How are thoughts and memories stored in our brains? How is new information integrated with existing information in our head? We don’t have the answer yet, but scientists are discovering that information is broken down and stored as a collection of basic ‘sub-concepts’, and the degree of overlap with existing sub-concepts leads to differences in how we organize information in our heads.
A Memory as a Collection of Smaller Ideas
The current model in neuroscientist circles- called the ‘distributed memory model’- is that a concept or an idea is stored as a collection of its attributes. Thinking about an idea, then, requires activation of each of the attributes associated with it. This is why thinking of one particular thing conjures up memories of other, similar things- there is simply a large overlap between the attributes of the two objects. You’re more likely to think of ‘soccer’ and ‘cricket’ in the same breath because of the many similarities inherent in how we understand them- sport, balls and, well, bad English teams.
Memory Recall by Clusters of Objects
Taking this one step further, scientists at the University of Pennsylvania and the Thomas Jefferson University have tried to look at memory recall to study how people store memories which are similar to each other. They plugged electrodes onto the brains of participants and gave them a list of words to read, recording patterns of electrical signals in the brain as the list was read. Upon being asked to recall words from the list, electrical recordings were taken again. The key was to see if participants tended to throw up words which were similar in meaning to each other rather than by the order in which they had been seen. And indeed, this was the case.
Brain Signals Encode Word Meanings
As participants were shown words, the researchers used mathematical techniques to map the part of the brain signal which could represent semantic components- the parts actually representing the meaning of the word. Similar components which were mapped upon observation of a word were shown to be activated when the same word was recalled. And moreover, the closer the overall pattern of the brain signal was to these semantic components, the more likely was the participant to remember words in order of similarity to each other (rather than by the order in which they were read).
Acquiring new knowledge involves integrating new information into our existing scheme of knowledge, and such maps are a way of determining exactly how we go about doing that. Perhaps this could also eventually explain how some of us are such bad quizzers- out brains just aren’t very good at plotting signals onto maps.
Also, the pattern of signals is reasonably well-conserved across different individuals. This means that there is a possibility of plotting out a map of ‘standard’ brain signals which could be used to predict thoughts which are still a few micro-seconds away. Going beyond this potential ‘mind-reading’, stimulating certain regions of the brain simultaneously could actually implant thoughts in a brain. Hypnosis has a potential competitor.