Master Thesis or Semester Project: Neural Mechanisms of Syntax

Background

The human brain possesses a remarkable ability to encode and process arbitrary rules, a capacity fundamental to language comprehension and production. Syntax, the set of principles governing sentence structure, represents a prime example of this capacity. This project aims to investigate how the brain encodes and implements syntactic rules, leveraging unique neural data and computational modeling.

 

Project Description

This project offers the opportunity to analyze high-quality intracranial neural recordings (local field potentials and single/multi-unit activity) obtained from human participants performing a controlled experiment designed to isolate neural activity related to syntactic processing. In this experiment, participants utter sentences that can correspond to two different meanings, depending on the selected syntactic manipulation. This allows us to specifically investigate how the brain encodes and processes changes in syntactic structure, while holding the acoustic and semantic content relatively constant.

 

Methodology

In this project, you will explore how neural activity (at LFP level, and if time allows, also at the single/multi-unit level) changes as participants switch between different syntactic interpretations of the same sentence. You will also develop simple computational models to generate testable predictions about the underlying neural processes involved in syntactic processing.

The project will use existing intracranial recordings (LFP and single/multi-unit activity) from five participants, with more recordings anticipated. The data that are already available have been largely preprocessed and are mostly ready for analysis. The data analysis involves examining changes in neural activity related to syntactic processing using statistical and decoding methods. The computational modeling involves developing simple models based on vector symbolic architectures to generate and test hypotheses about underlying mechanisms.

 

Requirements

- Proficiency in Python is essential.

- A solid foundation in mathematics (linear algebra) for understanding and implementing the analytical methods and computational models.

 

Contact

- Prof. Timothée Proix: proix@ini.ethz.ch

- Dr. Piermatteo Morucci (UNIGE)

- Prof. Nina Kazanina (UNIGE)

 

Starting date and duration

This project is currently available as a semester project or master thesis.

 

Related literature:

- Nelson et al. (2017). Neurophysiological dynamics of phrase-structure building during sentence processing. PNAS 114(18). http://www.pnas.org/lookup/doi/10.1073/pnas.1701590114

- Calmus et al. (2020). Structured sequence processing and combinatorial binding: neurobiologically and computationally informed hypotheses. Philosophical Transactions of the Royal Society B: Biological Sciences 375(1791). https://royalsocietypublishing.org/doi/10.1098/rstb.2019.0304