Is the brain able to capture a new temporal relationship between a motor action and its consequence?

This study investigated how the brain can learn new temporal relationships between the predicted and actual consequences of a descending motor command. We chose a long and a short delay to assess cognitive and sensorimotor adaptation, respectively. Cognitive adaptation, in contrast to sensorimotor adaptation, would use higher cognitive mechanisms requiring attention and working memory associated with temporal processing. A horizontal cable, attached at pelvis level, was used to suspend a load that tended to move participants backwards. In experiment 1, the participants (n = 10) controlled unloading of the cable, which induced a forward destabilization. Three blocks of trials were performed. The first (PRE) and third (POST) blocks contained 30 trials each. For these conditions, the mechanical consequences of cable unloading were directly transmitted to the participants. The second block (DELAY) contained 60 trials. In this condition, a constant 600-ms delay was introduced between the motor action and the mechanical consequences of unloading. To determine whether the duration of the delay or the number of trials could explain the absence of prediction, we tested eight different participants (experiment 2). These participants first performed 15 imposed unloading trials (IMPOSED). In this condition, the experimenter controlled the onset of the unloading. Then, the participants were exposed to 120 self-triggered trials; the mechanical consequences of cable unloading were directly transmitted to the participants. Finally, four participants were exposed to a delay of 300 ms, whereas the other four were exposed to a delay of 600 ms. All the participants performed 120 delayed trials. Results of the first experiment revealed that when the participants controlled the timing of unloading and there was no delay, they activated their gastrocnemius (GM) muscle before unloading (120 ms and 126 ms for PRE and POST condition, respectively). When the unloading occurred 600 ms later, however, participants did not anticipate the unloading; their GM muscle onset followed the unloading. Across trials, however, they adapted their balance strategy as they decreased the activity of their tibialis anterior, and GM muscle onset occurred earlier than during imposed unloading. Nonetheless, overall participants succeeded in decreasing their peak center of mass velocity and center of mass and center of pressure ranges. Results from the second experiment showed that GM muscle onset occurred earlier than during imposed unloading for both delayed unloadings. Moreover, data for the shorter delay (i.e., 300 ms) showed that participants adapted their balance strategy response; their GM muscle onset occurred before the delayed unloading. Although, increasing the number of 600 ms delayed unloading (i.e., 60 to 120 trials) did not allow the brain to anticipate the unloading. In summary, the results of the present study suggest that the brain could adjust to some extent its balance control strategy when the temporal relationship between the motor action and its consequence was delayed. Nonetheless, cognitive adaptation would be limited compared to sensorimotor adaptation.