Ever wonder how the brain does it? How do we take in so much information and somehow organize it into something useful?
Scientists at the University of California, Berkeley, have successfully created the first map illustrating how the brain categorizes the multitude of images received through our eyes daily. Through computational models using brain imaging data from subjects watching video clips, researchers have unveiled a ‘continuous semantic space’ mapping how the brain organizes categories of objects and actions across the cortex.
The study shows distinct areas of the brain handling specific categories of objects observed in our surroundings. While some category relationships are intuitive, such as humans and animals sharing a ‘semantic neighborhood,’ others like hallways and buckets show less obvious connections. Interestingly, despite individual differences, a similar semantic layout was observed among different people.
Utilizing functional Magnetic Resonance Imaging (fMRI), researchers monitored brain activity in five individuals watching video clips to analyze correlations and construct a model demonstrating how different cortical subdivisions respond to 1,700 categories of objects and actions.
Through sophisticated analysis techniques, the researchers mapped around 30,000 locations within the cortex to show how different categories of objects and actions are organized in the brain. The study revealed that the brain organizes categories like “humans” and “animals” in related areas, while unrelated categories like “eyeball” and “car” are stored separately. This innovative approach not only enhances our understanding of brain organization but also holds potential for applications in brain disorders diagnosis and treatment, as well as improving image recognition systems[4][7].
Categories are represented as smooth gradients covering much of the brain’s surface. Contrary to previous beliefs of distinct brain areas for each category, this study reveals that brain activity organization is based on category relationships. This efficient representation of diverse categories in a compact space suggests that not every category requires a separate brain area[1]. The brain uses various mechanisms to map information across its surface, indicating that categorization tasks engage multiple systems and are not confined to a single brain area[2].
An update from August 2019 shows the ability to determine what a person is reading based on what brain areas are lighting up.
UC Berkeley neuroscientists have unveiled groundbreaking interactive brain maps that reveal how reading or listening to classics like Melville’s Moby Dick or Tolstoy’s Anna Karenina stimulates similar cognitive and emotional brain regions. These maps, detailed in the Journal of Neuroscience, offer insights into the brain’s response to stories and words. By arranging words based on their semantic relationships, researchers could accurately predict which words activate specific brain regions. This innovative mapping technique covers a significant portion of the cerebral cortex, shedding light on how different categories of words trigger distinct brain areas. This research not only deepens our understanding of language processing but also holds promise for applications in learning, speech disorders, and interventions for conditions like dyslexia and auditory processing disorders.
One can suppose that it will long be much easier to read from the brain than to write to the brain. Once we can have direct write access to the brain, however, in the style of the Matrix, perhaps we can upload experiences or new skills. Imagine 10 years of school downloaded in just a few sessions at the local Brain Improvement Booth! It’s daunting to consider what kind of a world this might create.
Citations
[1] https://www.eurekalert.org/news-releases/832404
[2] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3709834/
[3] https://tech.cornell.edu/news/ai-generated-images-map-visual-functions-in-the-brain/
[4] https://www.extremetech.com/extreme/143816-scientists-discover-how-our-brains-categorize-map
[5] https://www.istockphoto.com/photos/brain-mapping
[6] https://www.nature.com/articles/d41586-023-03192-2
[7] https://news.berkeley.edu/2012/12/19/semanticspace
