Colloquium Cognitive Systems

Brain connections of words: A neurobiologically constrained model of semantic processing in sighted and visually deprived populations

Rosario Tomasello (FU Berlin)


Abstract. Understanding the meaning of words and their relationship with the outside world involves higher cognitive processes unique of the human brain. Despite decades of research on the neural substrates of semantic knowledge processing, a consensus about the functions and components of the semantic system has not been reached among cognitive neuroscientists. In the present talk, I seek to address the hotly debated questions of ‘how’ and ‘why’ specific brain regions are specialised in processing all types of meaning (i.e. semantic hubs) and why some of the brain’s semantic processes are category-specific to semantic types such as animals, tools or actions and how these semantic mechanisms are instantiated in case of deprivation, for example in the blind, where the visual cortex is recruited for language.
In the present talk, I will demonstrate that through the mutual effect of learning, cortical semantic areas and connectivity structure are sufficient to provide a direct and straightforward explanation to these questions and can explain the full range of semantic area activations seen in neuroimaging studies. A novel neurobiologically constrained model of spiking neurons that resemble the human cortical function is applied to describe the putative neural mechanisms at the cellular/synaptic level underlying word learning and semantic grounding in visually deprived and undeprived brains. The simulations demonstrate that meaningful linguistic units are represented in the brain in the form of cell assemblies that are distributed over both semantic hubs and category-specific regions spontaneously emerged through the mutual interaction of a single set of biological mechanisms acting within specific neuroanatomical structures. These biological principles acting together lead also to the spontaneous functional recruitment of the visual cortices for language processing in case of visual deprivation, as reported by a range of neuroimaging studies.

Bio. Rosario Tomasello´s main interest is on the neural basis of understanding and performing communicative action in the context of speech, gestures and intonation, taking into account social interaction, and common ground. Critical questions are: Which cortical areas are relevant for communicative acts processing and when in time they first emerge in the brain, and how are these cognitive processes instantiated in deprived conditions, such as aphasia and language impairments. Rosario is also interested in how different semantic word categoriesare processed, stored and acquired in undeprived and deprived populations.
From 2014 to 2016, he held a research assistant position in Neurocomputational Modelling of Language Learning for the interdisciplinary BABEL project between Plymouth, Manchester and Berlin universities; 2016-2019 Rosario held is PhD at the Brain Language Lab (Freie Universität Berlin), which was founded by the Berlin School of Mind and Brain (Humboldt Universität zu Berlin), where he also completed the doctoral M&B program. Currently, Rosario is holding a Post-doc position as a research fellowship on the Xprag DFG founded project - Brain Signature of Communications