Project 4

Representation and processing of constructions in the brain

This research project focuses on unravelling how language is stored and processed in the human brain. It centres on the intriguing concept that linguistic structures could be represented as complex, multi-scale network-like maps within our neural architecture. Recent findings in neuroscience indicate that regions such as the hippocampus and entorhinal cortex play a role in creating cognitive maps, primarily aiding in spatial navigation.

To explore this idea further, we plan to leverage computational models inspired by machine learning methodologies. These models will attempt to simulate the intricate representations of language found in these cognitive maps. The aim is to create computational analogs that mirror how the brain processes linguistic information. By comparing the activity of these models with real brain activity data, we aspire to refine and improve their fidelity, aligning them more closely with the brain’s language processing mechanisms.

Additionally, our approach involves conducting comprehensive multi-modal measurements during continuous speech perception. Departing from the conventional isolated word or sentence studies, we aim to investigate how the brain comprehends language within more natural, connected contexts. This shift towards examining language processing in real-world situations, like listening to audiobooks, offers a richer understanding of how our brains handle language in everyday scenarios.

Ultimately, the overarching goal of this endeavour is to establish connections between neuroscience and linguistics by uncovering how language might be represented within the brain’s intricate network-like structures. This research could significantly contribute to our understanding of language processing mechanisms and pave the way for novel insights at the intersection of neuroscience and computational linguistics.