From structure to sequence: A multinomial processing tree model of syntactic encoding

Area of Science:

• Psycholinguistics
• Computational Neuroscience
• Neurolinguistics

Background:

• Sentence production models propose two stages: syntactic structure construction and linearization.
• Existing models lack formal computational implementations.
• Aphasia research highlights syntactic error patterns linked to distinct processing deficits.

Purpose of the Study:

• To introduce a novel computational model (multinomial processing tree) for the two-stage framework of grammatical encoding.
• To explain syntactic error patterns in aphasia using this computational model.
• To investigate the neural substrates associated with distinct stages of syntactic production.

Main Methods:

• Developed a multinomial processing tree (MPT) model to operationalize the two-stage syntactic production framework.
• Analyzed discourse samples from individuals with aphasia, annotating distinct error types.
• Fit the MPT model to estimate individual processing abilities at hierarchical and linear stages.
• Correlated estimated abilities with observed error rates and mapped them to neural substrates using lesion analysis.

Main Results:

• The MPT model successfully estimated individual abilities in hierarchical encoding and linearization.
• Ability estimates correlated significantly with observed syntactic error rates.
• Hierarchical encoding ability localized to the posterior superior temporal sulcus and parietal cortex.
• Linearization ability was linked to posterior inferior frontal regions.
• Omission-related processes mapped to dorsal frontal regions and white matter.

Conclusions:

• Findings support a neurocomputational dissociation between hierarchical and linear stages of grammatical encoding.
• The study provides the first computational instantiation of a two-stage syntactic production model.
• This work bridges formal modeling and lesion analysis to enhance understanding of grammatical encoding architecture and its breakdown in aphasia.