Connecting preschoolers’ spontaneous speech patterns to future language skills: A three-year concurrent and longitudinal cohort study of canonical proportion as a developmental index

Purpose: Understanding how children’s spontaneous language behavior relates to standardizedmetrics of speech and language remains a crucial challenge in studies of language acquisition andclinical assessment. Traditional lab- and clinic-based paradigms are time- and resource-intensive,and may not fully capture a child’s underlying language competencies. This study aimed toinvestigate whether automated analysis of naturalistic, child-centered audio recordings can indexthe developmental trajectory of speech-language abilities from ages 3 to 5 years.Methods: A longitudinal design was employed with N=155 preschoolers, who were followedfrom age 3 to 5 years. Deep learning speech classification methods were used to computecanonical proportion—the proportion of a child’s speech that is produced as canonical syllables(well-formed syllables with a consonant–vowel transition, such as “ba”), a key marker of speechmotor control development and phonological representation building—from child-centered audiorecordings collected at ages 3 and 4. Standardized lab-based assessments of speech and languagedevelopment were administered one year later at ages 4 and 5.Results: Canonical proportion measures significantly predicted multiple dimensions of laterspeech-language development. The strongest longitudinal associations were observed forconsonant articulation skill and vocabulary size. Weaker, though still significant, relationshipswere found for meta-phonological skills, including phonological awareness and phonologicalworking memory.Conclusion: Findings suggest that children’s spontaneous, everyday speech productionpatterns provide moderate indices of subsequent language development across multiple domains.Automated measures such as canonical proportion hold promise for expanding and diversifyingapproaches to studying language acquisition and for complementing traditional lab-basedassessments.