The purpose of this study was to examine the impact of altered proprioception and reduced visual feedback on continuous motor sequence learning. A large body of literature demonstrates that information regarding body state is crucial for motor control.[4
] In this study, we sought to determine whether this was also the case for motor sequence learning. In the past, Rothwell and colleagues [34
] suggested that motor learning might be deleteriously impacted by absent proprioception via their case report of a deafferented individual. In this case, the authors reported that learning new, complex sequences of hand movements was difficult when deafferentation was present. We wondered if similar negative effects would be present for continuous motor sequence learning when proprioception was shifted by vibration. We discovered that in the short-term, peripherally altered proprioception and reduced visual feedback impacted motor performance; however, given two days of practice, the extraordinarily robust human motor learning system was able to overcome the challenge presented by shifted proprioceptive sensation and motor learning of a repeated continuous sequence occurred.
To our knowledge the present study represents the first experimental investigation of the impact of altered proprioception on continuous motor sequence learning. The experimental design employed for the current work differs from previous studies in several important ways. First, we used a continuous tracking task that required participants to use their entire upper extremity to produce movement. It has been suggested that investigation of complex movements (i.e. those that involve more degrees of freedom or greater muscle activation) are critical for understanding motor learning and behavior.[35
] Our task met those criteria by including multiple joints, greater movement excursion and longer movement patterns. Also, previous studies of proprioception have often employed discrete, reaching-type tasks for which participants likely already have at least a rudimentary motor plan.[2
] Our use of an entirely novel continuous tracking task allowed us to more fully examine novel motor sequence learning. Finally, we engaged individuals in two days of practice and a separate, delayed, retention test. In this manner learning versus performance improvements were clearly differentiated.[8
] Because no prior studies of the role of proprioception have employed a retention test design, it has not been clear whether altered proprioception would deleteriously impact motor learning.[2
We hypothesized that if veridical proprioceptive sensation was essential for sequence learning, peripherally altered proprioceptive information that did not reflect the true state of the limb would diminish both acquisition and retention of the repeated motor sequence. We discovered that the opposite was true; all participants were able to learn sequence-specific regularities as compared to random epoch performance. The finding that individuals can learn to accurately and continuously track a repeating sequence even when vibration was applied to the arm being used suggests that accurate and intact proprioception are not absolute prerequisites for encoding and consolidating movement regularities. We found that the group that practiced with altered proprioception (AV group) and minimal visual feedback was able to improve in the same manner as the group who experience only control vibration to the non-tracking limb. Additionally, we found that when vibration was re-introduced at retention, motor learning in the AV group was masked by altered performance; this effect was not observed for the CTL group. These findings were facilitated by our experimental design; had we stopped data collection after 1 day as past work has done we would not have noted the positive effect of task practice in overcoming altered proprioceptive feedback.
Cordo and colleagues [33
] have suggested that the dynamic position and velocity information supplied by proprioceptors may be important for the execution of movement sequences. Based on this, we posited that proprioception would also be critical for learning the spatio-temporal regularities of a repeated continuous sequence. Rather, we found that accurate proprioceptive information was not essential for learning our experimental task. Nor was continuous visual feedback. Recent work has reported that motor sequence learning can occur in a range of other environmental experiences. Overduin and colleagues [38
] demonstrated that sequence learning occurred independently of learning predictable shifts in the dynamic environmental state. Our work supports and extends these findings to show that motor sequence learning can occur despite changes in visual and peripheral proprioceptive information.
Simply becoming aware of the repeating sequence is one possible reason that the AV group was able to learn the continuous tracking task. It is certainly possible that observation of target movement was sufficient to stimulate learning. Indeed, sequence learning has been demonstrated following stimulus observation alone [39
] especially when individuals attend to the task.[40
] In accordance with these findings we cannot rule out the possibility that untrustworthy proprioception was compensated for by paying greater attention to target motion.
Another plausible explanation for our finding that altered proprioception did not diminish learning may be that accurate afferent sensation from more distal segments of the arm might have been preserved and exploited. We cannot totally rule out this possibility with the present experimental setup. Single joint elbow muscles as well as wrist and finger musculature, joint and cutaneous afferents were possibly spared from vibratory disruption (though several subjects reported "numbness and tingling" into the forearm and wrist). Furthermore, secondary spindle afferents appear to be relatively insensitive to vibration.[15
] The central nervous system could have preferentially attended to these signals for information regarding performance.
However, we suggest that the hypothesis outlined above cannot completely explain our results because this same "unaltered" afferent information did not overcome vibration-induced changes as shown by the limb position matching task in Experiment 1 or during reintroduction of vibration at Experiment 2 retention testing. These findings supply convergent evidence that vibration was disruptive to motor control. Based on these findings, it appears that vibration induced at least some shift in the afferent feedback from the shoulder and elbow spanning musculature to the central nervous system that altered motor output. Motor sequence learning appears to have occurred despite this shift in the veracity of limb proprioceptive sensation.
It has been previously noted that vision is critical when proprioceptive sensation is diminished or absent[34
] Ghez et al. [31
] reported that individuals with large fiber sensory neuropathy improved their aim on discrete reaching tasks when able to visualize arm position before movement. To explore the contribution of proprioception without the confound of visual feedback, we reduced visual information available to the participant via several controls. First, we occluded vision of the arm via draping. Next, we quickly faded feedback regarding cursor position over the first 20 trials to an intermittency exceeding that which Kao [27
] cited as being disruptive to continuous tracking. However, we chose to preserve some visual feedback to reduce cumulative error which might have obscured improved motor control associated with learning [43
] by displaying the arm position cursor for 200 ms at 1800 ms intervals. It is possible that even this minimal visual information may have allowed participants to evaluate their performance and adjust accordingly in the absence of trustworthy proprioceptive feedback. However, based on the past work of Kao [27
] we find this explanation of our conclusions highly improbable.
Our finding of preserved continuous sequence learning despite restricted visual feedback and altered proprioception reflects the dynamic and robust nature of a motor learning system that is able to compensate for inaccurate afferent information through redundant physiological and cognitive systems. One or some combination of all of the mechanisms proposed above may have facilitated learning for participants in this research. Though these findings do not directly support our original hypotheses that altered proprioception would disrupt motor sequence learning, they are not without precedent. Skill learning has been reported in dorsal rhizotomized monkeys.[44
] The juxtaposition between our findings and Taub et al.'s are in contrast to reports by others,[46
] who have reported disruption of skill learning following sensoricortical damage. These seemingly contradictory results may be a function of the difference between central and peripheral neural damage/disruptions. It remains to be seen if those with chronic sensory impairment resulting from damage to central sensory cortical or thalamic regions have difficulty learning new motor skills. Future work should consider this possibility in persons with medical conditions characterized by reduced proprioception.