Implicit learning of a perceptual-motor skill after stroke. (Research Report).

As a component of physical therapy interventions, motor learning is used to enhance skills. This process requires refinement of perceptual skills as well as motor skills; thus, the term "perceptual-motor skills" may be a most appropriate descriptor for this process. Cognitive processing skills are required as well. (1,2) A critical part of the cognitive process for perceptual-motor learning is the ability of the learner to comprehend and utilize the instructions provided to the learner. (3,4) Perhaps the most common use of instructions is to inform the learner about the goal of the task and what needs to be learned to achieve that goal. With these instructions, the learner can engage in explicit learning. In contrast, in implicit learning, the goal is not presented to the learner. The learner performs the task without awareness of what is being learned or, in some cases, without awareness that anything is being learned. (5)

"Implicit learning" is a broad term used to describe the acquisition of abstract knowledge without awareness of learning. (5) For example, abstract grammatical rules may be learned implicitly during the acquisition of language. (6) Procedural learning may be regarded as a subset of implicit learning in that it involves the learning of a perceptual-motor skill. (5)

One of the most common, paradigms used to study implicit learning is serial response time tasks. (7) Serial response time tasks have both perceptual and motor learning components and require the subject to respond to a stimulus, such as a light, with some motor response, such as touching it. Typically in research studies, the subject must attend to an array of 4 or more stimuli. The subject is instructed only to respond as quickly as possible to whichever stimulus lights up over a number of trials. What is not explained to the subject is that practice is organized in 2 ways: in 1 condition, the series of stimulus lights is presented in random order; in the other condition, a pattern for the presentation of stimulus lights is prescribed by the experimenter, and this pattern is repeated. Subjects are said to demonstrate learning of the pattern if their response time decreases during trials with the prescribed sequence and increases during subsequent trials with the random sequence. (8) Subjects are questioned after practice is complete, and if they report not having noticed any sequence or pattern, their learning is said to be implicit. Other subjects may gain explicit knowledge of the task and report that they noticed something about the task (eg, they may have an impression that some lights follow other lights). Some subjects may have recognition or recall regarding the actual pattern of the sequence and can recognize the pattern when it is displayed or can reproduce it without cuing. Performance by these individuals reflects explicit learning.

Work by Boyd and Winstein (9-12) was designed to examine the ability of people with stroke to learn by implicit and explicit means. In general, however, research that has examined motor learning in adults with stroke has focused on the instructions given to engage learners in explicit learning (13-16); that is, subjects were told what they were to try to learn. Overall, the results of these studies suggest that individuals with unilateral stroke can learn motor skills when instructions are provided to give them the opportunity to function at an explicit level. It is not known whether adults with stroke can learn a motor skill implicitly.

There is now substantial evidence (17-35) that implicit learning and explicit learning are controlled by different neural substrates. The functional neural network for implicit learning is thought to include the basal ganglia, (17) cerebellum, and prefrontal cortex. (18) The network for explicit learning is thought to include the temporal cortex, hippocampus, thalamus, (19,20) and frontoparietal cortex. (21) Learning differences in people with brain damage reflect, to a large extent, differences in neural substrates for implicit learning and explicit learning. For example, implicit learning, but not explicit learning, is preserved in some people with Alzheimer disease, (22-25) amnesia, (26) Huntington disease, (27,28) and multiple sclerosis. (29,30) In contrast, implicit learning is impaired in people with unilateral prefrontal cortex lesions, (31) Parkinson disease, (28,32,33) and cerebellar disease. (32,34,35)

Recently, it was shown that implicit learning is also dependent on the relationship between the side of the brain damage and the side of the effector, that is, the side of the limb performing the task. Implicit learning is preserved when adults with unilateral cerebellar lesions practice a task with the hand contralateral to the lesion, but this learning is impaired when practice is done with the hand ipsilateral to the lesion. (36) In contrast, people with unilateral damage to the supplementary motor area (37) or frontal area (38) have impaired implicit learning involving the contralateral upper extremity (UE). Finally, people with callosal dysfunction can learn implicit tasks involving one hand, but cannot do so when both hands are required to respond to stimuli. (38) Thus, it appears that the ability to learn implicitly depends on the integrity of the neuroanatomical substrates that subserve the performing UE.

Between 61% and 81% of strokes involve the anterior and middle cerebral arteries. (39) The primary clinical manifestation of these strokes is hemiparesis. Even after 6 months or more, 45% of those who survive stroke are left with some degree of hemiparesis. (40) Given that the primary deficits after stroke are on the side contralateral to the pathology, a unilateral stroke would leave the brain areas important for implicit learning intact if individuals performed a task with the UE ipsilateral to the lesion. Thus, it is reasonable to suggest that implicit learning would be preserved for people with unilateral stroke in the anterior circulation when performing with the UE ipsilateral to the lesion.

The primary purpose of our study was to examine the ability of adults with unilateral stroke to learn implicitly a motor skill using the UE ipsilateral to the side of brain damage compared with that of adults without stroke. If people with stroke can implicitly learn a serial response time task, performance would become faster during practice with a patterned, repeated sequence and would become slower with the introduction of a random sequence, and response time changes would be equivalent to those of people without stroke. If implicit learning is robust over time irrespective of unilateral brain damage, the speed of performance on a retention test 24 hours later should be similar between people with stroke and people without stroke.

A secondary purpose of our study was to investigate possible differences in the ability to learn implicitly relative to the severity of stroke. Given that the subjects performed the task with the UE ipsilateral to their brain damage, the severity of stroke should not affect implicit learning if implicit learning is dependent on an intact contralateral cortical substrate.

Method

Participants

We recruited 47 adults who had stroke and a comparison group of 36 adults without stroke matched for age and education (controls) (Table). To be included in the study, participants had to be: (1) at least 60 years of age, (2) living in the community, and (3) right-hand dominant (before stroke for those in the stroke group), as assessed by a handedness questionnaire. (41) Only subjects who were right-hand dominant were included to limit confounds in measures of performance of those with stroke related to brain laterality. In addition, those with stroke had to be at least 6 months postonset. Subjects were screened by trained research assistants to ensure that the subjects could follow 3-step commands and maintain a sitting posture independently. Exclusion criteria established by trained research assistants included having: (1) a history of any neurological disorder except for those in the stroke group, (2) impaired cognitive status (that is, a Mini-Mental Status Examination score of less than 18/30), (42) (3) pathology of the UE that would be used to perform the task if that limitation would affect the ability to perform the movement, and (4) uncorrected visual impairment. In addition, for those with stroke, exclusion criteria included apraxia, as defined by the Florida Apraxia Screen, (43) and stroke involving the brain stem or the posterior circulation system.

Adults with stroke were recruited from a pool of subjects who had participated in a previous study (n=27) and from local stroke support groups (n=20). The data set from the previous study included documentation made in medical charts within 14 days poststroke concerning the location of the lesion and the severity of the stroke, as defined by the Orpington Prognostic Scale. (44) As defined by the Orpington Prognostic Scale, (44) 18 of the 27 individuals with stroke from the previous study had mild stroke, and 9 had moderate stroke. Information about the severity of stroke was not available for the 20 participants from the stroke support groups. The presence of a unilateral lesion in the anterior circulation for those in the stroke support groups was determined from the participants' reports. Older adults with no known pathology or impairments were recruited from the Gray-hawk Panel (principal investigator: JMM), a pool of over 600 people who have agreed to participate in research.

Task and Equipment

The equipment used was a custom-made hand movement device (Fig. 1) consisting of 16 targets of 4 types of switches. Each type of switch was easily closed but required a different hand position (modeled after that of Harrington and Haaland (45)). We did not, however, specify a hand position for each type of switch. The first row consisted of 2.54-cm-diameter knobs that were closed by turning them to the right or left. The second row consisted of switches (the size of standard light switches) that were closed by moving them to the right or left. The third row consisted of 2.54-cm-diameter buttons that were closed by pushing them in. The fourth row consisted of switches (the size of standard light switches) that were closed by moving them up or down. A light was situated above each switch.

[FIGURE 1 OMITTED]

The hand movement device was linked to a computer that controlled the sequencing of the lights and recorded the time from the illumination of the first target light to the closure of the eighth switch (ie, an 8-target sequence). The lights for both the patterned sequence and the random sequence, proceeded from left to right and back again for the 8 targets. The first target light was in the first column, the second target light was in the second column, and so on, with the fourth and fifth target lights both appearing in the fourth column. The patterned sequence was randomly selected within the constraints of the order of columns. The target light remained illuminated until the correct response was made. Closure of the target switch (that is, the correct switch) caused the next light in the sequence to become illuminated. Thus, the subject was required to move from one switch to the next switch without resting.

Procedure

The rights of human subjects were protected. Participants signed a university-approved informed consent form that was part of a larger study evaluating attentional and learning capabilities of adults poststroke. The data from the motor learning task presented here are a subset of data collected. Participants with stroke used the UE ipsilateral to the side of brain damage; the UE that controls used was matched so that half used their left UE and half used their right UE. Participants practiced in 2 conditions: an 8-target patterned sequence and an 8-target random sequence. The participants were not informed that there were 2 conditions. The instructions were to close the switch corresponding to the light that became lit and to do so as quickly as possible. Practice was provided in an effort to ensure that the participant could close each switch properly.

Practice was conducted over 2 consecutive days and was organized into blocks. A block consisted of 10 repetitions of an 8-response sequence. A block was completed without interruption; that is, the closure of a target switch activated the subsequent target light to become illuminated. Subjects were given short rests of approximately 2 to 3 minutes as needed between blocks. On day 1, blocks 1 through 4 were the patterned sequence, blocks 5 and 6 were the random sequence and blocks 7 and 8 returned to the patterned sequence. The second day of practice was used as a measure of retention. Participants performed 2 blocks of the patterned sequence. Thus, participants performed a total of 100 8-response sequences; 80 of those were the patterned sequence.

A procedure commonly used to examine serial responses was used to assess explicit knowledge of the patterned sequence. In a recall test at the end of the retention measure on day 2, participants were asked whether they noticed anything about the task. Regardless of their response to this question, they were then informed that there was a patterned, repeated sequence embedded within their practice sessions. They were told to try to reconstruct that sequence by pointing to the switches (without the cuing of target lights) in the order that they thought represented the sequence. If they said they did not know, they were told to guess and to do the best they could.

Data Processing and Analysis

The sequence response time (SR) was recorded; this represented the time to complete a sequence of 8 stimuli and responses. Errors, such as moving the hand toward the incorrect switch or closing the wrong switch, were not counted individually but caused delays for the whole sequence and thus are reflected in the SR. The mean and standard error of the mean (SEM) of the SR were calculated for each 10 repetitions of an 8-response sequence to form blocks. The Statistical Package for the Social Sciences personal computer program (version 8.0) * was used for all analyses.

To examine differences between subjects with stroke and controls, 4 separate analyses of variance (ANOVAs) with a factor of group (stroke or control) and a repeated-measures factor of block were used to assess the following 4 items: learning of the patterned sequence, changes in performance with introduction of the random sequence, changes in performance with reintroduction of the patterned sequence, and retention. Effect size was calculated using an eta-square analysis ([[eta].sup.2]). (46) Values of [[eta].sup.2] of .01, .06, and .14 represent small, medium, and large effect sizes, respectively. (46) A main effect of group would indicate a difference in SR between those with stroke and controls. If implicit learning is preserved in individuals with stroke, then there would be a main effect of block such that SR would become shorter with practice of the patterned sequence, SR would become longer when the random sequence is introduced, and SR would become shorter again when the patterned sequence is reintroduced. If implicit learning is retained in both groups, then there would be no main effect of group. For each analysis, group X block interactions would indicate differences in implicit learning between those with stroke and controls.

Explicit knowledge of the patterned sequence was assessed using a chi-square analysis. If individuals with stroke gain explicit knowledge as well as controls, then there would be no association between group and yes-no responses. The longest series of correct responses that subjects were able to generate was analyzed with a one-way ANOVA with group (stroke or control) as the factor. If individuals with stroke gain explicit knowledge as well as controls, then there would be no group effect.

To investigate whether adults with moderate stroke could learn implicitly, the data from the 27 subjects whose stroke severity had been defined as mild or moderate (ie, a subset of the large stroke group analysis) were analyzed in a 2-way ANOVA with a factor of group (mild or moderate) and a repeated-measures factor of block, with the same block comparisons as those aforementioned. In all of these analyses, the level of significance was P<.05, except for post hoc analyses, in which paired t tests were completed with a Bonferroni correction. The Greenhouse-Geisser method, a statistical method that makes the necessary adjustments for repeated measures, (46) was used for all repeated-measures tests.

Results

Learning of the Patterned Sequence: Blocks 2 to 4

Means and SEMs for SR by group for the 10 blocks of practice are shown in Figure 2. Because it is likely that changes from block 1 to block 2 included improvements attributable to nonspecific learning of the task, we analyzed the block effects of blocks 2 to 4. Subjects with stroke decreased their SR over blocks 2 to 4 by 71 [+ or -] 21 milliseconds, representing a 5% improvement. Controls decreased their SR by 77 [+ or -] 12 milliseconds, representing a 7% improvement over these 3 blocks. There was a main effect of block (F= 15.795, P<.001, [[eta].sup.2]=.163), reflecting decreased SR with practice. There was also a main effect of group (F=13.146, P<.002, [[eta].sup.2]=.140), reflecting the longer SR of those with stroke. The absence of a group X block interaction suggests that both groups improved performance at similar rates.

[FIGURE 2 OMITTED]

Means and SEMs for SR by group for the 10 blocks of practice for those with mild and moderate strokes are shown in Figure 3. From block 2 to block 4, subjects with mild stroke decreased their SR by 77 [+ or -] 27 milliseconds, and those with moderate stroke increased their SR by 40 [+ or -] 45 milliseconds. There was a main effect of group (F=8.971, P=.006, [[eta].sup.2]=.264), and there was a group X block interaction (F=3.358, P=.049, [[eta].sup.2]=.118). Post hoc analyses of the group X block interaction did not reveal differences between any pair of blocks for either group.

[FIGURE 3 OMITTED]

Interruption of Performance With the Random Sequence: Blocks 4 and 5

With the introduction of the random sequence, the SR of those with stroke increased 156 [+ or -] 29 milliseconds, a 12% increase over their performance at block 4. Similarly, the SR of controls increased 145 [+ or -] 35 milliseconds, a 14% increase over their performance at block 4. There was a main effect of block (F=43.910, P=.000, [[eta].sup.2]=.352). Individuals with stroke were slower than controls, as evidenced by the main effect of group (F= 14.704, P=.000, [[eta].sup.2]=.154). There was no group X block interaction.

The SR of individuals with mild stroke increased 250 [+ or -] 49 milliseconds, and that of individuals with moderate stroke increased 95 [+ or -] 85 milliseconds. There was a main effect of group (F=5.662, P=.025, [[eta].sup.2]=.185), and there was a main effect of block (F= 14.247, P<.001, [[eta].sup.2]=.363).

Enhancement of Performance With Reintroduction of the Patterned Sequence: Blocks 6 and 7

The reintroduction of the patterned sequence at block 7 resulted in faster SRs for those with stroke (55 [+ or -] 21 milliseconds) and controls (29 [+ or -] 12 milliseconds), representing improvements from block 6 of 4% and 3%, respectively. There was a main effect of block (F=9.801, P=.002, [[eta].sup.2]=.108), and there was a main effect of group (F=13.931, P=.000, [[eta].sup.2]=.147). There was no group X block interaction.

Sequence response times were 138 [+ or -] 36 milliseconds faster at block 7 for those with mild stroke and 15 [+ or -] 38 milliseconds faster for those with moderate stroke. There was a main effect of group (F=5.797, P<.025, [[eta].sup.2]=.188), there was a main effect of block (F=7.146, P<.014, [[eta].sup.2]=.222), and there was a group X block interaction (F=4.610, P=.042, [[eta].sup.2]=.156). Post hoc analyses revealed that the interaction was attributable to the faster SR in block 7 than in block 6 for those in the mild-stroke group only (t=3.885, P=.001).

Retention of the Patterned Sequence: Blocks 9 and 10

Controls and those with stroke decreased their SR over the 2 blocks of retention. There was a main effect of block (F=121.949, P=.000, [[eta].sup.2]=.601), and there was a main effect of group (F= 11.781, P=.001, [[eta].sup.2]=.127).

The analyses of blocks 9 and 10 for those with mild and moderate strokes revealed a group X block interaction (F=6.522, P<.02, [[eta].sup.2]=.207). Post hoc analyses revealed that the SR for block 10 was faster than that for block 9 for both those with mild stroke (t=8.643, P<.001) and those with moderate stroke (t=8.034, P<.001).

Explicit Knowledge of the Patterned Sequence

With regard to whether participants noticed anything about the task, there was no difference between the participants with stroke and the control group, with most individuals stating that they noticed something (68% of those with stroke and 75% of controls). Similarly, there was no difference between those with stroke and the controls (2.2 [+ or -] 0.3 and 2.8 [+ or -] 0.4, respectively) for the largest number of correct single responses in a row recalled.

There was no difference in explicit knowledge of the patterned sequence between those with mild stroke and those with moderate stroke; two thirds of those with mild stroke noticed something about the task, and just over one half of those with moderate stroke did so. Subjects with moderate stroke did recall a larger number of correct responses in a row (3.7 [+ or -] 0.7) than did those with mild stroke (1.7 [+ or -] 0.4), as revealed by a one-way ANOVA (F=6.818, P<.02).

Discussion

To our knowledge, this is the first study to show that implicit learning of a perceptual-motor task may be preserved in people with unilateral stroke, particularly if the stroke is mild. As we hypothesized, individuals who had stroke and used the UE ipsilateral to the side of brain damage demonstrated the ability to learn implicitly an SR task as well as older adults without impairments, although the performance of those with stroke was consistently slower than the performance of the comparison group. The SRs of both groups improved with practice of a patterned sequence, were lengthened by the onset of practice of a random sequence, and improved again with reintroduction of the patterned sequence. Adults with stroke continued to be slower than controls in the retention phase of the study. Both groups showed faster SRs in the second block than in the first block on day 2.

Although rehabilitation after stroke often has been based on the learning of motor skills, only within approximately the last 8 years has research provided some guidance regarding strategies for constructing practice for people with brain damage. (47) Recently, the impact of instructions on rehabilitation has been discussed. (48) Instructional strategies can provide the learner with explicit knowledge about the features of a task that are essential for performance of the task (ie, what a person needs to learn to achieve the goal associated with the task). (3) Explicit learning is invoked when participants are searching for features, rules, or some structural property during practice and when they are able to say how they solved the problem. (49) An alternative instructional strategy does not provide the learner with knowledge about the critical properties of the task, and the skill is learned implicitly. (3) Implicit learning occurs when participants focus on specific items, such as searching for a target light, not rules, such as any relationship between successive target lights. (49)

In adults without neurological damage, there is evidence that explicit instructions can lead to poorer performance than implicit instructions. (50,51) What instructional set is appropriate for people with stroke? Our results demonstrate that people with stroke, particularly those with mild stroke, can learn under implicit-learning conditions, suggesting that a strategy with implicit instructions may be important in rehabilitation. Studies are needed to compare explicit and implicit strategies to determine whether there is an optimal instructional set for adults with stroke.

The differences in performance between the participants with stroke and the older adults without impairments were attributable to the slower SR of those with stroke. This finding is congruent with the findings of researchers who have demonstrated that movements with the UE ipsilateral to the side of brain damage are slower than those in controls. (15,16,52) In general, we suggest, as others have before, (15,16,52) that the ipsilateral UE movement slowing may be attributable to disrupted output from the damaged hemisphere influencing the control of both UEs. An alternative, but not exclusive, explanation is that both hemispheres are necessary for normal execution of unimanual movements, particularly if the movements are complex. (53-55)

Implicit learning may be particularly appropriate for some people with stroke but not all. Contrary to our hypothesis, the performance of subjects with moderate stroke did not change as the practice condition changed even though the subjects performed with the UE ipsilateral to the lesion. It appears that the subjects were not able to learn the task implicitly and that the only learning that occurred was a nonspecific learning that allowed them to decrease SR and increase speed from the first block to the second block of retention. Given the small number of subjects with moderate stroke, we suggest that conclusions be drawn with caution. Damage in the neostriatum impairs implicit learning, (32,56) and it may do so for both UEs. We did not have information about precise lesion locations. People with moderate stroke may have had greater involvement of the neostriatum than those with mild stroke.

Implicit learning is not a passive process, (57) nor is it unconscious. (58) The learner needs to be able to attend to the task that is practiced. (7) Adults with stroke may have deficits in attentional processing. (59) In our study, numerous trials were required. We chose a task that required different hand postures rather than the single hand posture that was required in the original serial response time task studies. (7) We did this to make the task more interesting, although we admittedly have no measures of whether subjects were interested or attentive.

The presence of implicit learning suggests that there is no need to ask learners questions during practice about what they are experiencing, (3) in sharp contrast to some of the techniques used in physical therapy. Whether there is a dichotomy in implicit learning and explicit learning is questionable. (60) Indeed, some authors (57,61) have argued that, with all learning, some parts of learning are explicit and some are implicit. An optimal strategy for motor learning within the context of rehabilitation may require some blend of explicit and implicit strategies that has yet to be defined, but at present this notion is speculative.

Implicit learning usually is thought to occur when changes in performance occur as the practice conditions change, such as a change from the patterned sequence to the random sequence in our study. The inclusion of a retention test is compatible with the motor learning literature in which learning is said to exist only with relatively permanent changes in behavior. (4) The results of our retention test suggest that there was considerable retention of implicit learning for adults with stroke and controls. Group differences persisted in the retention test performance with the patterned sequence. However, both groups increased the speed of performance from the first block to the second block of retention. This result is congruent with research that has shown that implicit learning is robust (49,62) and resistant to deterioration over time. (62) However, we believe that caution must be exercised when generalizing the benefits of implicit learning to all measures of motor learning. Although retention tests have demonstrated the effectiveness of implicit learning, transfer tests have revealed that implicit learning can result in negative transfer. (62) After implicit learning, performance deteriorated when subjects were asked to perform a task that was similar to but not exactly the same as the learned task. (62) There was no transfer test in our study; however, no data suggested that transfer test performance would be any different for adults with stroke versus adults without stroke.

Explicit knowledge of a task learned under implicit-learning conditions does not appear to be different between individuals with stroke and older adults without impairments. A large percentage of subjects answered "yes" to the question about whether they noticed anything about the task. This open-ended question did not provide us with any information about what they noticed but was chosen to avoid leading the respondents as much as possible. However, the longest correct series recalled by members of either group was only approximately 2 out of the 8 responses. Thus, although subjects appeared, on the basis of questioning, to have some explicit knowledge, their recall performance was limited.

The relationship between the presence of explicit knowledge and the level of implicit learning in a task practiced under implicit-learning conditions has been debated in the literature. (60,63) Explicit learning and implicit learning may occur in parallel, as suggested by the finding that those who train under explicit-learning conditions also gain implicit knowledge. (60) Other researchers, however, have regarded the relationship between explicit learning and implicit learning as "antagonistic" (see, for example, Buchner et a1 (63)).

There are several limitations to our study. Using reports from medical charts, personal reports, and examinations, we limited our participants with stroke to those with unilateral lesions in what appeared to be the distribution of the anterior circulation system. We suggest that the results from our study should not be generalized to people with pathology in the brain stem or posterior circulation system or to those with bilateral brain damage. The diagnosis of stroke documented in medical charts for our subjects was based on a combination of clinical signs and symptoms and radiological information. Reports from brain imaging were not available for each subject. Without this information, it is not possible to verify the precise lesion size and location. We acknowledge that a branch of the middle cerebral artery serves subcortical areas, including the basal ganglia. (64) In the absence of brain imaging reports, we cannot exclude the possibility that some of our participants with stroke had lesions that included the basal ganglia.

Another limitation of our study is that the participants with stroke were at least 6 months after onset of the stroke. We do not know whether implicit learning is intact during the early stages after stroke, when rehabilitation efforts are directed. This question remains to be answered. Our task was atypical and differed from the reaction time task used in serial response time paradigms. The nature of the task, that is, requiring different hand postures for responses, may have invoked more explicit-learning strategies than a task that requires only pushing buttons for each response. Finally, the individuals with stroke performed the perceptual-motor task with the UE ipsilateral to the lesion. Recently, it was shown that the primary sensorimotor cortex contralateral to the performing UE is active during implicit learning. (21) In this light, we would predict that people with stroke could not learn implicitly with the contralateral, more affected UE. This notion remains to be empirically tested.

Conclusion

Our results suggest that adults with unilateral stroke can learn implicitly, particularly if the stroke is mild. The clinician should consider the instructional set provided to people with stroke during interventions. In rehabilitation, we may often rely on verbalization to describe the goals of skills that we are teaching our clients. From our experience, we often see explicit instructions provided to clients on how to meet these goals. For example, in coming to a standing position, clients may be told to keep their head up and to lean forward so that their nose is over their knees. Although this teaching strategy may be successful, our study suggests that adults with mild stroke may learn perceptual-motor tasks through implicit-learning experiences. This may be particularly helpful for individuals who have difficulty processing extensive verbal instructions. Incorporating strategies to engage an individual in implicit learning may provide a successful avenue to the learning of motor skills during rehabilitation.

Table.
Characteristics of Participants

          Age (y)

Group     [bar]X   SD   Male/Female   Education (y)

Stroke    71       6    29/18         12.6
Control   73       6    15/21         13.1

                     Time Since Stroke (y)

Group     MMSE (a)   [bar]X   SD    Range

Stroke    27.5       3.6      5.1   0.6-28.4
Control   28.6

(a) MMSE=Mini-Mental Status Examination score (maximum=30).

* SPSS Inc, 444 N Michigan Ave, Chicago, IL 60611.

References

(1) Fitts PM. Perceptual-motor skills learning. In: Melton AW, ed. Categories of Human Learning. New York, NY: Academic Press Inc; 1964:243-285.

(2) Newell KM. Motor skill acquisition. Annu Rev Psychol. 1991;42: 213-237.

(3) Magill RA. Knowledge is more than we can talk about: implicit learning in motor skill acquisition. Res Q Exerc Sport. 1998;69:104-110.

(4) Schmidt RA. Motor Control and Learning. Champaign, Ill: Human Kinetics Inc; 1988.

(5) Green REA, Shanks DR. On the existence of independent explicit and implicit learning systems: an examination of some evidence. Mem Cognit. 1993;21:304-317.

(6) Shanks DR, St John MF. Characteristics of dissociable human learning systems. Behav Brain Sci. 1994;17:367-447.

(7) Nissen MJ, Bullemer P. Attentional requirements of learning: evidence from performance measures. Cogn Psychol. 1987;19:1-32.

(8) Willingham DB. Implicit motor sequence learning is not purely perceptual. Mem Cognit. 1999;27:561-572.

(9) Boyd LA, Winstein CJ. Serial reaction time learning in individuals with unilateral brain damage. Soc Neurosci Abstr. 1998;24:2116.

(10) Boyd LA, Winstein CJ. Implicit learning is diminished in individuals with a cerebrovascular accident [abstract]. Neurol Rep. 1998;22:179.

(11) Boyd LA, Winstein CJ. Explicit knowledge promotes implicit learning of the serial reaction time task. Soc Neurosci Abstr. 1999;25:1900.

(12) Boyd LA, Winstein CJ. Explicit knowledge promotes implicit learning while extended practice does not in individuals with unilateral stroke [abstract]. Neurol Rep. 1999;23:191-192.

(13) Hanlon RE. Motor learning following unilateral stroke. Arch Phys Med Rehabil. 1996;77:811-815.

(14) Platz T, Denzler P, Kaden B, Mauritz K-H. Motor learning after recovery from hemiparesis. Neuropsychologia. 1994;32: 1209-1223.

(15) Pohl PS, Winstein CJ. Practice effects on the less-affected upper extremity after stroke. Arch Phys Med Rehabil. 1999;80:668-675.

(16) Winstein CJ, Merians AS, Sullivan KJ. Motor learning after unilateral brain damage. Neuropsychologia. 1999;37:975-987.

(17) Rauch SL, Whalen PJ, Savage CR, et al. Striatal recruitment during an implicit sequence learning task as measured by functional magnetic resonance imaging. Hum Brain Mapp. 1997;5:124-132.

(18) Pascual-Leone A, Wassermann EM, Grafman J, Hallett M. The role of the dorsolateral prefrontal cortex in implicit procedural learning. Exp Brain Res. 1996;107:479-485.

(19) Mishkin M, Appenzeller T. The anatomy of memory. Sci Am. 1987;256:80-89.

(20) Squire LR. Mechanisms of memory. Science. 1986;232:1612-1619.

(21) Honda M, Deiber M-P, Ibanez V, et al. Dynamic cortical involvement in implicit and explicit motor sequence learning: a PET study. Brain. 1998;121:2159-2173.

(22) Deweer B, Ergis AM, Fossati P, et al. Explicit memory, procedural learning and lexical priming in Alzheimer's disease. Cortex. 1994;30: 113-126.

(23) Grafman J, Weingartner H, Newhouse PA, et al. Implicit learning in patients with Alzheimer's disease. Pharmacopsychiatry. 1990,23:94-101.

(24) Libon DJ, Bogdanoff B, Cloud BS, et al. Declarative and procedural learning, quantitative measures of the hippocampus, and subcortical white alterations in Alzheimer's disease and ischaemic vascular dementia. J Clin Exp Neuropsychol. 1998;20:30-41.

(25) Sabe L, Jason L, Juejati M, et al. Dissociation between declarative and procedural learning in dementia and depression. J Clin Exp Neuropsychol. 1995; 17:841-848.

(26) Baddeley A, Wilson BA. When implicit learning fails: amnesia and the problem of error elimination. Neuropsychologia. 1994;32:53-68.

(27) Bylsma FW, Rebok GW, Brandt R. Long-term retention of implicit learning in Huntington's disease. Neuropsychologia. 1991;29:1213-1221.

(28) Sprengelmeyer R, Canavan AGM, Lange HW, Homberg V. Associative learning in degenerative neostriatal disorders: contrasts in explicit and implicit remembering between Parkinson's and Huntington's diseases. Mov Disord. 1995;10:51-65.

(29) Beatty WW, Goodkin DE, Monson N, Beatty PA. Implicit learning in patients with chronic progressive multiple sclerosis. Int J Clin Neuropsychol. 1990;12:166-172.

(30) Rao SM, Grafman J, DiGiulio D, et al. Memory dysfunction in multiple sclerosis: its relation to working memory, semantic encoding, and implicit learning. Neuropsychology. 1993;7:364-374.

(31) Beldarrain MG, Grafman J, Pascual-Leone A, Garcia-Monco JC. Procedural learning is impaired in patients with prefrontal lesions. Neurology. 1999;52:1853-1860.

(32) Doyon J, Gaudreau D, Laforce R Jr, et al. Role of the striatum, cerebellum, and frontal lobes in the learning of a visuomotor sequence. Brain Cogn. 1997;34:218-245.

(33) Westwater H, McDowall J, Siegert R, et al. Implicit learning in Parkinson's disease: evidence from a verbal version of the serial reaction time task. J Clin Exp Neuropsychol. 1998;20:413-418.

(34) Molinari M, Leggio MG, Solida A, et al. Cerebellum and procedural learning: evidence from focal cerebellar lesions. Brain. 1997;120: 1753-1762.

(35) Pascual-Leone A, Grafman J, Clark K, et al. Procedural learning in Parkinson's disease and cerebellar degeneration. Ann Neurol. 1993;34: 594-602.

(36) Gomez-Beldarrain M, Garcia-Monco JC, Rubio B, Pascual-Leone A. Effect of focal cerebellar lesions on procedural learning in the serial reaction time task. Exp Brain Res. 1998;120:25-30.

(37) Ackermann H, Daum I, Schugens MM, Grodd W. Impaired procedural learning after damage to the left supplementary motor area (SMA). J Neurol Neurosurg Psychiatry. 1996;60:94-97.

(38) de Guise E, del Pesce M, Foschi N, et al. Callosal and cortical contribution to procedural learning. Brain. 1999; 122:1049-1062.

(39) Gresham GE, Duncan PW, Stason WB, et al. Post-Stroke Rehabilitation. Clinical Practice Guideline No. 16. Rockville, Md: US Department of Health and Human Services; 1995

(40) Gresham GE, Phillips TF, Wolf PA, et al. Epidemiologic profile of long-term stroke disability: the Framingham Study. Arch Phys Med Rehabil. 1979;60:487-491.

(41) Bryden M. Laterality: Functional Asymmetry in the Intact Brain. New York, NY: Academic Press Inc; 1982.

(42) Folstein MF, Folstein SE, McHugh PR. "Mini-Mental State": a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975; 12:189-198.

(43) Rothi LJG, Raymer AM, Heilman KM. Limb praxis assessment. In: Rothi LJG, Heilman KM, eds. Apraxia: The Neuropsychology of Action. Brain Damage, Behaviour and Cognition Series. Hove, England: Psychology Press/Elrbaum (UK) Taylor & Francis; 1997:61-73.

(44) Kalra L, Crome P. The role of prognostic scores in targeting stroke rehabilitation in elderly patients. J Am Geriatr Soc. 1993;41:396-400.

(45) Harrington DL, Haaland KY. Programming sequences of hand postures. J Motor Behav. 1987;19:77-95.

(46) Green SB, Salkind NJ, Akey TM. Using SPSS for Windows. 2nd ed. Chicago, Ill: SPSS Inc; 2000:224.

(47) Winstein CJ. Knowledge from results and motor learning: implications for physical therapy. Phys Ther. 1991;71:140-149.

(48) McNevin NH, Wulf G, Carlson C. Effects of attentional focus, self-control, and dyad training on motor learning: implications for physical rehabilitation. Phys Ther. 2000;80:373-385.

(49) Dienes A, Berrry D. Implicit learning: below the subjective threshold. Psychon Bull Rev. 1997;4:3-23.

(50) Green TD, Flowers JH. Implicit versus explicit learning processes in a probabilistic, continuous fine-motor catching task. J Motor Behav. 1991 ;23:293-300.

(51) Magill RA, Clark R. Implicit versus explicit learning of pursuit-tracking patterns. J Exerc Sport Psychol. 1997;19:S85.

(52) Winstein CJ, Pohl PS. Effects of unilateral brain damage on the control of goal-directed hand movements. Exp Brain Res. 1995;105: 163-174.

(53) Colebatch JG, Dieber MP, Passingham RE, et al. Regional cerebral blood flow during voluntary arm and hand movements in human subjects. J Neurophysiol. 1991;65:1392-1401.

(54) Dieber MP, Passingham RE, Colebatch JG, et al. Cortical areas and the selection of movement: a study with positron emission tomography. Exp Brain Res. 1991;84:393-402.

(55) Winstein CJ, Grafton ST, Pohl PS. Motor task difficulty and brain activity: investigation of goal-directed reciprocal aiming using positron emission tomography. J Neurophysiol. 1997;77:1581-1594.

(56) Knowlton BJ, Mangles JA, Squire LR. A neostriatal habit learning system in humans. Science. 1996;273:1399-1402.

(57) Whittlesea BWA, Wright RL. Implicit (and explicit) learning: acting adaptively without knowing the consequences. J Exp Psychol Learn Mem Cogn. 1997;23:181-200.

(58) Whittlesea BWA, Dorken MD. Implicit learning: indirect, not unconscious. Psychon Bull Rev. 1997;4:63-67.

(59) Tatemichi TK, Desmond DW, Stern Y, et al. Cognitive impairment after stroke: frequency, patterns, and relationship to functional abilities. J Neurol Neurosurg Psychiatry. 1994;57:202-207.

(60) Willingham DB, Goedert-Eschmann K. The relation between implicit and explicit learning: evidence for parallel development. Psychol Sci. 1999;10:531-534.

(61) Reed J, Johnson P. Assessing implicit learning with indirect tests: determining what is learned about sequence structure. J Exp Psychol Learn Mem Cogn. 1994;20:585-594.

(62) Lee Y, Vakoch DA. Transfer and retention of implicit and explicit learning. Br J Psychol. 1996;87:637-651.

(63) Buchner A, Funke J, Berry DC. Negative correlations between control performance and verbalizable knowledge: indicators for implicit learning in process control tasks? Q J Exp Psychol A. 1995;48A: 166-187.

(64) Carpenter MB. Core Text of Neuroanatomy. 2nd ed. Baltimore, Md:Williams & Wilkins; 1978:327.

PS Pohl, PT, PhD, is Associate Professor, Department of Physical Therapy, University of Florida, Gainesville, FL 32610 (USA) ([e-mail]). Address all correspondence to Dr Pohl.

JM McDowd, PhD, is Associate Professor, Department of Occupational Therapy Education, University of Kansas Medical Center, Kansas City, Kan.

DL Filion, PhD, is Assistant Professor, Department of Psychology, University of Missouri-Kansas City, Kansas City, Mo.

LG Richards, OTR, PhD, is Assistant Instructor, Department of Occupational Therapy Education, and Associate Scientist, Center on Aging, University of Kansas Medical Center

W Stiers, PhD, is Clinical Associate Professor, Department of Rehabilitation Medicine, University of Kansas Medical Center.

All authors provided concept/research design. Dr Pohl provided writing. Dr McDowd provided project management. Dr Filion provided data analysis. Dr Stiers provided consultation (including review of manuscript before submission). The authors acknowledge Pamela W Duncan, PT, PhD, FAPTA, for providing access to the Kansas City Stroke Study and research assistants Valorie Wells and Pam Flax-Laws for assistance in subject recruitment and data collection.

This study was supported by the University of Kansas Claude D Pepper Older Americans Center for Excellence in Aging Award AG 14635 (Principal Investigator: Stephanie Studenski; IDS-2 Core Leader: Dr McDowd) and by NCMRR/NICHHD/NIH Grant 5 KO1 HDO1187-01 awarded to Dr Pohl. Subjects were recruited from the Kansas City Stroke Study supported by VA Grant E8793 (Principal Investigator: Pamela W Duncan).

This study was approved by the Human Subjects Committee of the University of Kansas Medical Center.

This article was submitted June 19, 2000, and was accepted April 6, 2001.