Programs that employ task-specific training can do so for a variety of reasons. In the context of CIT, practice is designed to reverse the sub-acute conditioning that leads to decreased spontaneous use of an extremity, referred to as "learned-non-use" (Taub et al., 1994, 2003). By contrast, a training program for patients with diminished motor control and impaired functional ability is designed to promote skilled performance (Dean and Shepherd, 1997; Sunderland and Tuke, 2005; Winstein and Prettyman, 2005).
CIT protocols grew out of the behavioral model put forth by Taub and colleagues (1994) in which it is proposed that during the early post-injury phase (e.g., deafferentation), use of the limb is suppressed when spontaneous attempts to move it are unsuccessful (negative reinforcement). This conditioned response is "learned" and ultimately results in diminished spontaneous use. The design of CIT protocols is therefore directed towards the reversal of learned-non-use and the increase of spontaneous use of the hemi-paretic limb in individuals post-stroke. Since the goal of CIT is to promote spontaneous hand use and not necessarily to develop skilled use, the conditions of practice are designed directly from operant-conditioning principles and include the "shaping" procedure. With shaping, a behavior is progressively modified towards the goal through successive approximation and positive reinforcement. In contrast to a motor-learning-based approach, the "shaping" procedure as described within the context of CIT (Morris et al., 1997; Taub et al., 2003), does not address the known resource impairments of motor control, strength and coordination (Sunderland and Tuke, 2005).
Skinner (1968) taught us that shaping was a form of operant conditioning in which the probability of experimenter determined behaviors are "elicited" through reinforcement (reward or punishment). Using this procedure he shaped pigeons to peck a ping-pong ball over a net. Obviously, the pigeon is not aware that this is a game-like, goal-oriented behavior. In fact, the learner (i.e., pigeon in this case) is relatively passive in this process while performance is progressively "shaped" towards the behavioral objective (task goal) in small steps through reinforcement or reward (positive feedback).
The shaping procedure is designed around the elicitation of behavior and not the acquisition of a voluntary skill. In fact, the pigeon, or any animal, can be shaped without knowing or ever understanding the goal behavior. The shaping procedure stands in sharp contrast to the procedures employed when designing task practice to optimize motor skill learning in the context of neurorehabilitation. For skill acquisition, the learner practices under a set of active learning principles that are derived from more modern theories of learning and memory (Cahill et al., 2001) such as those reviewed in Chapter 2 ofVolume I. For example, an operant-conditioning model treats "augmented feedback" as a form of "reinforcement" or reward, while a skill-learning model treats "augmented feedback" as information about performance for cognitive processing (e.g., problem-solving) relevant to the preparation for the next practice trial.
If augmented feedback operates like positive reinforcement, designing practice with frequent rewards should enhance learning within an operant-conditioning-based approach. In contrast, if augmented feedback operates like post-response information that elicits cognitive processing and problem-solving, designing practice with a faded feedback schedule, where feedback is provided on progressively fewer trials, should enhance learning within a motor-learning-based approach. Table 7.1 compares and contrasts training principles derived from each of these two learning models (operant conditioning and voluntary skill) as they apply to the choice of task practice variables to enhance recovery. In the remaining sections of this chapter, we review the literature and expand the discussion of two of
Table 7.1. Comparison of training principles between operant-conditioning and motor-learning-based interventions.
Practice variable Operant-conditioning training principles
Motor-learning training principles
Amount and scheduling of practice
Lifting of learned suppression explains the increased use of the affected limb in real-world activities:
• Learned-non-use develops from negative reinforcement during the acute stage where non-reinforced behavior becomes suppressed.
• Successful performance and positive reinforcement are necessary to lift the suppression allowing the behavior to be expressed in a real-world environment.
Massed practice is essential for cortical re-organization and reversal of learned-non-use:
• "Massed" practice in CIT is the term used to mean intense or extensive practice that is necessary to reverse learned-non-use and leads to cortical re-organization.
• The optimal duration, intensity, or challenge (level of difficulty) of practice for enhancing functional recovery has not been determined.
Task progression Shaping of motor behavior is essential especially for patients with limited ability:
• Shaping is based on the idea of successive approximations.
• Guidelines for progression are performance based and not learning based.
Skill acquisition, motor program, or schema formation and the development of internal representations for action explain the increased functional use of the affected limb for purposeful, volitional activities:
• Automatic and implicit procedural knowledge develops with practice of motor tasks.
• Tasks are controlled more automatically and with less cognitive effort; this manifests as skill develops.
Physical practice is the most important variable for motor learning:
• Practice that challenges the learner is motivating and optimal for learning-dependent cortical re-organization.
• The term "massed" practice is contrasted with "distributed" practice where within a bout of practice the distribution of practice-rest is manipulated. In "massed" practice, there is little to no rest and performance decrements due to fatigue are generally not considered detrimental to learning.
Task progression is learning based and depends on an analysis of underlying motor control deficits (strength, coordination, etc.):
• Progression can be accomplished by manipulating a variety of variables depending on individual needs (e.g., speed, ROM, adding or freeing degrees of freedom, part-whole task practice).
• These progression techniques are recommended especially for the lower or beginning levels of skill acquisition.
• Task complexity and parameterization within a class of actions are important components of task progression.
Operant-conditioning training principles
Motor-learning training principles
Role of errors
Diversity of tasks leads to a more generalized benefit of practice:
• During shaping, no more than two sets of 10 trials of a given task should be practiced in a single day.
Constant and frequent feedback is necessary for optimal lifting of learned suppression:
• The informational content of feedback is de-emphasized while reinforcement and encouragement are emphasized.
Errors during performance are ignored:
• Error feedback serves as negative reinforcement and is detrimental to the reversal of learned suppression of behavior.
• Errors are viewed as punishment that leads to avoidance of the behavior.
Behaviors are "elicited" in this model. These behaviors can be shaped through successive approximations during practice with positive reinforcement: • Collaboration with the patient for task selection or engagement for self-management and development of self-efficacy are not directly addressed in this model.
Contextual variety enhances problem-solving and retention of skills:
• The task practice schedule is designed to challenge the cognitive operations important for future capability in variable contexts.
Reduced augmented feedback (KR, KP) frequency and faded schedules that promote problem-solving and the development of internal error-detection capabilities are more beneficial for learning than constant and frequent feedback:
• The informational content of feedback is emphasized and is distinct from the encouragement provided during practice.
Errors during performance are beneficial to learning and therefore should be provided as information feedback to the performer:
• Errors are viewed as information that can be useful for planning the next trial.
• Movement problems are effectively solved partially through the provision of error information (feedback).
The development of skill through practice is by nature embedded into a meaningful and social context:
• The choice of tasks to practice and the development of self-efficacy as performance improves are intertwined with this approach and are recognized mediators for self-management and maintenance after training ends.
KP: knowledge of performance; ROM: range of motion.
these practice variables, augmented feedback and task scheduling, as they relate to motor skill acquisition in neurorehabilitation.
Was this article helpful?