Daily, we make thousands of decisions: latte or Americano? Brown shoes or black? Though it feels like our choices are made only with our conscious mind, more mysterious forces are at work. Blood flow in our brain indicates how we will decide several seconds before we’re aware we’ve made up our own mind.
Previous research studied motor decisions, predicting which hand someone would choose to push a button. In a new study in PNAS Early Edition Chun Siong Soon, John-Dylan Haynes, and colleagues expand their previous research, reporting that even abstract decisions can be predicted by unconscious preparation activity. They suggest our brains may hold a “starting point” common to different types of choices.
“We already showed that free motor decisions can be predicted by neural activity in prefrontal and parietal regions occurring before we consciously make up our minds,” said Soon.
To test if the same, semi-spooky phenomenon seen in motor could be found in complex decisions, the researchers asked test subjects to do some math. Participants were faced with a stream of frames on a computer screen refreshed every second. When a subject felt the urge to add or subtract they first noted a letter in the center of the screen, marking the time of their decision. They then added or subtracted the central numbers on the two following screens. They then reported their answer, touching a button to indicate where they found their answer on the next frame. Each frame held a number in each corner, two possible correct answers (one if the subject decided to add, one if they decided to subtract) and two false answers. Once they indicated their decision had been made, they were shown a frame with four letter options to report when they felt their decision had been made (see image).
The researchers found that about 4 seconds before participants were conscious of their choice, their decision could be decoded by brain activity spotted with a functional MRI machine.
“Our finding was not completely surprising,” Haynes said, “but it extended the previous work by showing that such early prediction of choices also holds for more abstract decisions.”
Arithmetic, he explained, involves manipulating arbitrary symbols according to mathematical laws. While animals make many motor-based decisions, “no animal can perform such proper symbol-based mental calculation.”
The predictive power of their model is above chance, the researchers say, but not complete. They think it’s possible that the early activity only indicates a certain decision bias, not the choice itself. Or perhaps the fMRI scans are too limited in their resolution. The next step, they say, is to implant electrodes.
“If we could record invasive signals from populations of neurons there might be much more information,” said Haynes. “This is on the way.”