Data Science Investment Counter —Funding raised by UK data science companies in 2018.
£ 5.640 Million

Artificial intelligence shows how the brain understands objects in our world

Credit: Sasa L. Kivisaari

The brain’s ability to reconstruct facts from clues has been observed using brain scanning and artificial intelligence (AI) by researchers at Aalto University.

The method could be used to address the question of why people understand or perceive the same concept differently.

‘The organisation of meanings in the brain differs from person to person and can also affect how easy or hard it is for them to understand each other,’ says Riitta Salmelin, Professor at the university.

The research results may also play a role in detecting memory disorders.

‘Combining and understanding meaningful information seems to involve the same brain areas that are damaged in early Alzheimer’s disease. Therefore, the method we use may also be applied to the early detection of memory disorders,’ said Sasa Kivisaari, Postdoctoral Researcher at the Aalto University.

Published today in Nature Communications, the study analysed the brain reactions of test subjects after they were given three clues to help them guess what familiar objects the clues described.

They depicted well-known animals, vegetables, fruits, tools and vehicles. The familiar objects and concepts described in the clues were never presented directly to the test subjects.

The tests revealed that brain activation patterns contained more information about the features of the concept that had been presented as clues. This brought the researchers to the conclusion that the brain uses environmental clues in an agile way to activate a whole range of the target concept’s properties that have been learned during life.

The Aalto university program scanned an enormous amount of internet-based material to map the meaningful features associated with different concepts. Machine learning was used to create a model that describes the relationship between these features and brain activation patterns.

Based on such a model, brain activation patterns could be used to accurately deduce which concept the test subject was thinking of. For instance, the activation patterns could be used to infer whether the clues led the subject to think of a snake or a moose.

‘This is a very important skill in nature because it enables a quick response based on a small amount of information. For example, we automatically avoid a wiggly thing on a rocky shore because we know that a snake may be poisonous’.

Professor Riitta Salmelin’s research team studies the neural basis of processing of language and meaningful information at the Department of Neuroscience and Biomedical Engineering at Aalto University.


Co-working space and blog dedicated to all things data science.

Subscribe to our newsletter