Task switching after stroke.

Executive function is a complex, high-level cognitive domain that describes the ability to plan and regulate a task or combinations of tasks. (1) Task switching is one aspect of executive function. (1) In everyday experience, people often are required to switch activity between 2 or more tasks. Task switching comes at a cost to performance, a cost measured as a switch cost. (2) Switch costs are determined by subtracting performance tinder practice conditions in which no switching is required from performance under practice conditions in which switching is required. The difference reflects the added cognitive burden of task switching. (3) Switching tasks, compared with performing the same tasks in a repetitive manner without switching, results in a deterioration of performance that is reflected in slower movements or more errors in responses. (2)

Task switching may be completely self-initiated, may occur at predictable time intervals, or may follow a pattern, such as alternating between 2 tasks. For example, walking with a one-handed walker requires a pattern of moving each leg and the walker in a prescribed sequence. In this case, switching relies primarily on internal cues and is considered to be under endogenous control. (2) In contrast, the requirement to switch may be cued externally and may occur at unpredictable intervals, such as when a telephone rings while one is reading a book. The external cue provides imperative information about which task to perform. In this case, task switching is considered to be under exogenous control. (2)

The typical paradigm for measuring switching costs focuses on exogenous control. Subjects must respond to an unpredictable external cue to switch from one activity to another. Manipulations of interest include the time interval from the cue to the appearance of the stimulus and the time interval from a response to the appearance of the next cue (see Monsell (2) for a review). Task switching under endogenous control is less well studied, but switching costs have been reported for a voluntary task switch, an activity in which subjects are given general guidelines about switching, but they decide when to switch. (4) It has been suggested that alternating activity between 2 tasks, a form of endogenous control, may be particularly challenging because of the requirement to keep track of the sequence. (2)

A network of cortical and subcortical regions forms the neural substrate for successful task switching. (5) Key structures that contribute to the cognitive flexibility to switch tasks include the parietal cortex (6-8) and the frontal (13,5,8-12) and prefrontal (13,14) cortices. Given the vast connectivity of these cortical areas with other brain structures, (15) it is not surprising that dysfunction in a wide range of brain regions can result in difficulties in task switching. There have been reports of higher switching costs in subjects with various diagnoses than in healthy control subjects; these diagnoses include cerebellar disease, (16) Parkinson disease (3,17) (see Gurd and Ward (18) for an exception), Huntington disease, (19) schizophrenia, (20) and attention-deficit/hyperactivity disorder. (21)

There is a paucity of literature on the ability of people who have survived a stroke to switch tasks. The adverse impact of focal lesions in the prefrontal cortex secondary to tumor resection or stroke on task-switching abilities has been reported. (9) The most common stroke, however, is an ischemic event that follows the distribution of the middle cerebral artery. (22) Given the extended branching of the arterial system, the resultant brain damage can affect numerous functional networks. With subject criteria expanded on the basis of the variable distribution and extent of lesions typical in stroke, we showed that people who are long-term survivors of stroke have deficits in task-switching abilities. (23) Task switching can be part of many functional tasks and clinical activities that are incorporated into physical therapy interventions. Given that the majority of rehabilitation efforts occur in the first months after stroke onset, we sought to determine whether task-switching deficits are present in the subacute phase after stroke.

The overall objective of this study was to examine the effects of unilateral subacute stroke on task switching. Specifically, we had 4 purposes. First, we needed to verify that our experimental manipulation of a switch condition and a no-switch condition was effective. Second, we were interested in whether the task-switching deficits reported in people who are long-term survivors of stroke would be present at 1 and 3 months after stroke. We hypothesized that task-switching deficits would be evident in this subacute phase.

Our third purpose was to compare the switching ability for a task under endogenous control with that for a task under exogenous control, that is, an alternating switch task versus a cued switch task. Given the additional demands of a task under endogenous control, we hypothesized that the performance of adults who had had a stroke would be particularly impaired in the alternating switch task.

Our fourth purpose was to examine performance at 1 and 3 months after stroke. The typical rapid rate of recovery in the first few months after stroke onset led us to hypothesize that switch costs would decrease by 3 months after stroke but that group differences would persist. Finally, although the study was not designed to investigate the issue of laterality, we compared performance as a function of the hand used to complete the tasks and the side of brain damage.

Method

Participants

We recruited 46 adults who had had a stroke within the previous month and 38 adults without a history of stroke (Table). Participants with stroke were recruited from the Stroke Registry, a database of more than 1,000 people who had survived a stroke, who were 50 years of age and older, and who lived within 96 km (60 miles) of the university. The registry excluded people who had serious cardiac or organ system disease, had a known limited life expectancy, or had a preexisting disability. Details of the criteria for the Stroke Registry are provided by Duncan et al. (24) Adults 50 years of age and older were recruited from a registry of older adults maintained by the second author to serve as a control group. These adults live in the community and have expressed a willingness to participate in research.

To be included in the study, participants had to have a Mini-Mental State Examination score of at least 24. (25) They had to be living in the community at the time of testing and be right-hand dominant (premorbid for people with stroke). (26) Subjects were included only if they had a minimum of 20/125 near-vision acuity, as tested with the Lighthouse Near Acuity Test, (27) and adequate color vision to see the display, as tested with the Ishihara Test for Colour Deficiency. (28) Participants had to have no history of brain damage, except for stroke for the participants in the stroke group. Those with stroke were excluded if they had apraxia, as measured by the Florida Apraxia Screen. (29) Stroke severity was categorized with the Orpington Prognostic Scale; subjects with severe stroke were excluded. (30) Subjects with stroke were tested at approximately 1 month after stroke and again 2 months later. Control subjects also were tested twice; baseline testing was followed by a second testing session 2 months later.

Switch Tasks

Two tasks reported in our previous study were used to measure switching costs. (23) For each trial, regardless of task, 1 of 4 differently shaped stimuli appeared on a computer screen (ie, circle, square, triangle, and cross). Between 1 and 4 of the same shapes appeared on the screen. Responses were made on a custom-designed device that interfaced with a laptop computer to record the accuracy of the response and the response time to the nearest millisecond. The device consisted of 4 microswitches that were assigned to the 4 shapes and numbers 1 to 4. In both switching tasks, the stimulus for each trial remained on the screen until a response was made. Completion of a response triggered the beginning of the next trial.

The alternating switch task was a measure of endogenous control. In the no-switch condition, participants had to respond to either the number of objects or the shape of the objects. Practice was organized into blocks that included a 12-stimulus sequence requiring responses to the number of objects and a 12-stimulus sequence requiring responses to the shape of the objects. In the alternating switch condition, participants had to respond to the shape of the objects and then to the number of objects on consecutive trials and so on in an alternating fashion. The task required that participants remember the just-executed response (to number or shape) so that they could respond to the other stimulus dimension (either number or shape) on the subsequent trial. A block included 24 trials. Participants performed a block of the no-switch condition and then a block of the alternating switch condition and so on until a total of 6 blocks were completed.

The cued switch task was a measure of exogenous control. Participants were cued by a written word (ie, number or shape) presented on the computer screen to indicate whether the subject should respond on the basis of the number of objects or the shape of the objects. The cue was presented simultaneously with the stimulus. After a variable number of trials (8-12), a participant would receive a cue to switch to the other response. Responses were made in the same manner as in the alternating switch task. Participants performed sufficient trials to respond to 30 cues to switch.

Procedure

All subjects provided oral and written consent. Screening tests were completed as described earlier. The 2 switch tasks were carried out as part of a larger study that included a battery of tests of attention and learning, some of which are reported elsewhere. (31) The order of tests was randomized. To avoid performance deficits attributable to hemiplegia of the hand contralateral to the side of brain damage, subjects in the stroke group completed the responses with the hand on the same side as the side of brain damage. Twenty-two of the control subjects performed the tasks with their right (dominant) hand, and 16 of the control subjects performed the tasks with their left (nondominant) hand. To verify that the subjects understood the procedure, they performed 2 or 3 practice trials of each of the 2 tasks before data collection was begun. Subjects were instructed to be fast and accurate. Total testing time was approximately 2 hours, including intermittent rests to minimize any potential effects of fatigue. Subjects were tested in their homes in a quiet room without distractions.

Data Analysis

For the alternating switch task, accuracy was determined as the percentage of correct responses for the no-switch trims and the switch trials. Median response times were calculated for each subject from these 2 sets of trials. Switch costs for both accuracy and response times were determined by subtracting the values for the no-switch condition from the values for the alternating switch condition.

For the cued switch condition, the no-switch condition was defined by the 3 responses before the cue to switch. The switch condition was defined by the 3 responses just after the cue to switch. Examination of trials just before and after a switch for a task that does not follow an alternating or predictable pattern is compatible with the literature. (4,23) Accuracy was again represented as the percentage of correct responses, and median response times were calculated. Switch costs were determined by subtracting the values for the no-switch condition from the values for the switch condition.

Individual subject data were used to calculate group means. We expected negative numbers for accuracy, indicating lower accuracy in the switch condition than in the no-switch condition. In contrast, we expected positive numbers for response times, indicating slower times for the switch condition than for the no-switch condition. Variance is presented as standard errors (SEs).

To verify that the manipulation of our no-switch and switch conditions was effective, we completed 2 multivariate analyses of variance (MANOVAs) with a main effect of condition (no switch, switch) for accuracy and response time--1 for the alternating switch task and 1 for the cued switch task. If a main effect of condition was found, then follow-up univariate analyses were planned to determine whether the condition effect was significant for both response time and accuracy.

We performed univariate analyses of variance with a between-subjects factor of group (stroke, control), a within-subject factor of task (alternating switch, cued switch), and a repeated-measures factor of time (time 1, time 2) for switch costs for accuracy and response time. To define the locus of interactions, post hoc tests with a Bonferroni correction were completed. Our hypothesis that subjects in the stroke group would have greater switching deficits compared with subjects in the control group would be supported by a main effect of group. If an endogenous switch task was more demanding than an exogenous switch task, then a main effect of task would reveal that the alternating switch task was more difficult than the cued switch task. The hypothesis that an endogenous task would be particularly difficult for subjects who had survived a stroke would be supported by an interaction of group and task. A significant interaction of group and time would support the hypothesis that subjects in the stroke group experienced some recovery in their task-switching ability and showed decreased switch costs from the 1-month test to the 3-month test.

Although not a purpose of this study, it would be informative to the understanding of the laterality of brain function as well as clinical practice to know whether the performance of subjects with right-side lesions was different from that of subjects with left-side lesions. Side was entered as a factor in a separate analysis to provide insight into this potential difference. A main effect of side would reveal differences in performance attributable to hand dominance. An interaction of side and group would reveal differences attributable to the side of the lesion.

All analyses were done with the personal computer version of SPSS, * version 13.0. The Greenhouse-Geisser test was used for the repeated-measures factor of time.

Results

Effect of Condition

Performance on switch trials was different than performance on no-switch trials. In the alternating switch task, the accuracy rates were 95.8% (SE=0.9) for the no-switch condition and 86.1% (SE=1.5) for the alternating switch condition. Response times were shorter for the no-switch condition, with a mean of 1,175 (SE= 31) milliseconds, than for the alternating switch condition, with a mean of 2,097 (SE=80) milliseconds. The MANOVA revealed a main effect of condition for the alternating switch task (P<.001). The univariate analysis that followed revealed that the condition effect was significant for both accuracy (P< .001) and response time (P<.001).

In the cued switch task, the manipulation was effective, but only for response time. In the reverse of our expectation, accuracy rates were better for the cued switch condition, at 93.1% (SE=1.1), than for the no-switch condition, at 90.9% (SE= 1.2). Response times were shorter for the no-switch condition, with a mean of 1,190 (SE=34) milliseconds, than for the cued switch condition, with a mean of 1,591 (SE= 56) milliseconds. The MANOVA revealed that the difference was significant (P<.001). The univariate analysis revealed that there were differences between conditions for accuracy (P<.001) and response time (P<.001).

In the following sections, we describe the main effects of group, task, and time and any interactions on switch costs.

Accuracy Switch Costs

Means and SEs by group, task, and time are shown in Figure 1. There was a main effect of group (P< .002). Subjects with stroke had higher switch costs than control subjects, with accuracy costs of -6.1% (SE= 0.9) for the stroke group and -1.4% (SE= 1.0) for the control group. The main effect of task was significant (P<.001). Accuracy switch costs were higher for the alternating switch task, at -9.7% (SE= 1.4), than for the cued switch task, at 2.3% (SE=0.5). There was also an interaction of group and task (P <.001). Post hoc tests revealed that the stroke group had higher accuracy switch costs than the control group in the alternating switch task (P<.001) but not in the cued switch task. Accuracy did not change over the 2 test times; the main effect of time was not significant, and there were no significant interactions with time.

[FIGURE 1 OMITTED]

Response Time Switch Costs Means and SEs by group, task, and time are shown in Figure 2. The main effect of group was not significant; the switch cost of subjects with stroke, 673 (SE=54) milliseconds, was not different from that of control subjects, 639 (SE=60) milliseconds. The cost for the alternating switch task was higher than the cost for the cued switch task, at 921 (SE=63) and 390 (SE=27) milliseconds, respectively. There was a main effect of task (P<.001). Subjects in both groups showed decreased switch costs from time 1, at 695 (SE=46) milliseconds, to time 2, at 616 (SE=40) milliseconds, resulting in a main effect of time (P<.019). There were no significant interactions.

[FIGURE 2 OMITTED]

Right Versus Left Performance

Switch costs were higher for performance with the left (nondominant) hand for both control subjects and subjects with stroke. There were main effects of side for both accuracy (P<.02) and response time (P<.04). There was a higher switch cost in accuracy for performance with the left hand than with the right hand, at -5.5% (SE=0.9) and -2.1% (SE= 1.0), respectively. Similarly, there was a higher switch cost in response time for performance with the left hand than with the right hand, at 745.8 (SE=58.8) and 570.5 (SE=53.2) milliseconds, respectively. The lack of significance for group x side interactions suggests that switching costs were related to the hand used for the task, not the side of the lesion, at least in this small sample.

Discussion

The present study was designed to examine task-switching abilities in the subacute phase after stroke. Our hypothesis that subjects with stroke would have higher switch costs than control subjects was partially supported. Although switch costs for response times were not differentially affected by stroke, we found higher switch costs in accuracy for subjects with stroke at both 1 and 3 months after stroke. Our findings are compatible with previous, reports of task-switching costs for adults with brain damage from stroke (23) and other etiologies. (3,16,17,19-21)

Difficulties in switching between tasks for subjects after stroke, a deficit similar to what is known as perseveration in the clinical literature, (32) have been reported, but these paradigms have focused on the motor component of the tasks. Responses included changing hand postures in a predetermined pattern or in response to target lights. (33,34) Difficulties in task switching in studies such as those have been described as motor control deficits. In the present study, switching deficits were apparent even when the necessary motor responses were minimal and equivalent across switch and no-switch conditions and when subjects with stroke used the hand that was not primarily affected by the stroke. Our results suggest that the control deficits in switching after stroke may be best described not as motor deficits but as deficits in executive function.

A higher switch cost for subjects with stroke than for control subjects was found only for accuracy. Switch costs for response time were similar between groups. The lack of a group difference in cost for response time despite the difference in cost for accuracy may be related to our calculation of the median response time; only correct response times were included in deriving the median. There was a tendency for subjects with stroke to have overall longer response times than control subjects. Slower movements, even with the hand ipsilateral to the side of brain damage, can be expected. (35) The more salient issue for the present study is that switch costs were similar between groups for response time but were higher for subjects with stroke for accuracy. When faced with the dual goals of being accurate and fast, subjects may sacrifice one to accomplish the other. (36)

Our results suggest that subjects with stroke, faced with the demands inherent in the switch trials, were not able to maintain the same combination of speed and accuracy as control subjects. It is also possible that subjects with stroke did not slow their responses because they could not inhibit executing a response. The failure to automatically inhibit an incorrect response may be the primary dysfunction in task switching. (37) Of note, impulsiveness is one of the clinical manifestations of stroke, particularly for those with right hemisphere lesions. (38)

As we hypothesized, the alternating task that relied on endogenous control was more difficult than the cued task that relied on exogenous control, as revealed by longer response times and lower accuracy. Alternating between 2 tasks provides a predictable pattern for switching, but the internal cuing necessary to alternate between 2 tasks is thought to place a greater load on working memory than an externally cued switch. (2) The subject must remember the response just completed and remember the required pattern. In a cued switch task, a response is not dependent on the previous response. Each trial is independent of previous trials, and performers can rely on a cue (in this case, visual) to inform them of the appropriate response. Our results suggest that switching under endogenous control was more difficult than switching under exogenous control for both of our subject groups, particularly so for older adults with stroke.

It should be noted that the 2 switching tasks used in the present study cannot be unequivocally dichotomized as being under endogenous or exogenous control. A response made in the cued switch task must be self-initiated, even if the selection of the type of response is not necessary. Similarly, in the alternating switch task, there is some element of exogenous control, if only in the confines of the possible stimuli and responses. They are not selected from some unrestricted array of choices. It is possible that there is a continuum of control and switching abilities of older adults and that the abilities of adults with stroke vary along that continuum.

An unexpected finding was the higher accuracy rate for the switch condition than for the no-switch condition in the cued switch task. The mean percentages were high for both conditions, that is, over 90%, and differed by only 2%. It is possible that the cue to switch from number to shape or from shape to number alerted the participant to attend to the response and influenced accuracy. We have no data to support this conjecture, however, and the finding may be only spurious.

Our hypothesis that only subjects with stroke would be able to decrease switch costs over time was not supported. All subjects reduced switch costs for response time from time 1 to time 2 without a change in switch costs for accuracy. Over the 2-month period between testing times, subjects with stroke received no specific intervention targeted at the ability to switch tasks. The decrease in switch costs for response time for both groups suggests that control subjects and subjects with stroke benefited from the experience with the tasks during the first testing session.

Without brain images, we are unable to address questions regarding the relationship of lesion location or lesion extent to task-switching abilities. Switch costs were not related to the side of the lesion for our group of subjects, but further study is warranted. Higher switch costs for people with left-side brain damage have been reported. (39) There also is evidence that switch costs are high regardless of lesion side but that the reasons for the difficulty in task switching are different for people with right-side lesions than for those with left-side lesions. (9) We do not know whether our results can be generalized to other switching tasks, but it is interesting to note that tasks that require switches that are made at the discretion of the performer still reveal switch costs. (4) Voluntary switching relies strongly on endogenous control and may closely mimic the manipulation of activities in daily life. The inability of subjects with stroke to switch between 2 or more tasks may have far-reaching effects contributing to disability.

Many motor activities that physical therapists include in their interventions with their clients after stroke require the executive function of task switching. Performance may rely on endogenous control. For example, sequencing a one-handed walker during gait follows a pattern of switching the movement of each lower extremity and the movement of the upper extremity with the walker; transferring from a wheelchair to a desk chair follows a sequence of switching from one task to another in preparing the wheelchair and executing the movement. In each case, the client must complete one task and initiate the next task while keeping track of the appropriate sequence. Performance of some activities in an intervention may depend on exogenous control. Verbal cues may guide a series of exercises. Saying "walker, right foot, left foot" during gait training provides external cues that incorporate aspects of exogenous control into an activity that would normally rely on endogenous control.

The results of the present study suggest that people with stroke may be particularly challenged by activities that involve switching, especially when the activities rely on the internal cues necessary for endogenous control. The clinician should consider that performance deficits may not be purely motor in nature; the deficits may have their foundation in cognitive impairment. Analyzing a client's task-switching abilities across various tasks and under both endogenous and exogenous conditions may help the clinician understand the nature of the deficits and provide insight into potential interventions.

Conclusion

Subjects in the subacute phase after stroke showed deficits in task switching, an ability that is one aspect of the cognitive domain of executive function. Deficits are particularly evident in switching tasks that are under endogenous control and that rely on internal cues to enact the switch. Task switching is part of many functional skills and intervention activities in physical therapy. Clinicians working with adults after stroke should consider the impact of task-switching deficits on client performance.

Dr Pohl, Dr McDowd, Dr Filion, Dr Richards, and Dr Stiers provided concept/idea/ research design. Dr Pohl, Dr Richards, and Dr Kluding provided writing. Dr Filion and Dr Stiers provided data collection, and Dr Pohl, Dr Filion, Dr Stiers, and Dr Kluding provided data analysis. Dr McDowd provided project management, fund procurement, subjects, and facilities/equipment. Dr Filion, Dr Stiers, and Dr Kluding provided consultation (including review of manuscript before submission).

The study protocol was approved by the Institutional Review Board of the University of Kansas Medical Center.

This study was supported by National Institute on Aging grant P60 AG14635 (Kansas Claude D. Pepper Older Americans Independence Center; Stephanie Studenski, Principal Investigator). This material is the result of work supported with resources and the use of facilities at the Malcom Randall VA Medical Center, North Florida/South Georgia Veterans Health System, Gainesville, Fla.

* SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

References

(1) Reimers S, Maylor EA. Task switching across the life span: effects of age on general and specific switch costs. Dev Psychol. 2005;41:661-671.

(2) Monsell S. Task switching. Trends in Cognitive Neuroscience. 2003;7:134-140.

(3) Rogers RD, Sahakian BJ, Hodges JR, et al. Dissociating executive mechanisms of task control following frontal lobe damage and Parkinson's disease. Brain. 1998;121: 815-842.

(4) Arrington CM, Logan GD. The cost of a voluntary task switch. Psychol Sci. 2004; 15:610-615.

(5) Derfuss J, Brass M, von Cramon DY. Cognitive control in the posterior frontolateral cortex: evidence from common activations in task coordination, interference control, and working memory. Neuroimage. 2004;23:604-612.

(6) Barber AD, Carter CS. Cognitive control involved in overcoming prepotent response tendencies and switching between tasks. Cereb Cortex. 2005;15:899-912.

(7) Kimberg DY, Aguirre GK, D'Esposito M. Modulation of task-related neural activity in task-switching: an fMRI study. Brain Res Cogn Brain Res. 2000;10:189-196.

(8) Wylie GR, Javitt DC, Foxe JJ. Don't think of a white bear: an fMRI investigation of the effects of sequential instructional sets on cortical activity in a task-switching paradigm. Hum Brain Mapp. 2004;21:279-297.

(9) Aron AR, Monsell S, Sahakian BJ, Robbins TW. A componential analysis of task-switching deficits associated with lesions of the left and right frontal cortex. Brain. 2004;127:1561-1573.

(10) Derfuss J, Brass M, Neumann J, yon Cramon DY. Involvement of the inferior frontal junction in cognitive control: meta-analyses of switching and Stroop studies. Hum Brain Mapp. 2005;25:22-34.

(11) Forstmann BU, Brass M, Koch I, yon Cramon DY. Internally generated and directly cued task sets: an investigation with fMRI. Neuropsychologia. 2005;43:943-952.

(12) Rushworth MFS, Hadland KA, Paus T, Sipila PK. Role of the human medial frontal cortex in task switching: a combined fMRI and TMS study. J Neurophysiol. 2002;87:2577-2592.

(13) Dove A, Pollmann S, Schubert T, et al. Prefrontal cortex activation in task switching: an event-related fMRI study. Cogn Brain Res. 2000;9:103-109.

(14) Dreher JC, Grafman J. Dissociating the roles of the rostral anterior cingulate and the lateral prefrontal cortices in performing two tasks simultaneously or successively. Cereb Cortex. 2003;13:329-339.

(15) Saper CB, Iversen S, Frackowiak R. Integration of sensory and motor function: the association areas of the cerebral cortex and the cognitive capabilities of the brain. In: Kandel ER, Schwartz JH, Jessell TM, eds. Principles of Neural Science. 4th ed. New York, NY: McGraw-Hill Book Co; 2000:363.

(16) Berger A, Sadeh M, Tzur G, et al. Task switching after cerebellar damage. Neuropsychology. 2005;19:362-370.

(17) Woodward TS, Bub DN, Hunter MA. Task switching deficits associated with Parkinson's disease reflect depleted attentional resources. Neuropsychologia. 2002;40: 1948-1955.

(18) Gurd JM, Ward CD. Retrieval from semantic and letter-initial categories in patients with Parkinson's disease. Neuropsychologia. 1989;27:743-746.

(19) Aron AR, Watkins L, Sahakian BJ, et al. Task-set switching deficits in early-stage Huntington's disease: implications for basal ganglia function. J Cogn Neurosci. 2003;15:629-642.

(20) Meiran N, Levine J, Merian N, Henik A. Task set switching in schizophrenia. Neuropsychologia. 2000;14:471-482.

(21) Tamm L, Menon V, Ringel J, Reiss AL. Event-related fMRI evidence of frontotemporal involvement in aberrant response inhibition and task switching in attention-deficit/hyperactivity disorder. J Am Acad Child Adolesc Psychiatry. 2004;43:1430-1440.

(22) Brust JCM. Appendix C: Circulation of the brainl In: Kandel ER, Schwartz JH, Jessell TM, eds. Principles of Neural Science. 4th Ed. New York, NY: McGraw-Hill Book Co; 2000:1306.

(23) McDowd JM, Filion DL, Pohl PS, et al. Attentional abilities and functional outcomes following stroke. J Gerontol B Psychol Sci Soc Sci. 2003;58:P45-P53.

(24) Duncan P, Studenski S, Richards L, et al. Randomized clinical trial of therapeutic exercise in subacute stroke. Stroke. 2003; 34:2173-2180.

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

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

(27) Elam JH. Analysis of methods for predicting near-magnification power. J Am Optom Assoc. 1997;68:31-36.

(28) Ishihara S. Ishihara's Test for Colour Defficiency. Tokyo, Japan: Kanehara & Co Ltd; 1997.

(29) 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, United Kingdom: Psychology Press/Erlbaum (UK) Taylor & Francis; 1997:61-73.

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

(31) Pohl PS, McDowd JM, Filion D, et al. Implicit learning of a motor skill after mild and moderate stroke. Clin Rehabil. 2006; 20:246-253.

(32) Rubio KB. Treatment of neurobehavioral deficits: a function-based approach. In: Gillen G, Burkhardt A, eds. Stroke Rehabilitation. A Function-Based Approach. St. Louis, Mo: Mosby-Year Book; 1998: 334-352.

(33) Harrington DL, Haaland KY. Motor sequencing with left hemisphere damage. Brain. 1992;115:857-874.

(34) Kimura D. Acquisition of a motor skill after left-hemisphere damage. Brain. 1977;100: 527-542.

(35) Pohl PS, Winstein CJ, Onla-or S. sensory-motor control in the ipsilesional upper extremity after stroke [corrigendum appears in NeuroRehabilitation. 1997;9: 245-249]. NeuroRehabilitation. 1997;9: 57-69.

(36) Fitts PM. The information capacity of the human motor system in controlling the amplitude of movement. J Exp Psychol. 1954;47:381-391.

(37) Verbruggen F, Llefooghe B, Szmalec A, Vandierendonck A. Inhibiting responses when switching: does it matter? Exp Psychol. 2005;52:125-130.

(38) Rapport LJ, Webster JS, Flemming KL, et al. Predictors of falls among right-hemisphere stroke patients in the rehabilitation setting. Arch Phys Med Rehabil. 1993;74: 621-626.

(39) Mecklinger A, von Cramon D, Springer A, Matthes-von Cramon G. Executive control functions in task switching: evidence from brain injured patients. J Clin Exp Neuropsychol. 1999;21:606-619.

PS Pohl, PT, PhD, is Associate Professor, Department of Physical Therapy & Rehabilitation Sciences, University of Kansas Medical Center, Mail Stop 2002, 3901 Rainbow Blvd, Kansas City, KS 66160 (USA). Address all correspondence to Dr Pohl at: [e-mail].

JM McDowd, PhD, is Professor, Department of Occupational Therapy Education, and Associate Director for Research, Landon Center on Aging, University of Kansas Medical Center.

D Filion, PhD, is Associate Professor, Department of Psychology, University of Missouri, Kansas City, Mo.

LG Richards, PhD, OTR, is Research Health Scientist, Research Service, North Florida/ South Georgia Veterans Health System, Gainesville, Fla, and Associate Professor, Occupational Therapy Department, University of Florida, Gainesville, Fla.

W Stiers, PhD, is Assistant Professor, Department of Physical Medicine and Rehabilitation, John Hopkins University, Baltimore, Md.

P Kluding, PT, PhD, is Assistant Professor, Department of Physical Therapy & Rehabilitation Sciences, University of Kansas Medical Center.

[Pohl PS, McDowd JM, Filion D, et al. Task switching after stroke. Phys Ther. 2007;87:66-76.]

[C] 2007 American Physical Therapy Association

This article was received March 23, 2006, and was accepted August 21, 2006.

Invited Commentary Lara A Boyd

How often have you treated an individual with stroke who, based on your experience and expertise, should have regained function but did not. What is going on? Routinely we generate a list of reasons to explain why sometimes our clients fail to recover function when seemingly they should thrive: poor motivation, concomitant medical problems, lack of family support, inadequate rehabilitation interventions, and so on. What if we had better tools with which to categorize some of these patients? Maybe we do.

In their article, Pohl et al describe a cognitive deficit associated with disrupted executive function after stroke. Switching between tasks is an essential component of normal function and an ability that most of us take absolutely for granted. In their research, Pohl et al used an elegant method demonstrating that task switching is not a motor function but rather a cognitive function. More importantly for physical therapists, the ability to switch between 2 tasks was impaired by the most common distribution of stroke: the middle cerebral artery. This switching deficit, or "switch cost," was particularly acute when the switch was not externally cued (ie, it was under "endogenous" control); yet, endogenous control of switching between tasks is essential for independent function. What makes this work even more relevant to physical therapists is the finding that a switching deficit was noted in individuals in the subacute phase (1-3 months) after stroke; a period that corresponds to the typical time frame for rehabilitation.

Pohl and colleagues' work raises a larger issue confronting physical therapists across practice domains: impairments in executive function may severely limit physical rehabilitation outcomes. Generally, physical therapists are ill prepared to deal with cognitive problems. Our lack of training to recognize and understand how cognitive impairments affect therapeutic outcomes is further compounded by a paucity of rehabilitation-related research considering these complex problems. Thus, the data presented by Pohl et al illustrate an important point: in some cases, successful rehabilitation requires consideration of the interaction between the motor and cognitive systems. Indeed, neural connectivity confirms the integrated nature of motor and cognitive systems for both motor learning and the production of voluntary movement.

Recent functional magnetic resonance imaging data in young adults who are healthy illustrates the interaction between cognitive and motor regions of the brain during voluntary movements in response to external (exogenous) or internal (endogenous) cues. (1) Externally triggered, or exogenous, movements require online processing of sensory cues and invoke activity in a parietal-prefrontal network that both attends to and transforms sensory (in this case, visuospatial) information into a plan for action. In contrast, internally triggered, or endogenous, motor responses benefit from advance planning of movement. In the case of endogenous movements, a frontal-basal ganglia-supplementary and cingulate motor area circuit is activated. The key player for endogenous movements appears to be the basal ganglia, which function to guide learning and performance by determining how much and when motor or cognitive effort are required for successful task completion. (1) Reciprocal connections with the prefrontal, frontal, cingulate, and parietal cortices ideally situate the basal ganglia to exert a strong influence over both cognitive and motor function. (2-5) Indeed, the richness of the neuroanatomic connectivity between the basal ganglia and the cortex has led to speculation that the striatum may participate in the coordination of cognitive and motor information necessary for the production of voluntary movement. (6-9)

One limitation of Pohl et al and colleagues' work is the lack of detail concerning stroke lesion location. In this case, knowledge of lesion location would be informative for 2 reasons. First, it would expand our understanding of which brain regions are necessary for successful task switching. Second, and perhaps more important, knowing which brain regions are invoked during task switching could inform clinical practice. In general, relating stroke location with anticipated behavioral deficits enables the implementation of individualized and precisely formed therapeutic interventions. For example, focal damage to the parietal lobe might disrupt the transformation of sensory (ie, visuospatial) information into the correct plan for movement during task switching. In this case, switching between tasks during rehabilitation might be stimulated by other sensory cues (eg, auditory). Alternately, visuospatial information could be emphasized during physical therapy in an effort to stimulate neuroplastic change in the parietal lobe and associated regions. In either scenario, an understanding of both lesion location and the underlying brain-behavior relationship enables the genesis of informed rehabilitation interventions, specifically designed to address the unique deficits of each client.

Two details of the work by Pohl et al distinguish it as particularly enlightening for physical therapists. Once again, prominent deficits have been noted in individuals with stroke even when they use their less-involved, ipsilesional upper extremity for task completion. The robustness of the finding of ipsilesional deficits are striking and range from those relating to motor control, (10-14) to motor learning, (8,15-18) and now executive function. Taken together, these data indicate that, although motor control of the contralateral hemibody is most obviously affected by stroke, the effect of even unilateral brain damage is widespread and deleterious for both motor and cognitive function.

Secondly, Pohl et al found improvements in response time for task switching for individuals with stroke across the 2 testing time points (1 month to 3 months after stroke). It is possible that reductions in response time across the testing sessions related to physiologic recovery. I think this explanation is unlikely because the healthy control group also improved. More plausible is the possibility that both groups learned something about the task during the first session, which allowed them to improve their performance during the second test. Although not explicitly tested in this study, the incidental demonstration of learning suggests that patients may benefit from practicing skills that specifically require task switching.

Although not the focus of Pohl and colleagues' article, their data raise a host of questions concerning why individuals with stroke demonstrated impaired task switching. For instance, was impaired ability to pay attention the cause of higher switch costs after stroke? Similarly (and not mutually exclusive), were deficits in working memory (the short-term maintenance of information necessary to guide upcoming action (19)) after stroke to blame for poorer accuracy? Last, the authors suggest that impulsivity after stroke might explain their finding of preserved response time but sacrificed accuracy during task switching. However, they offer no independent measure of impulsivity or any of the other executive functions mentioned above. Future work will have to endeavor to address each of the possibilities outlined here as well as consider the effect of impairments in other domains of executive functions on motor recovery after stroke.

It is becoming increasingly important for both physical therapists and research scientists to acknowledge that a wide variety of cognitive factors may critically affect the rehabilitation of motor function, regardless of diagnosis. As knowledge is gained regarding the intricate interconnections between neural regions, it is apparent that damage to both motor and nonmotor brain regions can and does have widespread implications for the functional recovery of voluntary movement. Pohl et al demonstrate the effect of impairments in one domain of executive function: task switching. They also tacitly point out a critical need for more research that considers how stroke and other diagnoses disrupt the elegant relationship between cognitive and motor systems.

As our understanding of these factors expands, we will be contented with the need to design novel interventions that account for disruptions in the cognitive-motor interface and incorporate knowledge of lesion location when it is available.

References

(1) Elsinger CL, Harrington DL, Rat SM. From preparation to online control: reappraisal of neural circuitry mediating internally generated and externally guided actions. Neuroimage. 2006;31:1177-1187.

(2) Cavada C, Goldman-Rakic PS. Posterior parietal cortex in rhesus monkey, II: evidence for segregated corticocortical networks linking sensory and limbic areas with the frontal lobe. J Comp Neurol. 1989;287:422-445.

(3) Cavada C, Goldman-Rakic PS. Topographic segregation of corticostriatal projections from posterior parietal subdivisions in the macaque monkey. Neuroscience. 1991;42: 683-696.

(4) Middleton FA, Strick PL, Obeso JA, et al. New concepts about the organization of basal ganglia outputs. In: Advances in Neurology: Basal Ganglia and New Surgical Treatment of Parkinson's Disease. New York, NY: Lippincott-Raven; 1997: 57-68.

(5) Alexander GE, DeLong MR, Strick PL, Cowan WM. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annu Rev Neurosci. 1986;9:357-381.

(6) Poldrack RA, Clark J, Pare-Blagoev EJ, et al. Interactive memory systems in the human brain. Nature. 2001;414(6863):546-550.

(7) Harrington DL, Haaland KY, Rosenbaum DA, Collyer CE. Sequencing and timing operations of the basal ganglia. In: Timing of Behavior: Neural, Psychological and Computational Perspectives. Cambridge, Mass: MIT Press; 1998:35-61.

(8) Boyd LA, Winstein CJ. Providing explicit information disrupts implicit motor learning after basal ganglia stroke. Learn Mem. 2004; 11:388-396.

(9) Jennings PJ. Evidence of incomplete motor programming in Parkinson's disease. J Mot Behav. 1995;27:310-324.

(10) Pohl PS, Luchies CW, Stoker-Yates J, Duncan PW. Upper extremity control in adults post stroke with mild residual impairment. Neurorehabilitation and Neural Repair. 2000; 14:33-41.

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

(12) Pohl PS, Winstein CJ, OnlaOr S. Sensory-motor control in the ipsilesional upper extremity after stroke. NeuroRehabilitation. 1997;9:245-249.

(13) Velicki MR, Winstein CJ, Pohl PS. Impaired direction and extent specification of aimed arm movements in humans with stroke-related brain damage. Exp Brain Res. 2000;130:362-374.

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

(15) Boyd LA, Winstein CJ. Explicit information interferes with implicit motor learning of both continuous and discrete movement tasks after stroke. Journal of Neurologic Physical Therapy. 2006;30(2):46-57; discussion 58-49.

(16) Boyd LA, Winstein CJ. Implicit motor-sequence learning in humans following unilateral stroke: the impact of practice and explicit knowledge. Neurosci Lett. 2001 ;298:65-69.

(17) Boyd LA, Winstein CJ. Impact of explicit information on implicit motor-sequence learning following middle cerebral artery stroke. Phys Ther. 2003;83:976-989.

(18) Boyd LA, Winstein CJ. Cerebellar stroke impairs temporal but not spatial accuracy during implicit motor learning. Neurorehabilitation and Neural Repair. 2004; 18: 134-143.

(19) Lezak M, Howieson DB, Loring DW. Neuropsychological Assessment. New York, NY: Oxford University Press; 2004.

LA Boyd, PT, PhD, is Canadian Research Chair, Neurobiology of Motor Learning; Assistant Professor, School of Rehabilitation Sciences, University of British Columbia, T325-2211 Westbrook Mall, Vancouver, British Columbia, Canada V6T 2B5; and Adjunct Faculty, Department of Physical Therapy and Rehabilitation Science, University of Kansas Medical Center, Kansas City, KS. Address all correspondence to Dr Boyd at: [e-mail].

DOI: 10.2522/ptj.20060093.ic

Table.
Subject Demographics (a)

Characteristic           Stroke Group         Control Group

Age (y)                  71.4 [+ or -] 9.4    75.6 [+ or -] 6.5
No. of women/men         23/23                30/8
Education level (y)      12.5 [+ or -] 2.8    13.6 [+ or -] 2.4
Mini-Mental State
   Examination
  (maximum score: 30)    27.4 [+ or -] 3.1    28.5 [+ or -] 1.4
Performing hand
  (right/left),
  no. of subjects        21/25                22/16
Stroke severity
  (mild/moderate),
  no. of subjects        25/21

(a) Reported as mean [+ or -] standard deviation unless otherwise
indicated.