Motor sequence learning and intermanual transfer with a phantom limb.

Amputees with phantom limb sometimes report vivid experiences of moving their phantom. Is phantom movement only "imaginary", or, instead, it has physiological properties comparable to those pertaining to real movements? To answer this question, we took advantage of the intermanual transfer of sequence learning, occurring when one hand motor skills improve after training with the other hand. Ten healthy controls and two upper-limb amputees (with and without phantom-movement) were recruited. They were asked to perform with the right (intact) hand a fingers-thumb opposition sequence either in Naïve condition or after an active (Real condition) or a mental (Imagery condition) training with the left (phantom) hand. In healthy controls, the results showed different effects after active training (i.e., faster movement duration (MD) with stable accuracy) and after mental training (i.e., increased accuracy with stable MD). Opposite results between moving-phantom case and static-phantom case were found. In the Real condition, after an "active" training with her phantom hand, the moving-phantom case showed a faster performance of the intact hand. This transfer effect was not different from that found in healthy controls, actually performing the active training with an existing hand (Real condition), but, crucially, it was significantly different from both Imagery and Naïve conditions of controls. Contrariwise, in the static phantom case, the performance during the Real condition was significantly different from the Real condition of healthy controls and it was not significantly different from their Imagery and Naïve conditions. Importantly, a significant difference was found when the transfer effect in Real condition was compared between the two phantom cases. Taken together, these findings provide the first evidence that a phantom limb can learn motor skills and transfer them to the intact limb.