Examining the Relationship Between Early Feeding Difficulties and Language Development: A Longitudinal Study Authors: Simone Poulin, David H. McFarland, Stefano Rezzonico, Natacha Trudeau Published: 2026 ABSTRACT ABSTRACTIncreasing evidence suggests that interactive feeding is an important variable in language development, with early disruptions in feeding potentially having long‐term consequences. This longitudinal experiment characterized early patterns of feeding difficulties and language development in full‐term, typically‐developing infants across the developmental milestones of 8, 12, 18, 24, and 54 months. Parent‐directed assessments were used to assess feeding difficulties and language abilities between 8 and 24 months and a clinical language assessment was administered at 54 months. Results revealed that certain feeding difficulties, such as poor saliva control at 18 months, correlate negatively with expressive vocabulary size and may signal co‐occurring and later occurring less developmentally advanced language skills at 18 and 24 months. Others, such as food selectivity, are characteristic of children with less developmentally advanced language skills at 24 months. Together, these findings suggest that feeding difficulties may reduce opportunities for language stimulation and/or reflect underlying oral‐motor developmental vulnerabilities that also affect language development. Introduction One of the key goals in research on language/communicative disorders is to understand the foundational skills that scaffold speech and language development. Our own research has focused on interactional mother–infant physiological synchrony (e.g., McFarland et al. 2020) and the potential role of feeding as providing key moments of communicative interaction and language stimulation (Malas et al. 2015, 2017; McFarland, Poulin, et al. 2020). There are several reasons to consider the potential role of feeding as a key factor in building language skills. Caregiver–infant feeding is a crucial moment of social interaction (Feldman 2015; Kaye 1977) and language stimulation (Aukrust and Snow 1998; Leech 2021; Leech et al. 2018; Snow and Beals 2006). Verbal interactions during solid feeding provide particularly rich language input (Weizman and Snow 2001; Zimmerman et al. 2019), even compared to other interactive contexts such as toy play, often considered the gold standard language stimulation context. Maternal speech during mealtimes has greater lexical diversity (Zimmerman et al. 2019) and contains more rare words (Weizman and Snow 2001) than play. The higher frequency of rare words during mother–child mealtime interactions is correlated with children's later higher receptive vocabulary skills (Weizman and Snow 2001). Despite the potential of feeding as a setting for language enrichment, when feeding difficulties are present, this optimal context for language stimulation may be compromised, as disrupted mealtimes adversely impact caregiver–infant communicative interactions and consequently, early language development. For example, parent–child mealtime communicative exchanges tend to focus on corrective comments on feeding behaviors when feeding difficulties are present (Harding et al. 2013). This reduces children's exposure to the normally rich, diverse parental language input during mealtimes related to food characteristics and enjoyment in infants (Harding et al. 2013; Thompson et al. 2024) or abstract topics beyond the here‐and‐now in older children (Leech 2021; Leech et al. 2018; Snow and Beals 2006). Mother–child conflicts also increase in the presence of feeding difficulties, particularly those related to food selectivity (Lucarelli et al. 2003). Such conflicts may contribute to long‐lasting negative interactive cycles, in which a child's food selectivity and parents’ nonresponsive and coercive behaviors reinforce one another (Jansen et al. 2017; Mallan et al. 2018). Given that low parental responsiveness in parent–child interactions has been shown to be associated with lower child language outcomes (Levickis et al. 2023; Tamis‐Lemonda et al. 2001), food selectivity may interfere with typical language development by hindering important language‐learning opportunities tied to contingent, meaningful, and developmentally‐appropriate parental responses to children's communicative acts and behaviors during mealtimes. It is not surprising then, that early disruptions in feeding may negatively impact language development (Adams‐Chapman et al. 2013, 2015; Crapnell et al. 2015; Lainwala et al. 2020; Motion et al. 2001; Putnick et al. 2022; Somaraki et al. 2024; Wolthuis‐Stigter et al. 2017; Yardımcı‑Lokmanoğlu et al. 2024; Zimmerman and Rosner 2018). Prior history of feeding difficulties in infancy is associated with a range of later speech and language concerns in both preterm and full‐term children, including caregiver‐perceived speech‐language delays before Age 3 (Zimmerman and Rosner 2018) and limited expressive vocabulary (fewer than three words) as early as 18 months (Motion et al. 2001). These findings suggest that infants who experience early feeding difficulties are more likely to show less developmentally advanced language abilities over time. Population‐based cohort data reinforce this idea. In large cohorts of preterm and full‐term children, a greater number of feeding difficulties at 18, 24, and 36 months increased the odds of communication delays (likely including language delays), with the strength of the associations increasing with child age (Putnick et al. 2022), and late introduction of solids (after 10 months) predicted smaller vocabularies at 24 months (Somaraki et al. 2024). Evidence from preterm infants further underscores the prognostic value of early feeding behavior. Abnormal sucking patterns have been shown to predict poorer performance on standardized language tests through the toddler and preschool years (Wolthuis‐Stigter et al. 2017; Yardımcı‑Lokmanoğlu et al. 2024), including lower comprehension at Age 5 (Wolthuis‐Stigter et al. 2017). These associations extend to clinically significant outcomes, with early feeding difficulties increasing the likelihood of language delays by 18–26 months (Adams‐Chapman et al. 2013; Lainwala et al. 2020) and expressive and receptive vocabulary delays by 30 months (Adams‐Chapman et al. 2015). Retrospective clinical studies of children later diagnosed with developmental language disorders without co‐occurring neurodevelopmental disorders, motor deficits, or history of prematurity provide converging support. In these samples, roughly half of children (48%–53%) had a documented history of early feeding difficulties, most frequently characterized by food selectivity issues and challenges in food transition (Malas et al. 2015, 2017). Compared to the 20% general population estimates of Lindberg et al. (1991), children with developmental language disorders are 3.77 times more likely to have experienced early feeding difficulties (Malas et al. 2017), reinforcing the idea that such difficulties may have long‐term consequences for language development and may contribute to later language disorders. Additional evidence from neurodevelopmental conditions characterized by high rates of motor and communication difficulties, such as cerebral palsy, autism, and developmental language disorders, highlights a similar pattern of co‐occurrence. In cerebral palsy, lower expressive language abilities are predictive of more pronounced feeding dysfunction (Malandraki et al. 2022), while in autism the joint occurrence of feeding difficulties and speech‐language disorders exceed that expected from the independent prevalence of each (Neumeyer et al. 2019). Taken together, these converging findings across population‐based cohorts, preterm samples, clinical developmental language disorder groups, and neurodevelopmental conditions suggest that early feeding difficulties represent subtle developmental disruptions that may shape divergent language trajectories. Feeding difficulties may not only disrupt caregiver–infant interactions (Byrne et al. 2017; Jansen et al. 2017; Lindberg et al. 1996; Mallan et al. 2018), reducing opportunities for rich and responsive verbal input and joint engagement that support early language learning (Anderson et al. 2021; Fu and Zheng 2025; Madigan et al. 2019), they may also reflect underlying oral‐motor vulnerabilities that may constrain speech sound development (Nip et al. 2009), vocabulary (Kehoe et al. 2018), and grammar. Less mature oral movements such as slower or less coordinated movements of the tongue, the lip, or the mandible may impede accurate articulatory positioning (Green et al. 2002) and the acquisition of new speech sounds (Nip et al. 2009). Given that children tend to produce words containing sounds already in their phonetic repertoire (Schwartz and Leonard 1982; Stoel‐Gammon 1985), this may in turn restrict children's ability to learn new words (Kehoe et al. 2018). Early oral‐motor disruptions that affect vocabulary growth may in turn influence grammatical development, as vocabulary provides the foundational elements for grammar (Bates et al. 1988; McGregor et al. 2005). Parents also tend to provide less lexically diverse and syntactically complex contingent input in response to less developmentally advanced infant oral productions (Elmlinger et al. 2025). These oral‐motor‐based constraints may further negatively impact children's language development by disrupting the quality of linguistic input they receive. Consistent with the premise that oral‐motor vulnerabilities may constrain language development, research indicates clear associations between early oral‐motor skill and language outcomes. Lower abilities to imitate oral movements at 21 months have been found to correlate cross‐sectionally to lower expressive vocabulary and grammatical abilities at 21 months (Alcock and Krawczyk 2010) and longitudinally to lower performance on an omnibus test of expressive and receptive language abilities at 36 months (Alcock and Connor 2021). Oral‐motor vulnerabilities, along with broader gross and fine motor deficits (Brumbach and Goffman 2014; Hill 2001; Hsu and Bishop 2014; Sack et al. 2022; Sanjeevan and Mainela‐Arnold 2019; Ullman et al. 2020; Vuolo and Goffman 2018; Vuolo et al. 2017), have also been documented in older children with developmental language disorders, suggesting that the influence of oral‐motor vulnerabilities on language development may extend beyond infancy and toddlerhood. Compared to typically‐developing peers, children with developmental language disorders frequently demonstrate poorer control and coordination of the lips and tongue (Noterdaeme et al. 2002), as well as challenges with speech production and organization (Benham et al. 2018; Goffman 1999). Taken together, these findings from toddlers and older children with developmental language disorders suggest that feeding‐related oral‐motor vulnerabilities may play a role in shaping language trajectory both directly, by constraining speech sound production, vocabulary growth, and grammatical development, and indirectly, by reducing the quality of caregiver linguistic input, potentially contributing to the developmental relationship between feeding difficulties and language. Within this theoretical/experimental context, the present longitudinal study was designed to prospectively study developmental interactions between feeding difficulties and language development in typically‐developing children, with the goal of determining how age, differences in language development, and feeding difficulties interact over time. Such information will add to our comprehension of the developmental expression of the relationship between these seemingly diverse behaviors and the specific roles of feeding difficulties in shaping divergent language trajectories. Given the focus on a nonclinical population, feeding difficulties were operationally defined as difficulties “that are usually clinically subthreshold, often signaled by parents, and typically without serious medical or nutritional consequences” (McFarland, Poulin, et al. 2020, 22). This is in contrast to pediatric feeding disorders, which are age‐inappropriate feeding abilities and oral intake that results in significant medical, nutritional, and/or psychosocial consequences (Goday et al. 2019). Our specific experimental objectives were twofold. First, to characterize feeding difficulties in relation to language development by examining their occurrence across developmental age and language status and by assessing their cross‐sectional and longitudinal correlations with continuous indices of language development. Second, to determine whether feeding difficulties increased the likelihood that children's concurrent and later language skills would meet established criteria for classification as less developmentally advanced (Desmarais et al. 2008; Henrichs et al. 2011; Reilly et al. 2009; Thordardottir et al. 2010, 2011; Trudeau et al. 2008). Considering these objectives and research suggesting that food selectivity disrupts caregiver–infant mealtime interactions (Byrne et al. 2017; Jansen et al. 2017; Lindberg et al. 1996; Lucarelli et al. 2003; Mallan et al. 2018) and that underlying oral‐motor vulnerabilities can constrain speech sound and language development (Kehoe et al. 2018; Nip et al. 2009) and elicit simplified caregiver language input (Elmlinger et al. 2025), two experimental hypotheses were advanced. The first was that both food selectivity issues and feeding difficulties reflecting poor oral motor control, such as chewing difficulties or poor saliva management, would be among those most frequent in children with less developmentally advanced language skills and would correlate cross‐sectionally and longitudinally with continuous indices of language development. The second was that these feeding difficulties would also increase the likelihood that children's concurrent and later language skills would be less developmentally advanced. Methods Participants Participants were 8‐month‐old, typically‐developing infants from monolingual French‐speaking homes. Eight months was selected because it is the youngest age at which Quebec French language abilities can be assessed using a standardized parent‐report instrument (Trudeau et al. 2008). Restricting the sample to monolingual French‐speaking homes minimized the influence of language differences and multilingualism (Hoff et al. 2012; Thordardottir and Brandeker 2013). The questionnaire used for identifying feeding difficulties was developed and content validated in Quebec French only (McFarland, Poulin, et al. 2020). One hundred and forty infants between the ages of 7 months 3 weeks and 8 months 4 weeks were recruited through the birth lists of the Sainte‐Justine University Hospital Center, through 15 Montreal parent Facebook groups, a webpage of the University of Montreal, and through social and professional networks. Infants were included if (1) they were from monolingual French‐speaking families (i.e., exposed to French at least 80% of the time) (Trudeau et al. 2008; Trudeau and Sutton 2011), (2) born at term (37 weeks of gestation or later), and (3) without known or suspected craniofacial, visual, cognitive, or auditive impairments, neurodevelopmental delays or disorders, genetic disorders, or neurological lesions based on parent report and the French version of the Ages and Stages Questionnaire‐Third Edition (Council on Children With Disabilities 2006; Squire and Bricker 2016; Wilson et al. 2012). All recruitment and experimental procedures were approved by the ethics committee of Sainte‐Justine University Hospital Center (file number: 2016–990), and written consent was obtained from one of the primary caregivers. Three (2%) of the participants were later diagnosed with autism spectrum disorders as reported by parents at the 54‐month measurement period. This led to the exclusion of their data from all analyses. Feeding and/or language measurements were not performed for 11 (8%) children at 12 months, 13 (9%) children at 18 months, 26 (19%) children at 24 months, and 56 (41%) children at 54 months due to either parental withdrawal from the study (n = 1 at 12 months, n = 2 at 18 and 24 months, and n = 5 at 54 months) or nonresponsiveness to contact attempts (n = 10 at 12 months, n = 11 at 18 months, n = 24 at 24 months, and n = 51 at 54 months). As a result, the study sample included 137 infants at 8 months (98% retention rate), 126 children at 12 months (90% retention rate), 124 children at 18 months (89% retention rate), 111 children (79% retention rate) at 24 months, and 81 children at 54 months (58% retention rate). Presented in Table 1 are details of the study sample at 8 months and comparative Quebec population data (Statistics Canada 2017, 2024). There were no significant differences in either gender (χ 2(1) = 1.11, p = 0.29) or family size (χ 2(2) = 1.97, p = 0.37) between the study sample and Quebec population. TABLE 1Demographics of the study sample at the beginning of the project (8 months) and available comparative Quebec population data.Study sample at 8 months (N = 137)Available comparative Quebec population dataa Child genderFemale73 (53%)49%Male64 (47%)51%Adopted childYes0 (0%)—No137 (100%)—Family sizeOne child71 (52%)46%Two children46 (34%)38%Three children or more20 (15%)16%BreastfeedingYes125 (91%)—No12 (9%)—Breastfeeding duration≥ 8 months80 (64%)—Data not available1 (1%)—Age at solid food introduction< 4 months4 (3%)—4–5 months89 (65%)—6–7 months41 (30%)—Data not available3 (2%)— Note: — indicates that comparative Quebec population data were not available. a From Statistics Canada (2017, 2024).John Wiley & Sons, Ltd. Demographics of the study sample at the beginning of the project (8 months) and available comparative Quebec population data. Note: — indicates that comparative Quebec population data were not available. From Statistics Canada (2017, 2024). Procedures The Feeding Difficulties Questionnaire (McFarland, Poulin, et al. 2020) and the Quebec version of MacArthur–Bates Communicative Development Inventories (MBCDI) (Trudeau et al. 2008) were used to quantify feeding difficulties and language abilities, respectively, at 8, 12, 18, and 24 months (see Figure 1). The MBCDI was also used to determine children's language status at 12, 18, and 24 months, classifying them as having less developmentally advanced language skills or not according to established normative criteria (Desmarais et al. 2008; Henrichs et al. 2011; Reilly et al. 2009; Trudeau et al. 2008). At 54 months and because the MBCDI is suitable only for assessing language in children up to 30 months, the standardized sentence imitation task titled Le grand déménagement (Royle and Thordardottir 2003), was used to quantify language abilities and determine language status. The test was administered by the first author, a licensed speech‐language pathologist and native speaker of Quebec French, in a single testing session that lasted approximately 15–20 min. Standardized administration and scoring procedures were followed as outlined in Royle and Thordardottir (2003) and Thordardottir et al. (2010, 2011). Study timeline. Black boxes = language measures; white boxes = feeding measures. MBCDI = MacArthur–Bates Communicative Development Inventories; PLS‐5 = Preschool Language Scale‐Fifth Edition. aOnly 30 randomly‐selected children were to undergo a clinical assessment of their feeding abilities at 8, 12, and 18 months using the Schedule for Oral Motor Assessment. bOnly children with less developmentally advanced language skills at 24 months were to undergo a clinical assessment of their receptive language abilities at 24 months using the auditory comprehension subscale of the Preschool Language Scale‐Fifth Edition. cThese clinical evaluations could not be completed due to COVID. The initial plan was to supplement parent‐reported measures between 8 and 24 months with clinical assessments of feeding at 8, 12, and 18 months using the Schedule for Oral Motor Assessment (Reilly et al. 2000) and receptive language abilities at 24 months using the auditory comprehension subscale of the Preschool Language Scale‐Fifth Edition (Zimmerman et al. 2011) for subgroups of children. These clinical assessments could not be completed due to the COVID pandemic. Measurement Instruments and Data Preparation Feeding Difficulties Questionnaire The Feeding Difficulties Questionnaire (McFarland, Poulin, et al. 2020) is a parent‐report instrument designed to assess feeding difficulties directly from caregivers using rigorous methodological standards. The instrument was developed and content validated according to the guidelines for health measurement scales construction outlined by Streiner and Norman (2008), involving collaboration between clinical experts, elementary school teachers, and parents of young children (McFarland, Poulin, et al. 2020). This collaborative approach ensured that assessment items accurately measured the intended feeding constructs, utilized accessible language free from clinical jargon, and were comprehensible to parents with a Grade 6 reading level. The final instrument comprises 33 Likert‐type items (the focus of the study) and two open‐ended questions (McFarland, Poulin, et al. 2020). Thirty of the Likert‐type items assess individual feeding difficulties previously identified in retrospective studies of children with developmental language disorders (Malas et al. 2015, 2017), distributed across four domains: sucking difficulties (10 items), food transition difficulties (12 items), food selectivity difficulties (five items), and salivary control difficulties (three items). The remaining three Likert‐type items assess broader feeding‐related constructs including food appreciation, mealtime behaviors, and parental concerns regarding feeding, drawing on established measures in the literature (Archer et al. 1991; Crist and Napier‐Phillips 2001; Ramsay et al. 2011; Wardle et al. 2001; Wright et al. 2006). The open‐ended questions provide caregivers opportunity to report additional concerns or contextual information regarding their child's feeding patterns. All Likert‐type items employ a 5‐point response scale ranging from 1 (very rarely/not at all) to 5 (very often/absolutely). To minimize acquiescence bias, 11 of the 33 items used an inverted response scale (Streiner and Norman 2008); these responses were recoded to the non‐inverted scale prior to analyses. Parental responses were entered into a spreadsheet‐based data management system. Missing responses (657 out of 16,434 [4.0%]) or dual responses (eight out of 16,434 [0.05%]) were excluded from all analyses. The distribution of missing data was nonrandom, with the majority of the missing responses (610 of 657) occurring on the 10 sucking difficulty items, with frequency increasing with child age (10 missing responses at 8 months, 20 at 12 months, 180 at 18 months, and 400 at 24 months). This pattern likely reflects normative developmental changes in feeding. Although solid foods are typically introduced between 4 and 6 months (Arvedson et al. 2020; Fewtrell et al. 2017), breast milk and/or formula generally remain a primary source of nutrition until approximately 12 months (Health Canada 2025). By 18 and 24 months, however, many children are no longer bottle‐ and breastfeeding. Caregivers’ qualitative comments in the open‐ended questions of the questionnaire and direct communication with the first author indicated that these items were often omitted once their child had stopped bottle‐ and breastfeeding. Accordingly, and as suggested by one reviewer, responses to the 10 sucking difficulty items at 18 and 24 months were excluded from the analyses. Language Assessment Early language development (8–24 months): The Quebec French versions of the standardized parent‐report MBCDI (Trudeau et al. 2008) were used to measure expressive vocabulary development across the first 2 years of life (Desmarais et al. 2008; Reilly et al. 2009). The MBCDI: Words and Gestures was administered at 8 and 12 months and the MBCDI: Words and Sentences was administered at 18 and 24 months. Although infants at 8 months typically produce few or no words, the expressive vocabulary measure was included at this age to provide longitudinal assessment continuity across all measurement periods and to align with comparable longitudinal research of early communication development and related developmental domains (e.g., Cadime et al. 2017; Nip et al. 2011). For vocabulary quantification, expressive vocabulary size was calculated as the total number of words parents reported their child producing in the 408‐word vocabulary checklist of the MBCDI: Words and Gestures (at 8 and 12 months) and in the 680‐word vocabulary checklist of MBCDI: Words and Sentences (at 18 and 24 months) (Trudeau et al. 2008). This measure provided a continuous indice of language development throughout infancy and toddlerhood. The MBCDI was also used to determine whether children's language skills met established criteria for classification as less developmentally advanced at 12, 18, and 24 months. Children were classified as having less developmentally advanced language skills at 12 and 18 months if expressive vocabulary fell at or below the normative 10th percentile (Desmarais et al. 2008; Henrichs et al. 2011; Reilly et al. 2009; Trudeau et al. 2008). At 24 months, less developmentally advanced language skills were defined as expressive vocabulary at or below the 10th percentile and/or an absence of word combinations, determined by parental endorsement of “not yet” in response to the MBCDI's trichotomous choice question about word‐combining abilities (Charkaluk et al. 2019; Reilly et al. 2009). The inclusion of absent word combinations as a marker of less developmentally advanced language skills at 24 months was based on evidence linking this feature to later developmental language disorders (Rudolph and Leonard 2016) and its status as an indicator of early language vulnerability relevant to the investigation of feeding difficulties and language outcomes. The MBCDI could not be used to determine whether children had less developmentally advanced language skills at 8 months because a normative 10th percentile cut‐off for expressive vocabulary has not been established for this age group (Trudeau et al. 2008). Summary statistics for these language measures are reported in Table 2. TABLE 2Summary of MBCDI‐derived data at 8, 12, 18, and 24 months and sentence imitation task data at 54 months.MBCDI‐derived expressive vocabulary size n (%) of children with less developmentally advanced language skillsRangeM (SD)Median (IQR)8 months (N = 137)0–50.4 (0.9)0 (0–0)— a 12 months (N = 126)0–335.8 (6.7)3 (0–6.5)33 (26.19)18 months (N = 124)3–37870.3 (68.7)47 (24–92)18 (14.52)24 months (N = 111)17–616274.8 (153.0)286 (140–400)22 (19.82)Percentage scores from the sentence imitation task n (%) of children with less developmentally advanced language skillsRangeM (SD)Median (IQR)54 months (N = 81)39.5–10083.0 (12.9)85.8 (76.9–91.8)2 (2.5)Abbreviations: IQR = interquartile range, MBCDI = MacArthur–Bates Communicative Development Inventories. a The MBCDI could not be used at 8 months to determine whether children had less developmentally advanced language skills because no 10th percentile normative cut‐off for expressive vocabulary size has been established for this age group (Trudeau et al. 2008).John Wiley & Sons, Ltd. Summary of MBCDI‐derived data at 8, 12, 18, and 24 months and sentence imitation task data at 54 months. Abbreviations: IQR = interquartile range, MBCDI = MacArthur–Bates Communicative Development Inventories. The MBCDI could not be used at 8 months to determine whether children had less developmentally advanced language skills because no 10th percentile normative cut‐off for expressive vocabulary size has been established for this age group (Trudeau et al. 2008). Later language development (54 months): Language abilities at 54 months were assessed using the sentence imitation task Le grand déménagement (Royle and Thordardottir 2003), a standardized measure that simultaneously evaluates multiple components of language abilities, including speech perception, vocabulary knowledge, and grammatical competence (Klem et al. 2015; Thordardottir et al. 2010; Ward et al. 2024). The task involves presenting children with a narrative context containing target sentences, which they are instructed to repeat (Royle and Thordardottir 2003; Thordardottir et al. 2010, 2011). The total number of words repeated out of a total of 134 is calculated and converted to a percentage score, with lower scores indicating language skills at the lower end of the developmental spectrum (Ward et al. 2024). This continuous metric provided a single, integrated assessment of language abilities across phonological, lexical, and syntactic domains, thus allowing for comprehensive characterization of language development in the early school period. This task was also used to determine whether children's language skills met established criteria for classification as less developmentally advanced at 54 months. Children were classified as having less developmentally advanced language skills at 54 months if their percentage score on the sentence imitation task fell at or below the cut‐off established by Thordardottir et al. (2011). This cut‐off corresponds to the percentage score one standard deviation below the mean of the published normative sample (Thordardottir et al. 2010) and has the highest diagnostic accuracy for identifying developmental language disorders in Quebec French‐speaking preschool children (Thordardottir et al. 2011). Summary statistics for this language measure are reported in Table 2. Data Analyses Feeding Difficulty Patterns and Language Status Feeding difficulties were operationally defined as parental responses of four or five on any item of the Feeding Difficulties Questionnaire, consistent with prior work identifying “above threshold” feeding concerns in both typically‐developing (Hubbard et al. 2014; Schmitt et al. 2008; Toyama and Agras 2016) and neurodivergent children (Hubbard et al. 2014; McFarland, Poulin, et al. 2020; Schmitt et al. 2008). For the 10 sucking difficulty questionnaire items, the proportion of above threshold responses was calculated at 8 months for the full sample and, at 12 months, separately for children with and without less developmentally advanced language skills. For the remaining 23 questionnaire items, the proportion of above threshold responses was calculated at 8 months for the full sample and at 12, 18, and 24 months separately by language status. Cross‐Sectional and Longitudinal Associations Between Feeding Difficulties and Language Skills To examine associations between feeding difficulties and language outcomes, groups of feeding items that covaried were first identified using principal component analyses (PCA) on Feeding Difficulties Questionnaire responses (Benfer et al. 2017; Field 2018; Putnick et al. 2022; Samara et al. 2010). PCA with oblique rotation (direct oblimin), conducted in SPSS (version 27), yielded four empirically‐derived clusters: general mealtime difficulties, poor saliva management, difficulties in chewing and/or issues with food selectivity, and difficulties in sucking (see Table 3). These clusters were treated as aggregated variables in subsequent analyses to reduce the number of correlations and to limit collinearity among individual feeding items that could bias regression results. A child was classified as exhibiting a PCA‐derived feeding category if at least one item within that cluster received an above‐threshold response. For example, poor saliva management was coded as present if any of drooling, having difficulty controlling saliva, and/or having a lot of saliva in the mouth was rated 4 or 5. TABLE 3Summary of clusters of feeding items from the Feeding Difficulties Questionnaire (McFarland, Poulin, et al. 2020) identified by principal component analysisa.English translation of the itemsb Components1234General mealtimes difficulties*Question 20—My child has a good appetite. 0.81 0.030.20−0.02*Question 32—My child enjoys feeding times. 0.85 −0.02−0.11−0.08*Question 33—My child is easy to feed. 0.81 0.000.02−0.04Question 23—My child is a picky eater. 0.71 −0.04−0.15−0.13Question 19—My child is eating too slowly. 0.65 −0.030.010.14Question 31—I am worried about my child's eating. 0.65 0.27−0.020.14*Question 21—My child is eating like other child their age. 0.64 0.10−0.120.18Question 11—Introduction of purees is difficult. 0.62 0.07−0.20−0.11Question 22—My child is low weight for their age or their weight is diminishing. 0.57 −0.030.140.02Question 26—My child refuses to taste new foods. 0.50 −0.16−0.25−0.07Question 27—My child spits up foods. 0.50 −0.12−0.020.10Poor saliva managementQuestion 28—My child drools.0.05 0.91 0.05−0.02Question 30—My child has difficulty controlling their saliva.0.03 0.91 0.04−0.02Question 29—My child has a lot of saliva in the mouth.−0.10 0.91 −0.11−0.03Difficulties in chewing and/or issues with food selectivityQuestion 16—My child has difficulties masticating (chewing).0.040.03 −0.81 0.04Question 12—Introduction of pieces of food is difficult.0.150.08 −0.76 −0.02Question 17—My child swallows foods whole without chewing well.−0.11−0.04 −0.75 0.02Question 15—My child chokes.−0.08−0.14 −0.73 0.05Question 24—My child is sensitive to food temperatures or textures.0.20−0.02 −0.56 −0.11Question 14—My child gags.0.290.01 −0.54 0.04Question 25—My child only eats pureed or ground foods.−0.080.20 −0.53 0.05Difficulties in sucking*Question 2—My child has a strong suck.0.10−0.09−0.02 0.80 *Question 3—My child's sucking is constant during a feed.−0.01−0.080.04 0.74 *Question 1—My child correctly latches on to the breast or nipple.−0.030.020.03 0.68 *Question 7—My child finishes a feed within a reasonable time.0.010.10−0.14 0.66 Eigenvalues6.582.932.372.10% of variance26.3411.729.488.40 α Cronbach 0.880.910.800.72 Note: An inverted response scale was used for the items identified with an asterisk (*). Factor loadings above 0.40 appear in bold (Field 2018). a Principal component analysis was performed using oblique rotation (direct oblimin). The rotation converged in seven iterations. b Items are provided in English as a convenience for the non‐French‐speaking readership. Details of the original French items can be found in McFarland, Poulin, et al. (2020).John Wiley & Sons, Ltd. Summary of clusters of feeding items from the Feeding Difficulties Questionnaire (McFarland, Poulin, et al. 2020) identified by principal component analysisa. Note: An inverted response scale was used for the items identified with an asterisk (*). Factor loadings above 0.40 appear in bold (Field 2018). Principal component analysis was performed using oblique rotation (direct oblimin). The rotation converged in seven iterations. Items are provided in English as a convenience for the non‐French‐speaking readership. Details of the original French items can be found in McFarland, Poulin, et al. (2020). Two complementary statistical approaches were then used. First, point‐biserial correlations were computed to assess cross‐sectional and longitudinal associations between the presence/absence of the PCA‐derived categories of difficulties in sucking at 8 and 12 months and general mealtimes difficulties, poor saliva management, and difficulties in chewing and/or issues with food selectivity at 8, 12, 18, and 24 months and continuous language outcomes. These outcomes included (1) MBCDI‐derived expressive vocabulary size at 8, 12, 18, and 24 months and (2) percentage scores on the sentence imitation task at 54 months. Point‐biserial correlation was selected because it is appropriate for testing associations between dichotomous and continuous variables, and statistical significance was set at p ≤ 0.05. Second, logistic regression analyses were conducted in R (version 4.3.3) using the “glm” function of the “lme4” package (Bates et al. 2015) to quantify the odds of co‐occurring or later occurring less developmentally advanced language skills as a function of the PCA‐derived feeding categories. For these analyses, the presence/absence of the categories of difficulties in sucking at 8 and 12 months and general mealtimes difficulties, poor saliva management, and difficulties in chewing and/or issues with food selectivity at 8, 12, 18, and 24 months served as independent variables, while early language status (presence/absence of less developmentally advanced language skills) at 12, 18, and 24 months served as the binary outcome variables. Logistic regression was not carried out for the 54‐month language status because only two children (2.5%) met criteria for less developmentally advanced language skills at 54 months (see Table 2), a percentage of occurrence substantially lower than the 7.4% reported in previous epidemiological work (Tomblin et al. 1997). In cases of quasi‐complete separation, Firth logistic regression was substituted using the “logistf” function of the “logistf” package (Heinze et al. 2023) to obtain bias‐reduced estimates (Heinze 2006). Statistical significance was set at p ≤ 0.05, and results were reported as odds ratios with 95% confidence intervals. Results Feeding Difficulty Patterns and Language Status Shown in Figure 2 are the percentages of occurrence of each feeding difficulty assessed by the Feeding Difficulties Questionnaire across all participating infants at 8 months (Panel A) and for children with (striped bars) and without (solid bars) less developmentally advanced language skills at 12 months (Panel B), 18 months (Panel C), and 24 months (Panel D). Feeding difficulties were color‐coded according to the categories of feeding difficulties previously identified in Malas et al. (2015, 2017) and McFarland, Poulin, et al. (2020). The top three most frequently occurring feeding difficulties are highlighted to aid in characterizing the evolving patterns of occurrence of feeding difficulties across developmental ages and/or language status, with the assumption that those feeding difficulties most frequent in children with less developmentally advanced language skills may contribute to or reflect subtle developmental disruptions that shape divergent language trajectories. Percentages of occurrence of each feeding difficulty (FD) across all participating infants at 8 months and for children with (striped bars) and without (solid bars) less developmentally advanced language skills at 12, 18, and 24 months. At 8 months, the number of parental responses from different children (N) used to calculate the percentage of occurrence for each feeding difficulty ranged from 134 to 137. At 12 months, N ranged from 91 to 93 (without less developmentally advanced language skills) and 32–33 (with less developmentally advanced language skills); at 18 months, from 102 to 106 (without less developmentally advanced language skills) and 18 (with less developmentally advanced language skills); and at 24 months, from 84 to 89 (without less developmentally advanced language skills) and 21–22 (with less developmentally advanced language skills). Variability in N was due to missing data from omissions/oversights, dual responses on single items, and cessation of both bottle‐ and breastfeeding at 8 months (one child) and 12 months (two children without less developmentally advanced language skills). At 8 months (Figure 2A), the three most frequently occurring feeding difficulties were only eating pureed or ground foods (Question 25; 30% [41/137]), sensitivity to food temperatures or textures (Question 24; 26% [35/137]), and drooling (Question 28; 26% [36/137]). At 12 months (Figure 2B) and similarly for children with and without less developmentally advanced language skills, the two most frequently occurring feeding difficulties were sensitivity to food temperatures or textures (Question 24; without less developmentally advanced language skills: 17% [16/93]; with less developmentally advanced language skills: 24% [8/33]) and bottle‐ or breastfeeds not easily finished (Question 8; without less developmentally advanced language skills: 14% [13/91]; with less developmentally advanced language skills: 15% [5/33]). The third most frequently occurring feeding difficulty differed with language status, with picky eating (Question 23; 13% [12/93]) being the third most frequent for children without less developmentally advanced language skills while for children with less developmentally advanced language skills, there was a tie between spitting up foods (Question 27; 15% [5/33]) and not feeding themselves like other children their age (Question 21; 15% [5/33]). At 18 months (Figure 2C) and looking separately by language status, the three most frequently occurring feeding difficulties for children without less developmentally advanced language skills were sensitivity to food temperatures or textures (Question 24; 18% [19/106]), spitting up foods (Question 27; 17% [18/106]), and picky eating (Question 23; 15% [16/106]), all related to food selectivity. Among children with less developmentally advanced language skills, the three most frequently occurring feeding difficulties were drooling (Question 28; 22% [4/18]), having lots of saliva in the mouth (Question 29; 22% [4/18]), and difficulty controlling saliva (Question 30; 22% [4/18]), all related to poor saliva management. At 24 months (Figure 2D) and looking separately by language status, the three most frequently occurring feeding difficulties for children without less developmentally advanced language skills were sensitivity to food temperatures or textures (Question 24; 20% [18/89]), followed by a tie between picky eating (Question 23; 11% [10/89]) and not being easy to feed (Question 33; 11% [10/89]). Among children with less developmentally advanced language skills, the three most frequently occurring feeding difficulties were picky eating (Question 23; 23% [5/22]), sensitivity to food temperatures or textures (Question 24; 18% [4/22]), and difficulties with the introduction of purees (Question 11; 10% [2/21]). Cross‐Sectional and Longitudinal Associations Between Feeding Difficulties and Language Skills Point‐Biserial Correlation Analyses Presented in Table 4 are the results of cross‐sectional and longitudinal point‐biserial correlations between PCA‐derived categories of difficulties in sucking at 8 and 12 months and general mealtimes difficulties, poor saliva management, and difficulties in chewing and/or issues with food selectivity at 8, 12, 18, and 24 months and children's MBCDI‐derived expressive vocabulary size at 8, 12, 18, and 24 months. Significant negative correlations were found between the PCA‐derived category of poor saliva management at 18 months and MBCDI‐derived expressive vocabulary size at 18 months (rpb = −0.18, p = 0.05) and 24 months (rpb = −0.21, p = 0.03). These significant correlations were observed in a sample characterized by expressive vocabulary ranging from three to 378 words at 18 months (M = 70.3, SD = 68.7; median = 47, IQR = 24–92) and from 17 to 616 words at 24 months (M = 274.8, SD = 153.0; median = 286, IQR = 140–400) and where 14.5% (18/124) and 19.8% (22/111) of children had less developmentally advanced language skills at 18 and 24 months, respectively (see Table 2). No other PCA‐derived feeding category at any of the measurement periods was found to significantly correlate with MBCDI‐derived expressive vocabulary measures at 8, 12, 18, and 24 months. TABLE 4Cross‐sectional and longitudinal point‐biserial correlations between each PCA‐derived feeding category at 8, 12, 18, and 24 months and children's MBCDI‐derived expressive vocabulary size at 8, 12, 18, and 24 months and percentage scores from the SI task at 54 months.MBCDI‐derived expressive vocabulary size at 8 monthsMBCDI‐derived expressive vocabulary size at 12 monthsMBCDI‐derived expressive vocabulary size at 18 monthsMBCDI‐derived expressive vocabulary size at 24 monthsPercentage scores from the SI task at 54 months N rpb p N rpb p N rpb p N rpb p N rpb p General mealtimes difficulties at 8 months1370.010.881260.020.84124 0.11 0.231110.090.3781−0.010.92Poor saliva management at 8 months137−0.050.56126−0.090.30124−0.040.69111−0.020.8281−0.060.60Difficulties in chewing and/or issues with food selectivity at 8 months1370.150.081260.010.90124−0.060.53111−0.070.46810.050.69Difficulties in sucking at 8 months136−0.060.50125−0.110.24123−0.010.891100.100.3181−0.070.55General mealtimes difficulties at 12 months———126−0.100.28120−0.100.26109−0.030.74790.000.99Poor saliva management at 12 months———126−0.140.13120−0.020.86109−0.060.5579−0.030.82Difficulties in chewing and/or issues with food selectivity at 12 months———1260.030.78120−0.020.861090.010.90790.030.81Difficulties in sucking at 12 months———1240.080.371180.070.451070.040.6978−0.030.79General mealtimes difficulties at 18 months——————124−0.040.70111−0.060.5780−0.050.66Poor saliva management at 18 months—————— 124 −0.18 0.05 111 −0.21 0.03 80−0.100.40Difficulties in chewing and/or issues with food selectivity at 18 months——————1240.040.661110.090.33800.170.13General mealtimes difficulties at 24 months—————————111−0.010.94800.030.77Poor saliva management at 24 months—————————111−0.070.4780−0.170.13Difficulties in chewing and/or issues with food selectivity at 24 months—————————111−0.170.07800.070.52 Note: — represents correlations not tested. Significative correlations (two‐tails) are in bold. Negative correlations represent PCA‐derived feeding categories correlated with lower scores on the continuous indices of language development, whereas positive correlations represent PCA‐derived feeding categories correlated with higher scores on the continuous indices of language development. The number of children (N) included in each correlation varied due to nonresponsiveness to contact attempts at one or more measurement periods and missing responses from items about difficulties in sucking caused by cessation of both bottle‐ and breastfeeding at 8 months (one child) and 12 months (two children).Abbreviations: MBCDI = MacArthur–Bates Communicative Development Inventories, SI = sentence imitation.John Wiley & Sons, Ltd. Cross‐sectional and longitudinal point‐biserial correlations between each PCA‐derived feeding category at 8, 12, 18, and 24 months and children's MBCDI‐derived expressive vocabulary size at 8, 12, 18, and 24 months and percentage scores from the SI task at 54 months. 0.11 Note: — represents correlations not tested. Significative correlations (two‐tails) are in bold. Negative correlations represent PCA‐derived feeding categories correlated with lower scores on the continuous indices of language development, whereas positive correlations represent PCA‐derived feeding categories correlated with higher scores on the continuous indices of language development. The number of children (N) included in each correlation varied due to nonresponsiveness to contact attempts at one or more measurement periods and missing responses from items about difficulties in sucking caused by cessation of both bottle‐ and breastfeeding at 8 months (one child) and 12 months (two children). Abbreviations: MBCDI = MacArthur–Bates Communicative Development Inventories, SI = sentence imitation. Also presented in Table 4 are the results of point‐biserial correlations between PCA‐derived categories of difficulties in sucking at 8 and 12 months and general mealtimes difficulties, poor saliva management, and difficulties in chewing and/or issues with food selectivity at 8, 12, 18, and 24 months and the percentage scores from the sentence imitation task at 54 months. As can be seen, no PCA‐derived feeding category at 8, 12, 18, and 24 months was found to significantly correlate with the percentage scores from the sentence imitation task at 54 months. Logistic Regression Analyses Presented in Table 5 are the results from univariate logistic regression analyses that quantified the odds of less developmentally advanced language skills at 12, 18, and 24 months as a function of the PCA‐derived categories of difficulties in sucking at 8 and 12 months and general mealtimes difficulties, poor saliva management, and difficulties in chewing and/or issues with food selectivity at 8, 12, 18, and 24 months. As shown in Table 5, only the PCA‐derived category of poor saliva management at 18 months was significantly associated with increased odds of co‐occurring (p = 0.006) and later occurring (p = 0.03) less developmentally advanced language skills. Children with poor saliva management at 18 months were 6.41 (95% confidence interval = 1.45–24.40) times more likely to have co‐occurring less developmentally advanced language skills at 18 months and 4.07 (95% confidence interval = 1.07–15.08) times more likely to have later occurring less developmentally advanced language skills at 24 months. TABLE 5Results from univariate logistic regression analyses on the odds of having co‐occurring or later occurring less developmentally advanced language skills at 12, 18 and/or 24 months based on PCA‐derived feeding categories at 8, 12, 18, and 24 months.Less developmentally advanced language skills at 12 monthsLess developmentally advanced language skills at 18 monthsLess developmentally advanced language skills at 24 months N OR (95% CI) p N OR (95% CI) p N OR (95% CI) p General mealtimes difficulties at 8 months126 1.61 (0.66–3.81) 0.28124 0.30 (0.05–1.15) 0.13111 0.57 (0.15–1.71) 0.35Poor saliva management at 8 months126 1.22 (0.51–2.83) 0.64124 1.61 (0.55–4.50) 0.37111 0.58 (0.18–1.63) 0.33Difficulties in chewing and/or issues with food selectivity at 8 months126 0.79 (0.35–1.74) 0.56124 1.16 (0.42–3.20) 0.77111 0.85 (0.32–2.17) 0.73Difficulties in sucking at 8 months125 0.92 (0.13–4.26) 0.93123 0.31a (0.00–2.68) 0.35110 0.55 (0.03–3.34) 0.59General mealtimes difficulties at 12 months126 1.55 (0.66–3.59) 0.31120 1.23 (0.36–3.72) 0.72109 0.68 (0.18–2.07) 0.52Poor saliva management at 12 months126 1.29 (0.33–4.28) 0.69120 1.21 (0.18–5.14) 0.82109 0.37 (0.02–2.10) 0.36Difficulties in chewing and/or issues with food selectivity at 12 months126 0.82 (0.31–2.00) 0.68120 0.39 (0.06–1.51) 0.23109 0.35 (0.05–1.39) 0.19Difficulties in sucking at 12 months124 1.42 (0.29–5.73) 0.64118 0.78 (0.04–4.73) 0.82107 1.50 (0.21–7.15) 0.63General mealtimes difficulties at 18 months———124 0.61 (0.18–1.75) 0.38111 0.83 (0.29–2.19) 0.72Poor saliva management at 18 months——— 124 6.41 (1.65–24.40) 0.006 111 4.07 (1.07–15.08) 0.03 Difficulties in chewing and/or issues with food selectivity at 18 months———124 0.53 (0.12–1.76) 0.34111 0.15 (0.01–0.76) 0.07General mealtimes difficulties at 24 months——————111 1.55 (0.55–4.10) 0.39Poor saliva management at 24 months——————111 1.17 (0.17–5.30) 0.85Difficulties in chewing and/or issues with food selectivity at 24 months——————111 0.84 (0.25–2.42) 0.76 Note: — represents regressions not tested. Significative regressions are in bold. The number of children (N) included in each univariate regression varied due to nonresponsiveness to contact attempts at one or more measurement periods and missing responses from items about difficulties in sucking caused by cessation of both bottle‐ and breastfeeding at 8 months (one child) and 12 months (two children).Abbreviation: MBCDI = MacArthur–Bates Communicative Development Inventories. aLogistic regression was substituted to Firth's logistic regression to account for quasi‐complete separation.John Wiley & Sons, Ltd. Results from univariate logistic regression analyses on the odds of having co‐occurring or later occurring less developmentally advanced language skills at 12, 18 and/or 24 months based on PCA‐derived feeding categories at 8, 12, 18, and 24 months. 1.61 (0.66–3.81) 0.30 (0.05–1.15) 0.57 (0.15–1.71) 1.22 (0.51–2.83) 1.61 (0.55–4.50) 0.58 (0.18–1.63) 0.79 (0.35–1.74) 1.16 (0.42–3.20) 0.85 (0.32–2.17) 0.92 (0.13–4.26) 0.31a (0.00–2.68) 0.55 (0.03–3.34) 1.55 (0.66–3.59) 1.23 (0.36–3.72) 0.68 (0.18–2.07) 1.29 (0.33–4.28) 1.21 (0.18–5.14) 0.37 (0.02–2.10) 0.82 (0.31–2.00) 0.39 (0.06–1.51) 0.35 (0.05–1.39) 1.42 (0.29–5.73) 0.78 (0.04–4.73) 1.50 (0.21–7.15) 0.61 (0.18–1.75) 0.83 (0.29–2.19) 6.41 (1.65–24.40) 4.07 (1.07–15.08) 0.53 (0.12–1.76) 0.15 (0.01–0.76) 1.55 (0.55–4.10) 1.17 (0.17–5.30) 0.84 (0.25–2.42) Note: — represents regressions not tested. Significative regressions are in bold. The number of children (N) included in each univariate regression varied due to nonresponsiveness to contact attempts at one or more measurement periods and missing responses from items about difficulties in sucking caused by cessation of both bottle‐ and breastfeeding at 8 months (one child) and 12 months (two children). Abbreviation: MBCDI = MacArthur–Bates Communicative Development Inventories. aLogistic regression was substituted to Firth's logistic regression to account for quasi‐complete separation. Because univariate logistic regression analyses revealed no significant association between any PCA‐derived feeding category and language status at 12 months, and only one PCA‐derived feeding category was found to be associated with higher odds of less developmentally advanced language skills at 18 months and 24 months at the p ≤ 0.05 level (see Table 5), no further multivariate logistic regression analyses were conducted to control for the potential impact of multiple PCA‐derived feeding categories on the odds ratios. Comparison of Children With and Without 54‐Month Follow‐Up Data As suggested by one reviewer, additional analyses were conducted to examine whether early feeding difficulties and language abilities differed between children who completed the study and those who did not. Children in the 24‐month study sample (N = 111) were divided into two groups based on completion of the 54‐month language assessment: those with follow‐up data at 54 months (n = 80) and those without (n = 31). Mean and median MBCDI‐derived expressive vocabulary sizes and the proportion of children with less developmentally advanced language skills at 24 months were calculated for both groups. Percentages of occurrence for the PCA‐derived feeding categories at 8, 12, 18, and 24 months were also calculated for both groups. Group differences were examined using a Mann–Whitney test for the continuous variable and chi‐square tests for categorical variables. Table 6 presents summary statistics for the MBCDI‐derived language measures at 24 months for children with and without 54‐month follow‐up data. No significant group differences were observed in expressive vocabulary (Mann–Whitney U = 1147.50, p = 0.54) or in the proportion of children with less developmentally advanced language skills (χ 2(1) = 2.30, p = 0.13), although 41% (9/22) of children with less developmentally advanced language skills at 24 months were lost to follow‐up. Table 7 presents the percentages of occurrence for PCA‐derived feeding categories at 8, 12, 18, and 24 months for children with and without 54‐month follow‐up data. Similarly, no significant group differences were observed at any measurement period. TABLE 6Summary of MBCDI‐derived data at 24 months for children with and without 54‐month follow‐up language data.MBCDI‐derived expressive vocabulary size at 24 months n (%) of children with less developmentally advanced language skills at 24 monthsRangeM (SD)Median (IQR) Children with 54‐month follow‐up data (n = 80)a 40–616278.8 (142.4)294.5 (150.75–402.25)13 (16.25)Children without 54‐month follow‐up data (n = 31)17–604264.5 (179.5) 237 (109.5–385) 9 (29.0)Statistical significanceMann–Whitney U = 1147.50, p = 0.54 χ 2(1) = 2.30, p = 0.13Abbreviations: IQR = interquartile range, MBCDI = MacArthur–Bates Communicative Development Inventories. aThe number of children with data at both 24 and 54 months differs from the total number of participants at 54 months (N = 81) because one child provided data at 54 months but not at 24 months.John Wiley & Sons, Ltd. Summary of MBCDI‐derived data at 24 months for children with and without 54‐month follow‐up language data. Children with 54‐month follow‐up data (n = 80)a 237 (109.5–385) Abbreviations: IQR = interquartile range, MBCDI = MacArthur–Bates Communicative Development Inventories. aThe number of children with data at both 24 and 54 months differs from the total number of participants at 54 months (N = 81) because one child provided data at 54 months but not at 24 months. TABLE 7Percentage of occurrence of PCA‐derived feeding categories at 8, 12, 18, and 24 months in children with and without 54‐month follow‐up language data.Children with 54‐month follow‐up dataChildren without 54‐month follow‐up dataStatistical significance N n (%) N n (%)General mealtimes difficulties at 8 months80a 22 (28)317 (23) χ 2(1) = 0.28, p = 0.60Poor saliva management at 8 months80a 23 (29)3112 (39) χ 2(1) = 1.03, p = 0.31Difficulties in chewing and/or issues with food selectivity at 8 months80a 34 (43)3115 (48) χ 2(1) = 0.31, p = 0.58Difficulties in sucking at 8 months78b 4 (5)30b 4 (13) χ 2(1) = 2.25, p = 0.13General mealtimes difficulties at 12 months78b 19 (24)319 (29) χ 2(1) = 0.25, p = 0.61Poor saliva management at 12 months78b 7 (9)315 (16) χ 2(1) = 1.16, p = 0.28Difficulties in chewing and/or issues with food selectivity at 12 months78b 17 (22)316 (19) χ 2(1) = 0.08, p = 0.78Difficulties in sucking at 12 months77b 5 (6)30b 3 (10) χ 2(1) = 0.38, p = 0.54General mealtimes difficulties at 18 months80a 26 (33)3113 (42) χ 2(1) = 0.87, p = 0.35Poor saliva management at 18 months80a 10 (13)311 (3) χ 2(1) = 2.15, p = 0.14Difficulties in chewing and/or issues with food selectivity at 18 months80a 17 (21)316 (19) χ 2(1) = 0.05, p = 0.83General mealtimes difficulties at 24 months80a 19 (24)3113 (42) χ 2(1) = 3.60, p = 0.06Poor saliva management at 24 months80a 8 (10)311 (3) χ 2(1) = 1.38, p = 0.24Difficulties in chewing and/or issues with food selectivity at 24 months80a 17 (21)3111 (35) χ 2(1) = 2.40, p = 0.12 aThe number of children with data at both 24 and 54 months differs from the total number of participants at 54 months (N = 81) because one child provided data at 54 months but not at 24 months. bThe number of children (N) varied due to nonresponsiveness to contact attempts at one or more measurement periods and missing responses from items about difficulties in sucking caused by cessation of both bottle‐ and breastfeeding at 8 months (one child without 54‐month follow‐up data) and 12 months (one child with 54‐month follow‐up data, one child without 54‐month follow‐up data).John Wiley & Sons, Ltd. Percentage of occurrence of PCA‐derived feeding categories at 8, 12, 18, and 24 months in children with and without 54‐month follow‐up language data. aThe number of children with data at both 24 and 54 months differs from the total number of participants at 54 months (N = 81) because one child provided data at 54 months but not at 24 months. bThe number of children (N) varied due to nonresponsiveness to contact attempts at one or more measurement periods and missing responses from items about difficulties in sucking caused by cessation of both bottle‐ and breastfeeding at 8 months (one child without 54‐month follow‐up data) and 12 months (one child with 54‐month follow‐up data, one child without 54‐month follow‐up data). Discussion Interactive feeding is a key moment for social interaction and language stimulation (Feldman 2015; Kaye 1977; Leech 2021; Leech et al. 2018; Snow and Beals 2006; Weizman and Snow 2001; Zimmerman et al. 2019) and subclinical, parent‐reported feeding difficulties can create suboptimal language‐learning environment during mealtimes (Byrne et al. 2017; Harding et al. 2013; Jansen et al. 2017; Lindberg et al. 1996; Mallan et al. 2018) and/or reflect underlying poor oral‐motor control that shape early speech and vocabulary development (Kehoe et al. 2018; Nip et al. 2009) and caregiver language input (Elmlinger et al. 2025). The present longitudinal investigation provides evidence for developmental interactions between feeding difficulties and language in full‐term, typically‐developing infants, with both mealtime environmental and oral‐motor influences potentially contributing to this association as discussed below and summarized in Figure 3. Summary of hypotheses, findings, and explanation/implications. Results revealed that poor saliva management at 18 months was negatively correlated with expressive vocabulary size at both 18 and 24 months. Pattern analyses similarly revealed that, at 18 months, the three most frequently occurring feeding difficulties in children with less developmentally advanced language skills were all related to poor saliva management, a pattern that contrasts sharply with the food selectivity issues most frequently seen in children without less developmentally advanced language skills at the same age. Regression analyses further revealed that children with poor saliva management at 18 months were over six times more likely to be identified with less developmentally advanced language skills at 18 months and four times more likely at 24 months. These findings with typically‐developing children are generally consistent with previous research in toddlers and preschoolers with significant neurodevelopmental and motor disorders (Chang et al. 2012). The fact that poor saliva management at 18 months correlates negatively with expressive vocabulary size at 18 and 24 months and increases the likelihood of less developmentally advanced language skills both concurrently at 18 months and later at 24 months may reflect the contribution of underlying oral‐motor vulnerabilities to early language development. From a developmental motor perspective, children typically develop sufficient oral‐motor control to manage saliva effectively by 15–18 months (Chaléat‐Valayer et al. 2016; Fairhurst and Cockerill 2011; Meningaud et al. 2006). Therefore, poor saliva management at 18 months may indicate delays in oral‐motor control development and its association with adverse language outcomes may arise from the constraints these delays impose on sound inventory and syllable shapes (Nip et al. 2009) ultimately resulting in reduced early expressive vocabulary and grammar (Kehoe et al. 2018). Caregivers may also provide less diverse and complex language input in response to these less advanced oral productions (Elmlinger et al. 2025), potentially establishing a bidirectional loop that hinders communicative development. Taken together, these results suggest that poor saliva management at 18 months may represent subtle developmental disruptions that may shape divergent language trajectories even in the absence of known biomedical risk factors. Although pattern analyses revealed that picky eating was among the most frequently occurring feeding difficulties in both children with and without less developmentally advanced language skills at 24 months, it was particularly prominent in those with less developmentally advanced language skills, occurring more than twice as often and ranking as the most frequent feeding difficulty. The previous retrospective data of Malas et al. (2015, 2017) similarly revealed that food selectivity issues were among the most frequently documented feeding difficulties in the clinical files of older children with developmental language disorders. It is therefore possible that the negative impact of food selectivity on language development, through less lexically diverse parent input (Harding et al. 2013) and negative, emotionally charged mealtime interactive cycles in which both food selectivity and nonresponsive parental behaviors reinforce one another (Jansen et al. 2017; Mallan et al. 2018) may not be evident until later in development. This interpretation is consistent with that of Cardona Cano et al. (2015), who suggested that persistent or later‐onset picky eating may be more strongly shaped by parent–child interaction dynamics than early‐onset picky eating that resolves quickly. Skinner et al. (1998) also observed that children's emerging abilities to verbally communicate food dislikes between 16 and 20 months coincided with an increase in nonresponsive and coercive parental behaviors during mealtimes, further supporting the idea of later‐occurring impacts of food selectivity on language development. In fact, verbal expressions of food dislikes in later part of toddlerhood (Skinner et al. 1998) and increasing autonomy in self‐feeding (Arvedson et al. 2020; Carruth et al. 2004) may make food selectivity more salient and disruptive during mealtimes, which in turn may perturb established feeding routines, confront caregivers’ expectations that mealtime behaviors would be less of a battle with age (Harris et al. 2018), escalate their emotional stress and frustration (Harris et al. 2018; Wolstenholme et al. 2020), challenge their efforts to maintain responsive practices, and ultimately lead to less responsive parental behaviors and less lexically diverse parental input. Further evidence for the longer‐term negative effects of food selectivity on language development comes from the longitudinal work of Putnick et al. (2022), who reported that the associations between feeding difficulties (primarily related to food selectivity) and communication delays strengthened between 18 and 36 months. Future research should extend beyond the early developmental window examined here and use longitudinal or cross‐sectional designs to investigate how these dynamics unfold across later toddlerhood and early childhood. When looking at the correlations between early feeding difficulties from 8 to 24 months and later language abilities at 54 months, no significant association emerged. At present, we can only speculate as to why longer‐term developmental links were not observed. One possible explanation is that the influence of early feeding difficulties on language development is restricted to the first few years of life. However, this interpretation seems unlikely given prior findings showing high percentages of early feeding difficulties in children later diagnosed with developmental language disorders (Malas et al. 2015, 2017) and longitudinal evidence that abnormal sucking patterns predict reduced language comprehension at Age 5 in preterm children (Wolthuis‐Stigter et al. 2017). Further, negative mealtime interactional dynamics appear to intensify over time rather than diminish (Cardona Cano et al. 2015). Given that 41% of children with less developmentally advanced language skills at 24 months were lost to the 54‐month follow‐up, and that less developmentally advanced language skills at this age are associated with an increased likelihood of poorer later language outcomes (Hammer et al. 2017; Reilly et al. 2010), a more plausible explanation is that the sample assessed at 54 months included relatively few children with language abilities at the lower range of the developmental spectrum. This underrepresentation is reflected in the lower occurrence of children with less developmentally advanced language skills in our sample at 54 months (2.5%) compared with the 7.4% widely‐cited prevalence reported by Tomblin et al. (1997) despite our rather large experimental sample. Additional research with even larger samples extending beyond our target population of French‐speaking children from the province of Quebec and with extended follow‐up periods could be used to clarify any longer‐term consequences of early feeding difficulties on language development. In conclusion, this longitudinal study adds to the growing body of literature supporting developmental cross‐system interactions between feeding–swallowing and language across diverse pediatric populations and age groups (e.g., Adams‐Chapman et al. 2013, 2015; Crapnell et al. 2015; Lainwala et al. 2020; Malandraki et al. 2022; Malas et al. 2015, 2017; McFarland, Poulin, et al. 2020; Motion et al. 2001; Neumeyer et al. 2019; Somaraki et al. 2024; Wolthuis‐Stigter et al. 2017; Yardımcı‑Lokmanoğlu et al. 2024). It focused on a key developmental window (8–24 months) during which both feeding and language abilities rapidly emerge (Arvedson et al. 2020; Carruth and Skinner 2002; Chaléat‐Valayer et al. 2016; Delaney et al. 2021; Fairhurst and Cockerill 2011; Gisel 1991; Meningaud et al. 2006; Reilly et al. 2009; Rudolph and Leonard 2016; Trudeau et al. 2008), while also assessing language abilities at a later stage when they are more stable and complex (Bishop et al. 2017). A key methodological strength was the exclusion of known or suspected developmental vulnerability factors, which are known to independently influence both feeding and language development (Bishop et al. 2017; van Noort‐van der Spek et al. 2012; Pados et al. 2021; Reilly et al. 1996; Sharp et al. 2013) and shape their interactions (Adams‐Chapman et al. 2013; Malandraki et al. 2022; Neumeyer et al. 2019; Wolthuis‐Stigter et al. 2017). The demographic composition of the study sample further mirrored that of the Quebec population in terms of sex and family size (Statistics Canada 2017, 2024), and our language data from 8 to 24 months were consistent with previous studies (Bouchard et al. 2009; Reilly et al. 2009; Trudeau et al. 2008). The present findings indicate that specific patterns of feeding difficulties are associated with differences in early language development. These results suggest that feeding difficulties may disrupt language‐rich mealtime interactions and/or reflect underlying oral‐motor vulnerabilities that shape language development. Conflicts of Interest The authors declare no conflicts of interest.