How Brain Signals Track Abstract Decision-Making Beyond Sensory Input

A team at the University of Nevada, Reno has isolated a brain signal that tracks evidence for decisions built on arbitrary, newly learned rules — not just sensory input. According to PsyPost, researchers Arianna Thoksakis and Edward F. Ester tested whether the centro-parietal positivity, a scalp-measured voltage ramp long tied to evidence accumulation, generalizes beyond fixed visual categories. For circuit labs, the result reframes a single EEG marker as a flexible readout of decision-state dynamics rather than a stimulus-locked one.

The paradigm: rule without a cue

Thoksakis and Ester built a categorization task around hidden, participant-specific boundaries. Each volunteer viewed circles packed with hundreds of parallel lines and sorted them into two groups by pressing keys, receiving feedback after every trial. Each participant got a unique, invisible cutoff — for example, a 73-degree dividing line. The rule had to be inferred through trial and error, which means the evidence being accumulated is internally computed against a learned standard, not read directly from the visual field.

The signal under test

The readout is the centro-parietal positivity: a slow ramp in scalp voltage that climbs while a decision forms and peaks just before the response. Prior work has linked it to sensory evidence, to memory retrieval, and to universally shared visual categories. The open question was whether the same ramp would track evidence generated against a rule invented moments earlier.

Across 38 participants, the ramp tracked performance on the arbitrary-boundary task. The signal scaled with how much evidence each participant had accumulated relative to their personal cutoff, not just stimulus features. In operational terms, one electrical signature is reused across an open-ended class of decision computations — the marker is rule-locked, not stimulus-locked.

Bench checklist

If you replicate or extend this, isolate three parameters first:

• Boundary invisibility: confirm via post-task debrief that the participant never extracted the rule from the stimulus geometry itself.
• Feedback cadence: trial-by-trial feedback is required to drive rule acquisition. Log the per-participant learning curve; flat curves mean the internal reference never formed.
• Artifact control: the positivity lives in centro-parietal channels where eye-movement drift is a known contaminant. Apply slow drift correction and reject blinks before averaging.

The work appears as a preprint in The Journal of Neuroscience. For model-organism labs, the takeaway is methodological: a single scalp-level marker may index decision-state across rule types, which sets up cross-species comparisons of how flexible a circuit needs to be to support arbitrary categorization.