Prediction error is the brain's measurement of the gap between what was expected and what actually happened. It is not a generic "reward" signal — it is a delta. In the ONDA model two systems live on this signal: the Ventral Tegmental Area (motivational salience) and the anterior cingulate cortex (conflict and error monitoring).
Three States
- •Match — outcome equals prediction, baseline dopamine, no learning trigger
- •Positive error — reality exceeds the model, dopamine spike, pathway strengthens
- •Negative error — reality undershoots the model, dopamine dip, pathway weakens
Why It Matters
The ACC uses prediction error to flag conflicts (focus task vs. easy distraction), run cost–benefit on cognitive control, and trigger recalibration when drift from the protocol is detected.
In ONDA Life
Both the VTA Reactor and ACC Calibration protocols treat prediction error as a tunable signal — by controlling the input cadence (notifications, sugar, deep-work blocks) you control how the delta is generated and learned from.