ai4exceptionaled/Redmond-Sentence-Recall

Dataset Summary

The Redmond Sentence Recall (RSR) measures a child’s ability to repeat sentences that contain regular past tense forms and past participle forms (e.g., “He kicked” vs. “He was kicked”). This task helps identify language impairments, with each child repeating 16 sentences heard through headphones. The dataset includes anonymized audio recordings of these repetitions.

What makes the RSR dataset uniquely valuable is its focus on sentence recall using both regular past tense and past participle forms, which are key markers in diagnosing language impairments in children. This makes it a valuable resource for researchers and practitioners in speech and language processing, as it offers detailed, real-world data for studying sentence recall abilities, training machine learning models to detect impairments, and developing more precise diagnostic tools for clinical use.

Example Use Cases:

• Speech and language research: Analyze sentence recall and grammatical structure in children, comparing repetition patterns between those with typical language development and those with impairments.

• Speech processing models: Train and evaluate machine learning models to automatically assess sentence recall performance and detect language impairments.

• Clinical diagnostics: Support the development of diagnostic tools to identify language impairments, using sentence recall as a key indicator of language processing abilities.

If you prefer direct audio access, the .wav files can be also accessed at link.

Languages

English

Dataset Structure

We are releasing only the anonymized audio files to the public. The dataset includes a folder containing 1,058 audio files, each corresponding to a child's repetition of sentences in the task.

Source Data

The data was originally collected by Sean Redmond link and his research team. The collection focused on evaluating sentence recall as a diagnostic measure for identifying language impairments in children. A detailed manual outlining the data collection methodology and structure is available here.

Annotations

Annotations for the dataset, such as sentence-level transcriptions or metadata, are not being released at this time.

Personal and Sensitive Information

ll the data has been fully anonymized by Sean Redmond's group to protect the privacy of the participants. There are no personal identifiers in the released dataset.

Additional Information

Dataset Curators

The dataset is curated by Sean Redmond's group link and @xlab-ub link. Sean Redmond’s group anonymized and shared this dataset, and @xlab-ub's group prepared and curated it.

Licensing Information

This dataset is licensed under the CC BY-NC 4.0 license, which allows for non-commercial use with attribution. Please make sure to credit the creators when using this dataset in your work.

Citation Information

If you use this dataset in your research, please cite the following: Redmond SM, Ash AC, Christopulos TT, Pfaff T. Diagnostic Accuracy of Sentence Recall and Past Tense Measures for Identifying Children's Language Impairments. J Speech Lang Hear Res. 2019 Jul 15;62(7):2438-2454.

Contributions

For inquiries regarding the use of this dataset or its intended applications, please contact Jinjun Xiong link or Sean Redmond link.

Acknowledgements

This material is based upon work supported under the AI Research Institutes program by National Science Foundation and the Institute of Education Sciences, U.S. Department of Education through Award # 2229873 - National AI Institute for Exceptional Education link. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation, the Institute of Education Sciences, or the U.S. Department of Education.