Savings in motor adaptation

Imagine a screen cursor starts moving to one side of where you aim. You practice compensating, then practice with normal feedback until the earlier adjustment fades. When the distortion returns, you may learn to compensate faster. What survived the apparent reset: a remembered aiming strategy, an automatic adjustment, or both?[1] [2]

In a cursor task, compensation for distorted feedback can fade during normal-feedback practice, yet return faster when the distortion is reintroduced. Which retained process produces this savings? Deliberately aiming to one side of a target and automatically recalibrating a movement are different routes to better performance. The latter is called implicit adaptation. In tasks designed to isolate it, people can adjust even when told to ignore misleading cursor feedback. In the 2026 rotation experiment, fMRI brain scanning tracked relationships among regions' activity during learning and relearning. A returning pattern was strongest in the default mode network, a set of regions associated with internally directed and memory-related processing. The pattern's return is called reinstatement.[1] [2]

Remembered strategies and automatic adjustment contribute differently. Their causal roles are not fully separated.[1] [2]

The 2026 network result associates reinstatement with savings; it does not prove that reinstatement causes retrieval or separate deliberate aiming from automatic recalibration. A 2025 task isolating implicit adaptation found weaker relearning after intervening feedback, whereas removing that feedback prevented the attenuation. Thus an overall relearning advantage need not mean that every adaptation process became stronger.[1] [2]

Returning to ordinary performance does not necessarily mean that earlier practice has been erased. Faster relearning can reveal a memory that was no longer obvious in the movement itself. Separating remembered strategies from automatic changes helps explain what an apparent reset or improvement actually tells us about learning.[1] [2]

Visible compensation can fade while a memory of earlier practice remains available. When the cursor distortion returns, reusing that memory can help a person regain the adjustment more quickly. This connects the current state of a movement with what past experience leaves available for future action. Recent conflicting feedback can also hinder automatic recalibration, so the benefit of prior practice depends on which learning process is involved. Separating these mechanisms would prevent a gain in total performance from being mistaken for stronger implicit learning.[1] [2]