Speech processing requires extracting meaning from acoustic patterns using a set of intermediate representations based on a dynamic segmentation of the speech stream. Using whole brain mapping obtained in fMRI, we investigate the locus of cortical phonemic processing not only for single phonemes but also for short combinations made of diphones and triphones. We find that phonemic processing areas are much larger than previously described: they include not only the classical areas in the dorsal superior temporal gyrus but also a larger region in the lateral temporal cortex where diphone features are best represented. These identified phonemic regions overlap with the lexical retrieval region, but we show that short word retrieval is not sufficient to explain the observed responses to diphones. Behavioral studies have shown that phonemic processing and lexical retrieval are intertwined. Here, we also have identified candidate regions within the speech cortical network where this joint processing occurs.
This repository contains code used to perform the analysis for the paper:
[1] Gong, X., Huth, A. G., Deniz, F., Johnson, K., Gallant, J. L., & Theunissen, F. E. (2023). Phonemic segmentation of narrative speech in human cerebral cortex. Nature Communications, 14(1), 4309.
Dataset used in the paper can be found here
If you publish any work using the code in this repository, please cite the original publication [1], and cite this repository in the following recomended format:
[1] Gong, X., Huth, A. G., Gallant, J. L., & Theunissen, F. E.. Phoneme segmentation analysis. DOI: https://zenodo.org/badge/latestdoi/547471345
# clone the repository
git clone https://github.com/theunissenlab/phoneme_segmentation.git
cd phoneme_segmentation
python setup.py install
This repository contains a general pipeline for analysis of the publication [1]: