Explaining savings for visuomotor adaptation: linear time-invariant state-space models are not sufficient

Area of Science:

• Motor control and learning
• Neuroscience
• Human-computer interaction

Background:

• Motor adaptation is crucial for tool use and adapting to bodily changes.
• Motor learning memory can manifest as savings, accelerating relearning after a perturbation.
• Assessing savings is simplified when initial sensory errors match during adaptation and re-adaptation phases.

Purpose of the Study:

• To investigate the theoretical prediction that a two-rate, linear time-invariant state-space model (SSM(LTI,2)) explains savings for the counterperturbation (CP) paradigm but not the washout (WO) paradigm.
• To empirically test the SSM(LTI,2) model's ability to explain savings in visuomotor rotation tasks under both CP and WO conditions.
• To explore the underlying mechanisms of motor adaptation savings, particularly whether it solely relies on linear time-invariant dynamics.

Main Methods:

• Utilized a planar reaching task with visuomotor rotation perturbation.
• Employed both counterperturbation (CP) and washout (WO) paradigms to assess motor adaptation savings.
• Applied a two-rate, linear time-invariant state-space model (SSM(LTI,2)) for theoretical and empirical comparison.

Main Results:

• Found comparable savings in motor adaptation for both CP and WO paradigms.
• The SSM(LTI,2) model successfully explained savings for the CP paradigm to some extent.
• The SSM(LTI,2) model failed to explain savings for the WO paradigm, indicating limitations in capturing the full phenomenon.

Conclusions:

• Savings in visuomotor rotation adaptation are not solely attributable to linear time-invariant dynamics.
• Motor adaptation savings likely involve meta-learning processes, characterized by changes in system parameters across experimental phases.
• The findings challenge existing models and suggest a more complex, adaptive mechanism underlying motor learning memory.