Cortico-striatal language pathways dynamically adjust for syntactic complexity: A computational study

Area of Science:

• Neuroscience
• Computational Linguistics
• Cognitive Science

Background:

• Fronto-striatal circuits are increasingly recognized for their role in language perception.
• The striatum is implicated in attention, linguistic rule computation, and phonological short-term memory.
• A dichotomy of ventral semantic and dorsal phonological streams in spoken language processing may extend to cortico-striatal perception.

Purpose of the Study:

• To investigate the hypothesis that different parallel language pathways contribute separately to canonical and non-canonical sentence comprehension.
• To analyze a computational model of frontal-striatal syntax circuits to understand sentence processing.
• To quantify sentence comprehension abilities by examining model performance under varied parameters.

Main Methods:

• Utilized a modified computational reservoir computing model of language circuits, previously developed by Dominey and coworkers.
• Analyzed the model's performance on a thematic role assignment task under various parameter regimes.
• Manipulated parameters influencing language input decay and temporal dynamics of word representations.

Main Results:

• Modeling results indicate that separate cortico-cortical and cortico-striatal circuits might be differentially recruited for processing sentences of varying syntactic complexity.
• The study quantified abilities in comprehending both canonical and non-canonical sentences.
• Findings suggest distinct neural circuit recruitment based on syntactic difficulty.

Conclusions:

• Separate neural circuits, cortico-cortical and cortico-striatal, may be differentially involved in processing sentences with varying syntactic complexity.
• Alternatively, a single neural circuit might adapt dynamically to adjust for syntactic complexity.
• These findings contribute to understanding the neural basis of sentence comprehension and language processing.