Rachel R. Romeo, Joshua Segaran, Julia A. Leonard, Sydney T. Robinson, Allyson P. Mackey, Anastasia Yendiki, Meredith L. Rowe, and John D. E. Gabrieli. 3/24/2018. “Neural correlates of the "30 million word gap": Children's language exposure is related to white matter structure.” In Cognitive Neuroscience Society. Boston, MA, USA.Abstract
Behavioral research has shown that the quantity and quality of young children’s language input predicts their later linguistic ability, and that children from lower socioeconomic status (SES) backgrounds receive less language exposure than their higher SES peers, which translates into a measurable gap in children’s language skills. The present study investigated which structural neural mechanisms underlie this input-output relationship. Forty SES-diverse children aged 4-6 years completed verbal and nonverbal standardized assessments, followed by a diffusion-weighted imaging (DTI) scan. Families then completed two full weekend days of real-world audio recordings from the child’s perspective, from which three measures were derived: the number of words spoken by any adult, the number of child utterances, and the number of conversational turns between the child and any adult. Behaviorally, SES was strongly correlated with both language exposure and verbal and nonverbal scores. However, when SES was partialled out, only conversational turns predicted additional variance in children’s verbal scores, even when additionally controlling for adult words and child utterances. Furthermore, the number of conversational turns, independent of SES, was positively correlated with the fractional anisotropy (FA) of the left arcuate fasciculus, which connects “Broca’s” and “Wernicke’s” area. This suggests that some qualitative aspect of dialogic communication has a greater impact children’s brain and behavior than the sheer volume of adult speech. To our knowledge, this is the first evidence directly linking children’s language environments with a brain structure known to underlie language development, and which may in turn contribute to the SES language gap.
Rachel R. Romeo, Julia A. Leonard, Sydney T. Robinson, Martin R. West, Allyson P. Mackey, Meredith L. Rowe, and John D. E. Gabrieli. 2/2018. “Beyond the “30 million word gap:” Children’s conversational exposure is associated with language-related brain function.” Psychological Science. Publisher's VersionAbstract

Children’s early language exposure impacts their later linguistic skills, cognitive abilities, and academic achievement, and large disparities in language exposure are associated with family socioeconomic status (SES). However, there is little evidence about the neural mechanism(s) underlying the relation between language experience and linguistic/cognitive development. Here, language experience was measured from home audio recordings of 36 SES-diverse 4-6 year-old children. During a story-listening fMRI task, children who had experienced more conversational turns with adults—independent of SES, IQ, and adult/child utterances alone—exhibited greater left inferior frontal (Broca’s area) activation, which significantly explained the relation between children’s language exposure and verbal skill. This is the first evidence directly relating children’s language environments with neural language processing, specifying both environmental and neural mechanisms underlying SES disparities in children’s language skills. Furthermore, results suggest that conversational experience impacts neural language processing over and above SES and/or the sheer quantity of words heard.

Rachel R. Romeo, Julia A. Leonard, Sydney T. Robinson, Allyson P. Mackey, Meredith L. Rowe, and John D. E. Gabrieli. 10/6/2017. “ Linking Language Exposure to the Brain: Structural and functional neural mechanisms underlying the relationship between children’s language exposure and their linguistic abilities.” In Many Paths to Language. Nijmegen, The Netherlands.
Rachel R Romeo, Joanna A Christodoulou, Kelly K Halverson, Jack Murtagh, Abigail B Cyr, Carly Schimmel, Patricia Chang, Pamela E Hook, and John DE Gabrieli. 6/2017. “Socioeconomic status and reading disability: Neuroanatomy and plasticity in response to intervention.” Cerebral Cortex. Publisher's VersionAbstract

Although reading disability (RD) and socioeconomic status (SES) are independently associated with variation in reading ability and brain structure/function, the joint influence of SES and RD on neuroanatomy and/or response to intervention is unknown. In total, 65 children with RD (ages 6–9) with diverse SES were assigned to an intensive, 6-week summer reading intervention (n = 40) or to a waiting-list control group (n = 25). Before and after, all children completed standardized reading assessments and magnetic resonance imaging to measure cortical thickness. At baseline, higher SES correlated with greater vocabulary and greater cortical thickness in bilateral perisylvian and supramarginal regions—especially in left pars opercularis. Within the intervention group, lower SES was associated with both greater reading improvement and greater cortical thickening across broad, bilateral occipitotemporal and temporoparietal regions following the intervention. Additionally, treatment responders (n = 20), compared with treatment nonresponders (n = 19), exhibited significantly greater cortical thickening within similar regions. The waiting control and nonresponder groups exhibited developmentally typical, nonsignificant cortical thinning during this time period. These findings indicate that effective summer reading intervention is coupled with cortical growth, and is especially beneficial for children with RD who come from lower-SES home environments.

Julia A Leonard, Rachel R Romeo, Sydney T Robinson, Allyson P Mackey, and John DE Gabrieli. 2017. “Can we improve children’s learning through interventions? Evidence from a parenting intervention, exploratory play, and social learning about effort.” In Society for Research in Child Development. Austin, TX, USA.
Rachel R Romeo, Julia A Leonard, Sydney T Robinson, Meredith L Rowe, Allyson P Mackey, and John DE Gabrieli. 2017. “Children’s language exposure predicts neural structure and function during language processing, independent of SES.” In Society for Research in Child Development. Austin, TX, USA.
Megumi E Takada, Julia A Leonard, Rachel R Romeo, Sydney T Robinson, Allyson P Mackey, and John DE Gabrieli. 2017. “Cognitive and neural correlates of mathematical reasoning across math proficiency levels.” In Society for Research in Child Development. Austin, TX, USA.
Joanna A Christodoulou, Rachel R Romeo, Kelly K Halverson, Abigail B Cyr, Jack Murtagh, Patricia Chang, Allyson P Mackey, Pamela E Hook, and John DE Gabrieli. 2017. “Individual differences in intervention response: Socioeconomic status and reading disability as predictors.” In Association for Psychological Science. Boston, MA, USA.
Rachel R Romeo, Julia A Leonard, Sydney Robinson, Meredith L Rowe, Allyson P Mackey, and John DE Gabrieli. 2016. “Children’s language exposure predicts neural activation during language processing.” In Society for Neuroscience. San Diego, CA, USA.Abstract

Research provides increasing evidence that a child’s experience early in life shapes their neuroanatomy and corresponding cognitive functions. One critical experience is language exposure; the quantity and quality of the language children hear during these early years impacts their overt language abilities throughout childhood and into adulthood. However, there is little evidence that directly relates children’s language experience to specific aspects of the underlying brain function. Here, we examined how variation in spoken language experience at home is related to variation in brain functions associated with spoken language comprehension. A socioeconomically-diverse group of families with typically-developing, native English-speaking children in pre-Kindergarten and Kindergarten (ages 4 to 6 years) were recruited. Both children and their primary caregiver completed behavioral assessments of their verbal and non-verbal cognitive abilities. Children then completed a functional magnetic resonance imaging (fMRI) session, including a task involving listening to simple stories and the same stories played backwards. Finally, families completed two full days of real-world audio recordings from the child’s perspective (via a child-worn audio recorder), from which two measures were calculated: the number of words spoken by an adult, and the number of conversational turns between the child and an adult. While the number of adult words is a purely quantitative measure, ‘conversational turns' is both quantitative and qualitative, because it implies a balanced, time-locked discourse. By contrasting activation for typical speech to incomprehensible backwards speech, we identified brain regions related to higher-level language comprehension, i.e., the semantic and syntactic processing necessary to understand the stories. This contrast yielded a significant positive correlation between the peak number of hourly conversational turns at home and activation in Broca’s area, formed by left pars triangularis and pars opercularis regions. This relationship held when controlling for measures of socioeconomic status (SES). Behaviorally, conversational turns also correlated with children’s scores on a standardized language assessment. Similar but statistically insignificant relations were found with simple number of words spoken by an adult at home. This suggests that language exposure is not only related to children’s behavioral verbal abilities, but also to the brain function underlying language processing; furthermore, it suggests that conversational quality has more impact than sheer numbers of words.

Outi Tuomainen, Valerie Hazan, and Rachel R Romeo. 2016. “Do talkers produce less dispersed phoneme categories in a clear speaking style?” Journal of the Acoustical Society of America, 140, 4, Pp. EL320-EL326. Publisher's VersionAbstract

This study investigated whether adaptations made in clear speaking styles result in more discriminable phonetic categories than in a casual style. Multiple iterations of keywords with word-initial /s/-/ʃ/ were obtained from 40 adults in casual and clear speech via picture description. For centroids, cross-category distance increased in clear speech but with no change in within-category dispersion and no effect on discriminability. However, talkers produced fewer tokens with centroids in the ambiguous region for the /s/-/ʃ/ distinction. These results suggest that, whereas interlocutor feedback regarding communicative success may promote greater segmental adaptations, it is not necessary for some adaptation to occur.

Rachel R Romeo, Joanna A Christodoulou, Abigail Cyr, Kelly K Halverson, Jack Murtagh, Patricia Chang, Pamela E Hook, and John DE Gabrieli. 2015. “Children’s socioeconomic status influences their response to reading intervention.” In American Speech-Language Hearing Association. Denver, CO, USA.Abstract

We explored the relationship between socioeconomic status (SES) and language/cognition in children with reading disability (ages 6-9). Despite lower-SES children initially exhibiting significantly lower non-verbal, oral language, and reading scores, they benefitted from an orthographic imagery-based intervention to a greater degree than higher-SES participants. Clinical implications will be discussed.

Rachel R Romeo, Joanna A Christodoulou, Abigail Cyr, Kelly Halverson, Jack Murtagh, Patricia Chang, Allyson P Mackey, Pamela E Hook, and John DE Gabrieli. 2015. “Impact of SES on brain and behavior in children with dyslexia receiving intervention.” In Society for the Scientific Study of Reading. Kona, HI, USA.Abstract

Purpose: Although socioeconomic status (SES) strongly predicts academic performance, cognitive ability, and brain structure and function, the impact of SES has yet to be explored in children with dyslexia. We asked how the brain structure of children with dyslexia varies with SES, and if there are SES-related differences in response to intervention in both brain and behavior outcomes.

Method: Children with dyslexia (ages 6-9) were randomly assigned to receive an intensive (at least 100 hours over 6 weeks) summer reading intervention (n=39) or to a waiting-list control group (n=24). Before and after intervention/waiting, children completed MRI and assessments of cognition, language, and reading. SES, based on parents’ reported education and occupation, was correlated with whole-brain cortical thickness and volume of anatomical regions of interest at pre and post. 

Results: Amongst pre-intervention students with dyslexia, lower SES was associated with thinner cortex and a volumetrically smaller left inferior frontal region (Broca’s area). Furthermore, SES explained a significant percentage of variance in Broca’s volume (p<0.01), whereas no behavioral scores contributed significantly to these individual differences. Amongst intervention participants, more low-SES children exhibited improvement than their higher-SES peers—even when controlling for initially lower scores. Longitudinal thickness and volume gains corroborated this result. 

Conclusion: Despite lower-SES children with dyslexia having significantly lower reading scores and smaller associated brain regions than their higher-SES peers, they benefit from intervention—behaviorally and neurally—to a greater degree. This suggests that intensive reading intervention might be even more effective for these dually at-risk children.

Rachel R Romeo and Daniel Swingley. 2015. “Word recognition, phonological specificity, and SES: a longitudinal word-recognition study of toddlers.” In Society for Research in Child Development. Philadelphia, PA, USA.Abstract

Primary caregiver education is positively correlated with the quantity and quality of children’s linguistic input (Hart & Risley, 1995; Rowe, 2012), which in turn correlates with their phonetic discrimination (Author, this symposium). Furthermore, children’s socioeconomic status (SES)—to which parent education largely contributes—correlates with their online word-level processing and vocabulary by 18 months of age (Fernald et al., 2013). To what extent might variation in word recognition relate to phonological representation of familiar words? In general toddlers seem to have phonetically well-specified lexical representations (e.g., Swingley & Aslin, 2002), yet it remains unclear what factors affect development. Our goal was to define the trajectory with which children develop mature phonetic representations, as assessed longitudinally using a language-guided looking task. We hypothesized also that higher-SES children might develop this specificity earlier. Monolingual English-learning children participated at 14.5 months (n=34), 17.5 months (n=30), 21 months (n=28), and 24.5 months (n=25). Parents reported their highest education level attained, presented here as two groups: “higher-SES” (mother completed college or higher, n=21) and “lower-SES” (less than college degree, n=13). Children viewed pairs of pictures of familiar objects—words that the children reportedly knew—and one object was named, (e.g., “Where is the baby?). Each of 16 objects was named twice—once pronounced correctly and once “mispronounced” by varying the voicing or place of articulation of the initial consonant (e.g., baby became paby). If mispronunciation hinders recognition, this evidences well-specified knowledge of the word’s correct phonetic form (Swingley & Aslin, 2000). Word recognition accuracy was computed as the percentage of time participants fixated on the target picture within 0.36-2 seconds after target word onset. Speech processing speed/efficiency was computed as the mean time after target onset at which children shifted their gaze from distractor to target. The difference in both measures between pronunciation conditions represented the child’s mispronunciation sensitivity (MPS). Unsurprisingly, children’s accuracy increased and speed decreased (both p<0.001) as they aged (Figure 1), and accuracy was higher for correctly-pronounced targets than mispronounced ones (p<0.001). However, a significant MPS did not appear until 17.5 months (p<0.001). Higher-SES children averaged 8% greater accuracy than lower-SES children (Figure 2) across all ages and conditions (p<0.05), recognizing mispronounced words as accurately as their low-SES peers recognized correctly-pronounced words, and at marginally faster speeds (p=0.062). Furthermore, while high-SES children reliably recognized correctly-pronounced words (p< 0.001) and exhibited MPS (p<0.05) by 14.5 months, low-SES children did neither until 17.5 months (both p<0.05). From 17.5 months and onward, MPS was equivalent across the SES groups. Though we had predicted slower phonetic development in low-SES children, individual differences were primarily found at the word level, extending evidence of language processing differences down to 14 months. In each SES group, MPS was evident at the first age at which word recognition was reliable: 14 months for higher SES, 17.5 months for lower. Together, this suggests that poorer word recognition performance in lower-SES toddlers is not caused by insensitivity to phonological features, but is likely caused by other disparities in children’s linguistic environments.


Rachel R Romeo, Valerie Hazan, and Michèle Pettinato. 2013. “Developmental and gender-related trends of intra-talker variability in consonant production.” The Journal of the Acoustical Society of America, 134, 5, Pp. 3781-3792. Publisher's VersionAbstract

This study investigates the effect of age and gender on the internal structure, cross-category distance, and discriminability of phonemic categories for two contrasts varying in fricative place of articulation (/s/-/∫/) and stop voicing (/b/-/p/) in word-initial tokens spoken by adults and normally developing children aged 9–14 yr. Vast between- and within-talker variability was observed with 16% of speakers exhibiting some degree of overlap between phonemic categories—a possible contribution to the range of talker intelligibility found in the literature. Females of all ages produced farther and thus more discriminable categories than males, although gender-marking for fricative between-category distance did not emerge until approximately 11 yr of age. Children produced farther yet also much more dispersed categories than adults with increasing discriminability with age, such that by age 13, children's categories were no less discriminable than those of adults. However, children's ages did not predict category distance or dispersion, indicating that convergence on adult-like category structure must occur later in adolescence.

Valerie Hazan, Rachel R Romeo, and Michèle Pettinato. 2013. “The impact of variation in phoneme category structure on consonant intelligibility.” In Acoustical Society of America, 133: Pp. 3386. Montreal, QC, Canada. Publisher's VersionAbstract

 Newman et al. [J. Acoustic. Soc. Am. 109, 1181-1196 (2001)] suggested that phoneme identification accuracy and speed for a given talker was affected by the degree of variability in their production of phoneme categories. This study investigates how intra-talker variability in the production of two phoneme contrasts varies with age and gender, and how this variability affects speed of perceptual processing. Multiple iterations of tokens differing in initial consonants (/s/-/ʃ/, /p/-/b/) were collected via picture elicitation from 40 adults and 31 children aged 11 to 14; measures of within-category dispersion, between-category distance, overlap, and discriminability were obtained. While females produced more discriminable categories than males, children produced farther yet more dispersed-and thus similarly discriminable-categories than adults. Variability was contrast-specific rather than a general talker characteristic. Tokens with initial /s/-/ʃ/ from pairs of adult and child talkers varying in between-category distance or overlap were presented for identification. The presence of overlap had a greater effect on identification accuracy and speed than between-category distance, with strongest effects for adult speakers, but reaction time correlated most highly with within-category dispersion. These data suggest that talkers who are less consistent in their speech production may be perceived less clearly than more internally consistent talkers.