This is a Python-based simulation project investigating how coupling to external sensory stimuli affects the ability of an intrinsic (neural-)oscillator model to track/process speech.
In the current project, we adapted the STiMCON model by Ten Oever and Martin (2021) (DOI: https://doi.org/10.7554/eLife.68066), which included an intrinsic oscillator network of word nodes unaffected by external inputs. To couple the intrinsic oscillator model with external sensory stimuli, we introduced the Stuart-Landau process, which allowed us to quantify the coupling strength with a scalar value K.
We found that weak sensory coupling/entrainment, compared to no or strong coupling, helps the oscillator better balance the use of internal language knowledge with the tolerance for unexpectedness in speech processing.
The repository consists of scripts belonging to simulations of the Stuart-Landau process, a new Oscillator module (which we used in STiMCON to accommodate the Stuart-Landau process), the adapted STiMCON model, and result generation & plots.
stuart_landau_STiMCON.py:
This script implements the Stuart-Landau process over a predefined time period.
oscillator.py:
This script defines an oscillator with parameters.
STiMCON_core_v4.py
Core script for the STiMCON model which has all the low-level code, adapted from Ten Oever and Martin (2021).
AdaptedSTiMCON_PredictiveFeedback_RhythmicInput.py
Implementing extended STiMCON when presenting isochronous input, and saving the data of feedback activation per word node.
AdaptedSTiMCON_PredictiveFeedback_RandomisedInput.py
Implementing extended STiMCON when presenting non-isochronous input, and saving the data of feedback activation per word node.
AdaptedSTiMCON_VarRhy_AmbiguousInput_Rhythmic.py
Implement extended STiMCON when presented with isochronous input, and save the results of ambiguous input categorisation across
stimulus onset delays, degrees of ambiguity, and stimulus frequency
AdaptedSTiMCON_VarRhy_AmbiguousInput_Randomised.py
Implement extended STiMCON when presented with non-isochronous input, and save the results of ambiguous input categorisation across
stimulus onset delays, degrees of ambiguity, and stimulus frequency
AdaptedSTiMCON_plotting_PredictiveFeedback.py
Analysing and plotting data of feedback activation per word node by implementing extended STiMCON with rhythmic (isochronous) and randomised (non-isochronous) sensory inputs.
AdaptedSTiMCON_plotting_AmbiguousInput.py
Analyse and plot data of ambiguous input categorisation implementing extended STiMCON with isochronous and non-isochronous sensory inputs
All the original scripts of Ten Oever and Martin (2021) used/called in the scripts above can be found: https://github.com/sannetenoever/STiMCON/tree/main.