Cortical Representation of Abstract Semantic Knowledge

It is very useful to know things: That Simmons et al (2010), used a Riser detection task as their control condition or that my daughter’s school will be closed on the 25th of April because it was the day Italy was reunified. This form of knowledge has a profound impact on our ability to function; both in our daily and professional lives, yet we know very little about how it is manifest by the brain. CRASK is a five-year ERC-funded project to understand how the coordinated activity of our semantic system creates this complex factual knowledge of the world. It will do this by breaking down the way our brains combine different conceptual elements into increasingly complex peices of knowledge.

A primary focus of this project is the relationship between the brain's core semantic system and other cortical regions which contribute further knowledge. The brain's semantic system is a network of regions more active when we engage with semantically richer stimulus (Binder et al, 2009) and that contains supramodal representations of semantic concepts (Fairhall and Caramazza, 2013). Outside of this system, additional brain regions that contain selective representations relating to specific categories or domains of knowledge.

Phase one: Our rich knowledge about other people.

In the first phase of CRASK, we explored how different elements of the semantic system contribute to diverse flavours of world knowledge. We used a special class of stimuli, other people.

Seeing and thinking about other people produces a strong response in our own brains. But this is not the only reason other people make great stimuli, they are also associated with a rich wealth of knowledge. In one study, we have people make judgments around five different types of information, biographic, social, nominal, physical and episodic knowledge, that we can know about other people. We found that, rather than one elements being responsible for one function, we see that different types of access are associated with different patterns across the network (figure, left) and, that these subtle modulations in the core semantic system are accompanied by the recruitment of different brain regions depending on the type of information that is being accessed (figure, right). 

Subtle variations underlie varied semantic access in the core semantic system (left), which is accompanied by robust changes in cortical activity outside of this network (right). (Aglinskas and Fairhall, 2020).

Phase Two: How do we combine concepts into knowledge?.

Multivariate pattern analysis of the selective response in the left Insula (Fairhall, 2020)

What can you learn about the brain by having Italians think about food? It turns out, a surprising amount.

Italians not only love the taste of food, a feature typical of the object, but also associate food with a rich wealth of cultural knowledge.
This knowledge was used as part of the next stage of CRASK, where we expand our research beyond knowledge about people to knowledge in general. We found that when people read words and think about the taste of a well-known food dish, we see activation in a part of the brain that is associated with the experience of taste, called the insula. The study found that this region continued to be active when people accessed cultural geographic knowledge about food – here the region of Italy from which the dish originated – and was now accompanied by activation in place selective regions of the brain. Thus, the brain region associate taste are active when we think about food – continuing even when taste is completely irrelevant to the thought process - potentially contributing to the richness of our everyday experience as we think. At the same time, when we access other, atypical forms of knowledge, such as the geographic origin of a food dish, the brain recruits additional regions specialised for that knowledge. In this way, the human brain can contain it’s rich and diverse knowledge of the world by flexibly combining representations in brain regions typical and atypical for that category of concept. This study is one of a range of studies in this phase two, where we look at how we combine different conceptual elements into the diverse and complex knowledge that we posses.  

Phase Three: Complex Knowledge in the Real World.

This work is motivated by the importance of rich semantic knowledge in modern life. In the final stage of CRASK, we have moved to aspects of semantic knowledge that impact upon daily life. We are examining the brain dynamics associated with transient failures in semantic access, the successful versus unsuccessful learning of new semantic knowledge, those differences between individuals which make us better or worse at knowing things and changes in the functioning of the representational system associated with later life.