FluencyBank English MichiganChild Corpus

	Soo-Eun Chang Department of Psychiatry University of Michigan sooeunc@med.umich.edu		Megan Sheppard Department of Psychiatry University of Michigan mesheppa@med.umich.edu
	Saralyn Rubsam Department of Psychiatry University of Michigan sara@speech-languagesolutions.com		Kristin Hicks Communication Sciences and Disorders Michigan State University hickskr@msu.edu

Browsable transcripts

Download transcripts

Media

Citation information

1. Treleaven, S., Sheppard, M., Bernstein Ratner, N., Chang, S-E. (2026). Expressive language growth among young children who do and do not stutter. J Communication Disorders, 10.1016/j.jcomdis.2026.106659. https://www.sciencedirect.com/science/article/pii/S0021992426000432
2. Chang S-E, Zhu DC (2013). Neural network connectivity differences in children who stutter. Brain, 136(Pt 12), 3709-3726. https://pubmed.ncbi.nlm.nih.gov/24131593/

Other publications relevant to this corpus include:

1. Chang S-E, Zhu DC, Choo AL, Angstadt M (2015). White matter neuroanatomical differences in young children who stutter. Brain, 138(Pt 3), 694-711. https://pubmed.ncbi.nlm.nih.gov/25619509/
2. Chow HM, Chang S-E (2017). White matter developmental trajectories associated with persistence and recovery of childhood stuttering. Human Brain Mapping, 38(7), 3345-3359. https://pubmed.ncbi.nlm.nih.gov/28390149/
3. Chow HM, Garnett EO, Koenraads SP C, Chang S-E (2023). Brain developmental trajectories associated with childhood stuttering persistence and recovery. Developmental Cognitive Neuroscience, 60, 101224. https://pubmed.ncbi.nlm.nih.gov/36863188/

Project Description

• Participants ranged in age from 3:0 to 13:11
• 208 speech samples available from 120 CWS
• 186 speech samples available from 95 controls
• Longitudinal speech samples were collected at intervals of approximately 12–18 months.

This corpus contains speech and language samples collected across multiple studies conducted by the University of Michigan Speech Neurophysiology Lab under the direction of Dr. Soo-Eun Chang. These studies include A longitudinal study of neural network development in children who stutter and Neural oscillations underlying speech perception and production in childhood stuttering. These studies were supported by NIDCD grants R01DC011277 and R01DC018283 awarded to PI Chang. This corpus includes only samples for which a parent provided signed consent authorizing data sharing through TalkBank.

Collectively, these projects aim to advance understanding of the neural mechanisms underlying developmental stuttering. This work focuses on identifying how differences in brain structure, functional connectivity, neural timing, and genetic factors contribute to the onset, persistence, or recovery of stuttering.

The corpus includes both cross-sectional and longitudinal speech samples from children who stutter and non-stuttering peers. Participants were enrolled across early childhood through early adolescence, with enrollment ages varying by study. Some cohorts were recruited near the onset of stuttering (ages 3–6 years), while others were enrolled later in childhood (ages 7-13 years). Longitudinal datasets include between two and six assessment time points.

The types of speech samples collected vary across studies but typically include conversational speech with a speech-language pathologist, a narrative elicitation task using Frog, Where Are You? or Frog Goes to Dinner, and an oral reading sample when age-appropriate. For a subset of participants, conversational speech samples with a parent or guardian are also available.

All participants were evaluated by a licensed and certified speech-language pathologist (CCC-SLP). All participants demonstrated speech, language, hearing, and cognitive abilities within normal limits, with the exception of stuttering in the clinical group. A battery of speech-language and cognitive assessments was completed for each participant, and repeated each year of the longitudinal study.

Additional details about the studies included in this corpus can be found in the key publications referenced above.

This spreadsheet summarizes the available speech sample types for each participant at each data collection time point. It also includes each participant's age at every time point, providing an overview of the longitudinal availability of speech samples across the dataset.