How the Brain Accumulates Evidence to Make a Decision
Post by Natalia Ladyka-Wojcik
The takeaway
Intracranial electroencephalography was used in a large group of pre-surgical epilepsy patients to identify where in the brain perceptual decision-making begins and to show how brain activity might accumulate to support the strength and accuracy of decisions.
What's the science?
Perceptual decision-making involves selecting one choice from a set of alternatives based on incoming information from our senses. For example, a football player is engaged in perceptual decision-making when judging which player to pass the ball to during a game. When we make these decisions, our brain is also planning the corresponding motor actions, like the player gripping the football behind his head to throw it to his teammate. Recent research has identified parts of the brain involved in making these decisions, even in complex situations when the appropriate motor action might not be known in advance, but it remains unclear where exactly in the brain this process actually starts. This week in Nature Human Behavior, Gherman and colleagues pinpoint the signals in the brain responsible for these abstract decision-making processes.
How did they do it?
Previous research has found that brain activity linked to decision-making builds gradually over time, proportionally to the strength of incoming sensory information. As this evidence accumulates, it reaches a fixed threshold just before the animal makes a response and this threshold predicts both the animal’s choice and response time. In human neuroimaging research where resolution at the single neuron level is often not feasible, past studies have instead relied largely on functional magnetic resonance imaging (fMRI) to investigate brain regions involved in perceptual decision-making. However, fMRI has low temporal resolution, meaning that it doesn't show changes in brain activity quickly enough to demonstrate if there is a signal related to evidence accumulation. To overcome this limitation, the authors measured brain activity directly using intracranial EEG in a group of pre-surgical epilepsy patients who were asked to perform different perceptual decision-making tasks. This technique allows for measurement of a larger portion of the brain compared to single-neuron recordings and with better temporal resolution than fMRI.
First, the authors showed patients two simultaneous random-dot patches and asked patients to report the direction of the moving dots. The researchers measured the patients’ responses and the time it took to press a button to make their decision while recording their brain activity using intracranial EEG. Some patients also had to do the same task but with verbal responses after a delay, helping the researchers find brain activity related to perceptual decision-making, not just motor responses.
What did they find?
The authors found a widely distributed network of brain regions associated with perceptual decision-making, including the prefrontal cortex, parietal cortex, as well as inferior temporal and insular regions. Importantly, this network of brain regions showed high-frequency activity consistent with evidence accumulation in decision-making (whereas lower-frequency activity tends to be associated with motor preparation signals). This work suggests that activity in these regions gradually builds up until enough evidence is accumulated to make a decision (here, in response to the direction of dots) even before the patient presses a button or verbalizes their decision. The authors observed a gradual buildup of activity following the onset of sensory evidence at a rate that scaled with the strength of that sensory evidence. This activity reflected both the choice accuracy and response times of patients.
What's the impact?
This study found that activity in a distributed network of brain regions accumulates in response to sensory information to enable perceptual decision-making. It is the first to use intracranial EEG to directly measure high-frequency activity in these regions in humans, enabling an investigation into brain responses for decision-making even before motor response planning.