How does the brain track fast-changing evidence during a decision?
Each new piece of visual evidence travelled through the brain in three stages, from visual to parietal to motor areas, with the posterior cingulate cortex involved throughout.
Source
Representation of Perceptual Evidence in the Human Brain Assessed by Fast, Within-Trial Dynamic Stimuli
Study at a glance
- Design
- Human experiment — Within-subject MEG reaction-time task: a single dot jumped every 100 ms and participants judged which of two targets was its centre; dot positions were regressed against sensor and source signals.
- N
- N=34 · 34 analysed adults (37 recruited; 1 dropped after training, 2 for eye artefacts/bad channels); each saw the same 480 trials.
- Population
- Healthy right-handed adults aged 20-35 from a Leipzig participant pool
- Outcome
- Trial-by-trial correlation of MEG signals (sensor and source level) with momentary evidence, accumulated evidence and choice
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
While their brain activity was recorded with MEG, 34 adults watched a white dot jump to a new position every 100 ms and decided as quickly and accurately as they could which of two targets (framed as beehives) it was centred on. The horizontal position of each dot was the decision-relevant evidence, while its vertical position served as an irrelevant control. The researchers correlated these dot positions, and their running sum, with MEG signals and with reconstructed brain sources.
What they found
Participants averaged 78% accuracy, and their choices followed the running sum of evidence (correlation around 0.7) far more than any single dot. Brain signals tracked each new dot in three phases: around 120 ms in visual cortex, around 180 ms in parietal cortex, and from roughly 300 to 500 ms in motor and premotor areas. The irrelevant vertical position was represented much more weakly in the late phase, and accumulated evidence was represented persistently in motor, premotor and posterior cingulate regions. The build-up before the response resembled these evidence signals, differing from the response itself mainly in stronger motor-cortex correlations.
The limits
What it doesn't show
The design used a pre-selected set of slow-response stimuli, so the pattern of dot influence partly reflects stimulus choice rather than natural decision behaviour. Momentary and accumulated evidence are mathematically linked, so their brain correlates cannot be fully separated. MEG source estimates have limited spatial precision, and the authors cannot rule out tiny eye or finger movements contributing to motor-area signals. The weak parietal accumulation signal may mean parietal neurons were too sparse to detect with MEG rather than that parietal cortex does not accumulate evidence.
Key terms
- Momentary evidence
- The information provided by a single new sample, here one dot position.
- Accumulated evidence
- The running total of evidence so far; in accumulation models a decision is made when it reaches a bound.
- Drift-diffusion model
- A model in which noisy evidence is added up over time until it hits one of two decision thresholds.
- Magnetoencephalography (MEG)
- A method that measures the tiny magnetic fields produced by brain activity with millisecond timing.
- Source reconstruction
- Estimating which brain regions generated the signals recorded by sensors outside the head.
- Posterior cingulate cortex
- A medial parietal region that here tracked evidence in all three processing phases.
Flashcards
0 of 11 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
In which order did momentary evidence appear across brain regions?
Common questions
Why change the evidence every 100 ms instead of showing a stable stimulus?
Rapid, controlled changes let researchers see how each individual piece of evidence is processed over time, rather than only an average build-up across the trial.
Why include the vertical dot position at all?
It is visual information that is irrelevant to the choice, so comparing it with the horizontal position separates plain seeing from using information to decide.
Does this show parietal cortex doesn't accumulate evidence?
Not conclusively. Parietal signals were brief and weaker than expected, but the authors note that MEG may simply miss a small population of accumulating neurons.
More on Perceptual decision