decades, developmental scientists have debated why infants begin to help others. In 2006, Science published a groundbreaking paper by Felix Warneken and Michael Tomasello. They found that most 18-month-olds helped an unfamiliar adult, for instance by handing back a dropped pen or paperclip to the adult. One key question was whether infants developed the ability to help others without any encouragement, praise, or other kinds of support from caregivers. One possibility was that infants began to help before their caregivers encouraged helping. Another possibility was that caregivers actively encouraged helping from around the age when infants began to help. This fundamental question about the origins of human prosociality cannot be settled by laboratory experiments as it concerns what happens in children’s daily lives. These lives—needless to say—do not take place in laboratories. To know whether or how caregivers encourage infants to help, we conducted naturalistic observations and parent interviews with families around the San Francisco Bay Area in Northern California. It turned out that, already by the first birthday, the parents in our studies frequently encouraged their infants to help by putting toys away after a play session, cleaning a surface, or even watering plants. In most of the situations in which infants helped, they received encouragement, praise, thanking, or some combination thereof from their caregivers. (Note that neither naturalistic nor experimental methods are sufficient by themselves: In addition to the naturalistic evidence on caregiver encouragement, we needed experimental evidence to know that adult encouragement actually had a causal effect, increasing infants’ helping.) Scholarly debates about the origins of infants’ helping offer just one illustration of how key questions about infant development call for naturalistic data about infants’ everyday lives. Calls for naturalistic methods have a history. Decades ago, Urie Bronfenbrenner, Alfred Baldwin, and others argued for the importance of studying children’s everyday social contexts. Around 1980, perhaps reflecting the influence of these scholars, 2-3 percent of articles in major developmental psychology articles contained the word “naturalistic” in their title, abstract, keywords, or other descriptors. Any effects of these admonitions seem to have waned, however. In the period from 2007 to 2017, the percentage of “naturalistic” articles had dropped below one percent.

Today, the future looks brighter for naturalistic methods. Technological advances have made naturalistic research both richer and less time-consuming. Automated recording devices for

speech, portable head-mounted cameras, and automated video processing all make it more manageable to collect and analyze large amounts of unstructured audio and video data. Thanks to more portable recording technology, such as the LENA devices for recording language input and children’s vocalizations, researchers collect naturalistic data from infants’ homes without even being there. By leveraging technological advances to record and analyze rich observational data, developmental scientists have made major contributions to our understanding of infants’ visual attention, word learning, and movement patterns.

Early in the 20th century, many countries used a gold standard to regulate their money. In the United States, the gold standard meant that a dollar could be converted into a fixed amount of gold. The gold standard was eventually abandoned, but not until it had hamstrung the central banks, exacerbating the Great Depression. Despite the demise of the actual gold standard, the metaphorical gold standard has remained a term of praise for experimental methods. But the field of developmental science may no longer need a methodological gold standard—if it ever did. Instead, a developmental science that judiciously combines experimental, naturalistic, and other tools can capitalize on the strengths of each.

Still, naturalistic methods hold a unique position in our toolbox, whether we study infants or Martians. Developmental research questions are nearly always about what happens in children’s everyday lives. How do infants learn language? How does locomotion affect cognitive development? Why do infants begin to hit and help others more from the first to the second year? Infants do not typically acquire language, begin to crawl, or change their tendencies to hit or help others inside the confines of a laboratory. Rather, these processes mainly occur in the homes, playgrounds, daycare centers, and other places where infants spend their days. When we bring infants into the lab, we wish to shed light on developmental processes that typically occur outside the laboratory. When we study infants’ everyday lives in their homes, we directly observe the very phenomena we ultimately seek to understand.

About the Author

by Andrea Sander-Montant and Krista Byers-Heinlein

Illustration by Andrea Sander-Montant

In Frankenstein, written by Mary Shelly in 1818, the doctor described his experiment like this: “I collected the instruments of life around me, that I might infuse a spark of being into the lifeless thing that lay at my feet. It was already one in the morning; the rain pattered dismally against the panes, and my candle was nearly burnt out, when by the glimmer of the half-extinguished light, I saw the dull yellow eye of the creature open; it breathed hard, and a convulsive motion agitated its limbs”. 

The story of Frankenstein exemplifies the romantic myth of the lone eccentric scientist working on matters unknowable to the rest of us. In a sense this myth, as countless others, stems from a real phenomenon of science being a mystical entity behind an impenetrable castle. In the past, the way we conducted science, and eventually communicated the results, was mainly aimed at other scientists and experts, and as a result, knowledge got shrouded by things like technical jargon and pay-per-view access. However, recent efforts like the open science movement have spearheaded a path towards making science more accessible to a broad audience (Sanjana, 2021).

The birth of open science

The idea behind open science started forming in the early 2000’s and in January 2012, the term “open science” was featured for the first time in a New York Times article (Lin, 2012). The term referred to a movement that came as a response to the “reproducibility crisis” in psychology, which psychology researchers were increasingly aware of, but which, of course, also existed in most other disciplines (Open Science Collaboration, 2015). The reproducibility crisis refers to the phenomenon whereby researchers are too often unable to reproduce findings reported in previously published research. For instance, Ionnidis (2005) estimated that more than half of published research may be false. The core principles of the open science movement relate to transparency, accessibility/openness, and collaboration as ways to make science more reproducible and within reach for all interested parties. 

As the way we do science changes to become more collaborative and open, the way science is communicated must change as well. Historically, science communication has been left up to two main groups: scientists with varying degrees of communication skills and journalists with varying degrees of scientific skills. But what if science communication could also be collaborative? Here we provide two real-life examples from the field of developmental psychology.

Example 1: Many babies, one collaborative press release

The ManyBabies Consortium came together in the context of previous and concurrent collective efforts to replicate research in the social and cognitive fields of psychology, such as ManyLabs, ManyPrimates, and the Psychological Science Accelerator. These international networks pool their resources, expertise, and ultimately, their data, to surmount the main challenges facing replicability, such as getting a sufficiently large dataset, testing participants from diverse backgrounds, and standardizing research practices across laboratories. 

In the ManyBabies 1 project, researchers tested thousands of babies from around the world to investigate infants’ preference for infant-directed speech (i.e., baby talk), and the team jointly published a peer reviewed journal article (ManyBabies Consortium, 2020). Next, the researchers were faced with another challenge: communicating their findings with the rest of the world. 

For the ManyBabies 1 project, the two lead authors took a traditional approach to publicizing their findings, with each author coordinating with their own university’s media relations team to draft a press release, which was then disseminated via standard channels like newspapers  and tv stations. ManyBabies 1 Bilingual, a spinoff from the main project that looked at bilingual infants’ preference for infant-directed speech, decided to experiment with a more open and collaborative approach to sharing their findings. The lead author worked together with the media relations team at her university to draft a press release, and 11 collaborators volunteered to act as liaisons to their own institutions’ media/press offices. On the designated day, the press release went out simultaneously across multiple countries and languages. In the end, the findings were shared widely across English, Spanish, French, and German language media, making the findings accessible to a wider audience. 

This collaborative approach extended the open science spirit of the project itself to the way it was disseminated. Media interviews were not only done with the lead author, but also with many other collaborators who had individual perspectives on their findings and how these relate to their own cultural milieu. 

Example 2: Communicating the big pictures through a collage of collaborators

In communicating developmental science, a key goal should be for caregivers, educators and policy-makers to be able to access and utilize evidence-based  information to advance societal goals. However, more often than not, the only accessible information are second- or third-hand interpretations of research findings. One of the issues that perpetuates this is hyperspecialization, whereby researchers become experts in increasingly narrow topics.

In 2020, a group of developmental scientists decided to tackle this problem by building on some of the lessons learned from open science to create more impactful science communication. The group created Kotoboo comics, a freely available platform where fun comics are used to introduce lay audiences to evidence-based tidbits of developmental science. What Kotoboo’s science communication model embraces from open science is the principle of collaboration. Each scientist brings a piece of expertise to what later becomes a sort of joyful big-picture science potluck. Secondly, the model abides by the open science mandate of “openness”, offering its audience the opportunity to freely access research that would be otherwise walled-off by jargon or by pay-per-view journals. It also makes science accessible by using dissemination channels that are more familiar to lay audiences, like posting on social media.  For example, the image below is from a comic that explains how speaking to babies helps them learn language, as the rhythm in caregivers’ speech provides clues on how to segment words. In this example, Kotoboo’s expert on prosody took the lead, but the rest of the team collaborated to offer insights from the fields of infant-directed speech, bilingualism, and semantics to make sure the message was about the bigger picture language learning and not the narrow field prosody.

Image by Clara Daoud. Reproduced from Kotoboo Comics with the author’s permission

Take-home messages

Making science accessible through effective communication is a task of monstrous proportion that would have been challenging even to Dr. Frankenstein. However, here we have offered two real-life examples from developmental science where scientists have successfully infused a spark of being into their science communication by applying the open science principles of openness and collaboration. In the future, as science moves further away from the impenetrable castle, the opportunities for openness and collaboration will become increasingly common. We hope that as developmental scientists, we take these opportunities to grow closer to one another as colleagues and to our audience!

About the Authors

by Kristin A. Buss & Frances M. Lobo

We have all faced the feedback that our samples may not be representative, either for our papers or our grant submissions. Gone are the days where we could publish without critiquing our convenience samples, which for most of us meant largely white, middle-class families that surrounded our college towns. You may already have read Lisa Oakes’ blog post (“Representing babies in science: How we describe our samples is important”), where she discusses the convenience sample issue and how many of us, especially those of us trained back in the 90’s (or earlier) didn’t think we needed to worry about diversity in our studies or give any thought to demographics. Lisa discusses several important points (calls for action), and our post addresses some of the issue by laying out a way to increase recruitment of hard-to-reach families.

This is an exciting time in our field. There is a push for approaches that are anti-racist, that include families of all backgrounds, that pay serious attention to BIPOC and LGBTQ families, that consider neighborhood and community factors. The energy around this is palpable and long overdue. If you are like many researchers, you’ve been having difficulty recruiting diverse samples. Kristin has been working on this issue ever since her arrival at Penn State in 2006. In this blog post, we share some lessons we have learned and best practices to ensure you capture a wide range of variability in the families that you recruit.

We have done this work by building a shared community-university partnership research initiative at Penn State called Parents and Children Together [https://pact.la.psu.edu]. In building these connections, we learned a great deal about the barriers to recruiting (e.g., mistrust of research and institutions, deficit-based approach seen as critical of families of color). We’ve learned that you need a more relationship-building, one-on-one approach. The solution is simple, but takes a lot of humility, effort and time. Gone are the days of researchers having all the answers and bringing in families to confirm their ideas. Communities – parents, grandparents, guardians, community leaders, teachers, pastors – have a lot to teach us about our science and we just have to listen. Engage the families that you are interested in, and the communities in which they live, and become partners with them in learning about infant development. Make sure these families and communities know they are your partners; and make the effort to learn about who they are, where they live, and why infant development matters to them for all research projects.

These are the steps we have taken to build these relationships, improve our recruitment methods, and in turn, our science.

1. Build institutional support and work together with other investigators.

Building relationships with families and community partners, especially in areas where investigators may not live, takes time and effort. We created a task force of investigators who were willing to do the work to build relationships in the community and work in partnership to meet research goals. Additionally, over the last decade, having institutional support has given investigators resources to be able to build partnerships with community agencies, forge a community advisory board, and utilize a research site off-campus to meet with families in their own community.

2. Become informed about the participants in your project and the community in which they live.

Historically, communities have been used by research institutions. We believed it was important to learn from community partners what research questions were important to them, and to disseminate findings and give back to them. We wanted to build trust through this community engagement.

Unfortunately, we didn’t live in the community from which we hoped to recruit families.
For this reason, we needed a community liaison who could teach us about the community and help us form and maintain partnerships with agencies there that were working with families. Carmen Henry-Harris stepped into this role and is compensated for her time and expertise. She lives in Harrisburg, and community members know who she is and trust her as a gate-keeper of the community.

Additionally, the community agencies work with families regularly and are attuned to their needs. We began asking them what they were seeing as areas of concern for families, and where they would appreciate some help or some answers. Through these conversations, we continue learning about families and the community and we allow this to inform our science. For example, in a conversation with Capital Area Head Start teachers, we learned that teachers were concerned by the higher levels of stress and anxiety they were seeing in their preschoolers. In partnership with them, we conducted a pilot study with families to better inform how we may intervene with families to improve children’s outcomes as they transition to school.

3. Get feedback from community partners and other investigators about your project and the design and method you are proposing for it.

Study paradigms and the instruments we use to collect data have often been validated using convenience samples. Receiving feedback from others on all aspects of the project is important. For example, when Kristin first presented her risk room paradigm used to measure children’s dysregulated fear to PACT, community agencies rightfully questioned why all of the objects that were meant to scare children were black in color. This was a protocol that was validated with White families without consideration of how more marginalized groups might respond to the materials. Although other investigators pushed back on Kristin’s ideas to change the materials, ultimately there were no differences observed after she did so.

It is so easy to just listen to the community, and it has a meaningful impact. We have designed new instruments, changed protocols, adjusted recruitment strategies, changed the language we use, and continued to strengthen our research designs based on community input, even for projects that weren’t originally designed as community-engaged projects.

4. Ask the community for help with recruitment.

Community agencies that hear about our research projects have often volunteered to share our recruitment materials with the families they serve or have invited us to speak with parents directly at activity nights.

5. Give back to the community.

Our community agencies are partners in the research process. For this reason, we engage in community outreach to give back and speak with families about temperament, emotions, and the development of anxiety. We make sure to disseminate our findings to them and listen to their feedback about how to interpret the results. Additionally, given that grant budgets can be tight, using institutional support and collaborations helps us pool our efforts and help all the projects simultaneously. For example, all PACT researchers are expected to share responsibility to attend community events, and contribute to materials (e.g., handouts summarizing all projects that are shared in the community). Although these efforts take time, they are all low-cost ways to engage the community and build relationships.

We hope these steps offer you some concrete steps that you can take that may help with recruitment.
We would be remiss if we didn’t mention the work of the many amazing BIPOC scholars who are leading the way in understanding the interplay between families and their communities (Mayra Bamaca, Gabbi Livas-Stein, Deborah Rivas-Drake, Dawn Witherspoon, to name a few); and the broader literature on community engagement and community-based participatory research that has guided our approach.

The main takeaway is to get to know the communities you are working in. Rather than assuming we are the experts on infants and their families, we need to realize they are the experts and that we must learn from them.

About the Authors