What do you think I am, a mind reader?
This is Sandra Tsing Loh with the Loh Down on Science, saying, well, not yet, anyway.
But, someday, UC Berkeley neuroscientist Brian Pasley just might be.
Pasley asked fifteen hospital patients being treated for epilepsy to be in a study. Because the patients already had a net of electrodes laid across their brains, their neural activity was easy to record.
Which is just what Pasley did as he played them a series of words.
Later, he matched the activity in the brain's STG region to information about the frequencies of the words.
Then he wrote a computer program to recreate such frequencies from brain activity.
The technique is called stimulus reconstruction. Its quality depends on making the right assumptions about which sound information--frequency, pitch, rhythm--the specific brain region being recorded is keying into.
Listen: The first word is what patients heard, and the second is the computer's reconstruction:
Not perfect yet. But when it is, maybe we'll find Waldo!
The Loh Down on Science, online, at lohdown.org. Produced by 89.3 KPCC and the California Institute of Technology, and made possible by TIAA-CREF.
Follow us on Twitter at LohDown.