Sage Journals: Discover world-class research

Abstract

Background. Phase III trials of rehabilitation of paresis after stroke have proven the effectiveness of intensive and extended task practice, but they have also shown that many patients do not qualify, because of severity of impairment, and that many of those who are treated are left with clinically significant deficits. Objective. To test the value of 2 potential adjuvants to normal learning processes engaged in constraint-induced movement therapy (CIMT): greater distribution of treatment over time and the coadministration of d-cycloserine, a competitive agonist at the glycine site of the N-methyl-D-aspartate glutamate receptor. Methods. A prospective randomized single-blind parallel-group trial of more versus less condensed therapy (2 vs 10 weeks) and d-cycloserine (50 mg) each treatment day versus placebo (in a 2 × 2 design), as potential adjuvants to 60 hours of CIMT. Results. Twenty-four participants entered the study, and 22 completed it and were assessed at the completion of treatment and 3 months later. Neither greater distribution of treatment nor treatment with d-cycloserine significantly augmented retention of gains achieved with CIMT. Conclusions. Greater distribution of practice and treatment with d-cycloserine do not appear to augment retention of gains achieved with CIMT. However, concentration of CIMT over 2 weeks (“massed practice”) appears to confer no advantage either.

Keywords

stroke hemiparesis randomized controlled trial d-cycloserine distributed practice constraint-induced movement therapy

Phase III trials of rehabilitation of paresis following stroke^1-4 have demonstrated the effectiveness of particular rehabilitation strategies and their impact on long-term outcome. However, they have also shown that despite the intensity and dose of the treatments used, impact was modest, and many patients were left with significant impairment. Long-term impairment is even greater in patients who did not have sufficient baseline motor function to qualify for these trials. Thus, any adjuvant to therapies leveraging experience-dependent neuroplasticity, even if it had a modest effect, could be of clinically significant value. In this article, we report the impact of 2 potential adjuvants: 1 pharmacologic (d-cycloserine) and 1 consisting in the potential benefits of greater distribution of treatment over time. Their adjuvant value was tested in a trial employing constraint-induced movement therapy (CIMT): a thoroughly vetted rehabilitation treatment for upper extremity paresis that reliably yields, on average, a clinically significant benefit, thereby qualifying as a behavioral engine for testing adjuvant treatments.⁵ It is the reliable benefit of CIMT that can potentially be amplified by an adjuvant treatment.

Distributed Practice

Massed practice is traditionally defined as continuous practice conducted over a single day in which the stimuli or skills to be retained are presented without any intervening irrelevant stimuli or activities. Distributed practice is traditionally defined as practice that is broken up into 2 or more sessions, each separated by an interlude of minutes, hours, or days, or as single-session practice in which repetitions of target stimuli are separated by 1 or more nontarget stimuli. In neurorehabilitation, target stimulus repetitions are typically separated by many intervening nontarget stimuli, and limits of human endurance dictate that sessions be conducted over multiple days. Thus, no rehabilitation therapy conforms to traditional definitions of massed practice, and our research focused rather on whether therapeutic efficacy is optimized by more or less condensed treatment. Studies of CIMT for upper extremity paresis following stroke may have cultivated the notion that the case for large amounts of highly condensed practice has largely been proven. However, this hypothesis has never been tested.

Studies contrasting the relative impact of massed practice and distributed practice on learning rate and knowledge retention date back to 1885.^6-10 Literally hundreds of investigations have been conducted involving a range of knowledge and skill types. With very few exceptions, distributed practice has been found to be superior to massed practice in its effects on knowledge retention, whether the training involved factual knowledge, simple motor skills, or more complex tasks involving acquisition of factual knowledge, motor skills, and the development of strategies. This superiority has come to be known as the spacing effect. A small number of studies have tested the relative value of distributing learning in normal subjects over long intervals, such as days, weeks, or months—epochs that might be germane to neurorehabilitation.^11-17 In all, superior retention was achieved with more distributed practice. Several studies have demonstrated that delayed recall is superior with distributed practice in populations of brain-injured participants, including patients with the amnestic syndrome,¹⁸ traumatic brain injury,^19-21 and multiple sclerosis.²² However, in all these studies, distributed practice involved repetition intervals of minutes, not days, and the learning tasks were constituted as tests of the principle, not part of a rehabilitation program. The testing of the distributed practice effect employing distribution over weeks in the context of a rehabilitation program, as described in this report, appears to be novel.

A number of psychological theories have been offered to explain the spacing effect,^10,23-27 and there is good evidence of molecular mechanisms as well.²⁸ One group of theories (contextual variability theories) has a clear bearing on training distributed over days to weeks. One particular contextual variability theory—the component levels theory of Glenberg—has stood the test of time and received extensive empirical validation.^25,26 Glenberg posited that at times of learning, 3 levels of information are encoded: substantive, structural context, and general context. Substantive information might consist of the semantic features of the stimulus or the procedural components of the skill. Structural context refers to the substance and arrangement of other stimuli in the training set (eg, other words in a word list to be memorized or other stimulus pairs in a native word–foreign word learning paradigm). It might also consist of parametric variations of a skill being trained (eg, shooting basketballs from different distances and different positions). General context refers to attributes of the training environment (eg, room, trainer, training equipment), affective state of the subject, travel entailed to the training test site, the subject’s familial situation at the time, people accompanying the subject to the training site, and so on. Glenberg posited that recall is aided to the extent that there are contextual commonalities between the time of training and the time of recall. Training multiple times at substantial intervals (less condensed practice) aids recall by increasing the probability that there will be elements of commonality, structural but particularly general, between the contexts at one or more training sessions and the context at time of retrieval—elements that will aid retrieval.

Greater distribution of training also allows for more intervening nights of sleep, during which both declarative and procedural memories are consolidated. CIMT likely engages both declarative and procedural memory mechanisms. Sleep is emerging as a major engine of neuroplasticity in general and memory consolidation in particular.^29-31 Bell et al³² found that intervening sleep and spacing appeared to make independent contributions to long-term retention.

The first hypothesis motivating this study was that, based on the theoretical rationale for the superiority of more distributed (less condensed) practice and the extensive empirical evidence from studies in normal subjects, more distributed practice would be more effective in brain injured subjects.

D-cycloserine

It is likely that 6 or more months postinjury, normal learning mechanisms (experience-dependent plasticity) are primarily responsible for recovery. Learning, whether declarative or procedural, represents alteration of neural connection strengths. This occurs through the processes underlying long-term potentiation (increase in neural connection strengths) and long-term depression (decrease in neural connection strengths).³³ The N-methyl-D-aspartate (NMDA) glutamate receptor, which gates a sodium-calcium channel, has been identified as a key component of the mechanisms underlying long-term potentiation. Normal learning processes mediated via the NMDA glutamate receptor can potentially be amplified by potentiation of NMDA glutamate receptor–mediated processes. D-cycloserine, a drug long used as an antibiotic in the treatment of tuberculosis, is a high-affinity partial agonist with moderate specificity at the glycine-B (strychnine insensitive) receptor of the NMDA-glutamate receptor complex.^34,35 It is therefore capable of potentiating NMDA receptor channel activity without activating glycine-A (strychnine sensitive) receptors in the brainstem and spinal cord, which have inhibitory effects on neuronal firing. Extensive studies in normal animals (primarily rodents but also nonhuman primates) and animals with various lesions do suggest that d-cycloserine can enhance learning in a large variety of tasks. In studies of normal human subjects, d-cycloserine has been shown to potentiate declarative³⁶ and procedural memory formation³⁷ and to improve memory function in normal subjects treated with the anticholinergic drug scopolamine.³⁸

D-cycloserine has never been tested as an adjuvant to rehabilitation of stroke-related impairment. Our second hypothesis in the present trial was that d-cycloserine would be an effective adjuvant to CIMT for upper extremity paresis after stroke—by virtue of its mechanism of action, the extensive evidence of adjuvant effects on learning in animal studies, and its reported effects on declarative and procedural memory formation in human subjects.

We tested these 2 hypotheses in a prospective parallel-group randomized controlled trial employing a crossed design—d-cycloserine/placebo (double blinded) × more versus less condensed practice (single blinded)—in which all subjects received 60 hours of CIMT but those in the more condensed group received therapy 6 hours a day, 5 days a week, for 2 weeks while those in the less condensed group received therapy 2 hours a day, 3 days a week, for 10 weeks. This randomized controlled trial is best characterized as a phase IIa study: it was designed to determine the presence of an effect of the 2 interventions and to establish parameters for a power analysis for a subsequent larger trial. It was not expected that the results of this trial would achieve statistical significance.

Methods

Participants

Participants were recruited from inpatient and outpatient populations at the Malcom Randall VA Medical Center (Gainesville, Florida) and the Brooks Rehabilitation Hospital (Jacksonville, Florida), through newspaper articles and advertisements, and via contact with stroke support groups. The study was conducted at the VA Rehabilitation Research and Development Brain Rehabilitation Research Center at Malcom Randall VA and at outpatient clinics of the Brooks Rehabilitation Hospital. Twenty-four participants with moderate upper extremity paresis due to a stroke that had occurred at least 6 months prior were recruited to the study. Participants who had experienced 1 or more strokes on the same side of the brain were included unless there was a history of a clinical ischemic or hemorrhagic event affecting the other hemisphere or there was imaging evidence of more than a lacune or minor ischemic demyelination affecting the other hemisphere. Participants had to be between the ages of 18 and 80 years. There were no restrictions related to sex, ethnicity, or handedness.

Participants had to meet the following motor and functional criteria to enter the study:

Active motions of the wrist and hand: 20° of wrist extension from a relaxed flexed position, 10° of extension of any 2 digits at any joint, and 10° of thumb extension at either joint. All active motions had to be repeated 3 times within 1 minute.

Passive range of motion: 90° of flexion and abduction and 45° external and internal rotation at the shoulder, 45° elbow supination and pronation, elbow extension limited by no more than 30°, wrist extension to at least neutral position, and digit extension limited by no more than 30°. Participants were not required to exhibit any active shoulder or elbow motion.

Ability to sit independently for at least 2 minutes. Ability to ambulate was not required.

Motor Activity Log³⁹ score of 3 or less.

Exclusionary criteria included a history of a learning disorder, mental retardation, drug or alcohol abuse, more than very minor head trauma, subarachnoid hemorrhage, lobar cerebral hemorrhage, severe chronic obstructive pulmonary disease, cardiac dysrhythmias requiring medical treatment, serious medical illness, dementia or other neurodegenerative disease, multiple sclerosis, epilepsy, medically resistant depression prior to stroke, schizophrenia, severe visual impairment, pregnancy, breast feeding, and participation in CIMT within the prior 12 months. Serum creatinine documented within the prior 6 months had to be ≤1.5 mg/dL (d-cycloserine is renally cleared). Aphasia and other cognitive deficits did not preclude inclusion so long as subjects were sufficiently sentient to be able to understand the potential risks and benefits of the study, personally provide informed consent, and understand and cooperate with the treatment.

This study was approved and monitored by the Investigational Review Board of the University of Florida Health Science Center. All participants provided informed consent.

Randomization and Drug Treatment

This was a prospective randomized parallel-group study employing a 2 × 2 design: d-cycloserine/placebo × more/less condensed therapy. Subjects were randomized to 1 of the 4 treatment groups by the research pharmacist. Double blinding of drug treatment was maintained until all participants had undergone their final outcome evaluation. Because therapists could not be blinded to distribution of treatment, all outcome measurements were conducted by blinded evaluators who had no other role in the trial.

Participants took d-cycloserine (50 mg) by mouth on therapy days only, 4 hours before beginning therapy, to ensure maximal blood and brain levels of the drug over the treatment session. D-cycloserine and identical placebo containing capsules were prepared by a compounding pharmacy. The drug was used under an Food and Drug Administration investigatory drug exemption. Compliance was ensured through pill counts conducted by study therapists.

All participants underwent 60 hours of CIMT. More condensed treatment consisted of 6 hours per day for 10 consecutive weekdays. Less condensed treatment consisted of 2 hours per day, 3 days a week (Monday, Wednesday, Friday), for 10 consecutive weeks.

CIMT^40,41 was conducted by an occupational therapist, physical therapist, or physical therapy assistant. Short breaks in treatment were provided in all groups when needed. The actual time in therapy was recorded for each session. Total time in therapy, through the course of the trial, was documented to be similar for all groups: more condensed, placebo, 59.2 ± 1.8 hours (range, 55.5-60); more condensed, d-cycloserine, 60.5 ± 0.1 hours (range, 60-60.05); less condensed, placebo, 58.6 ± 2.7 hours (range, 54-61); less condensed, d-cycloserine, 59.3 ± 1.9 hours (range, 55.5-60.33). Participants contracted to wear the mitt on the unaffected hand 90% of their total waking hours throughout the course of treatment, regardless of whether it was a treatment day, except when use of that hand was required for considerations of safety, hygiene, and specifically agreed on activities, such as driving to and from the lab. A home diary was used to document daily compliance in wearing the mitt. Unfortunately, on examination, the diaries were insufficiently complete to enable quantitative assessment of compliance. All subjects, regardless of group assignment, went home following treatment. Following completion of CIMT, subjects were contacted by research therapists only to schedule follow-up testing.

CIMT included massed repetition, task practice with shaping (approach of a desired motor or behavioral objective by small steps and successive approximations), and intensive timed activities. The daily interventions were designed around a menu of functional activities that incorporated variations of strength, endurance, coordination, dexterity, and range of motion. Interest inventories and role checklists were used to incorporate the participant’s unique personality traits and activity interests into the menu. The participants gave daily input regarding the activities that they performed in the laboratory and when they returned home. In this way, a variety of purposeful and meaningful activities were incorporated with attention to unique limitations. Encouragement was provided continuously. Therapists sought to prevent participants from failing by providing assistance as necessary or changing the task. If an activity was strongly disliked or appeared to be too difficult, it was eliminated from the menu for the next day.

Outcome Measures

The primary outcome measure was time on the Wolf Motor Function Test (WMFT) for the affected limb.^39,42 Secondary outcome measures included grip strength, the Fugl-Meyer Motor Scale–Upper Extremity,⁴³ the Motor Activity Log–quality,³⁹ the Stroke Impact Scale (version 2.0; participation—ability to participate in activities that are meaningful and give purpose to life; recovery—analog scale rating of degree of recovery from stroke),⁴⁴ Caregiver Strain Index,⁴⁵ and the Geriatric Depression Scale.⁴⁶ Grip strength was measured 3 times during each administration of the WMFT and the 3 values averaged. The WMFT, Motor Activity Log, and Fugl-Meyer were obtained at baseline, end of treatment, and 3 months after completion of treatment. At each assessment, the WMFT was conducted twice and the times averaged. The Stroke Impact Scale and the Caregiver Strain Index were obtained at baseline and 3 months after completion of treatment. At the end of treatment, participants were asked to guess whether they had been receiving drug or placebo.

Statistical Analysis

Statistical analyses were conducted with SAS 9.2 (SAS Institute, Cary, North Carolina). Means ± SDs were computed for continuous baseline variables, and outcomes and percentages were computed for binary variables for the 4 treatment groups. The balance of baseline characteristics for the 2 group comparisons (d-cycloserine vs placebo; more vs less condensed) was compared with a t test for continuous variables and χ² or Fisher’s exact tests for categorical variables.

A general linear model was used to assess the impact of drug (d-cycloserine vs placebo) and distribution of practice group (more versus less condensed) and their interaction on WMFT score at 3 months after the end of treatment, controlling for baseline WMFT score. A paired t test was used to compare baseline with 3-month posttreatment WMFT scores for the entire group. Three months was chosen as the outcome point of interest because the impact of the distribution of practice effect theoretically affects retention of knowledge and skill, not acquisition of knowledge and skill. An intention-to-treat analysis was employed. Missing values were handled through last observation carried forward. Effect sizes, denoted as the semipartial statistic ${\hat{η}}^{2}$ , were calculated to determine the proportion of variance accounted for by the independent variable being tested, based on its sum of squares and the total corrected sum of squares. No statistical corrections were made for multiple comparisons.

Results

Of the 24 participants who entered the trial, 22 completed it (Figure 1). Baseline values for the comparison groups are detailed in Tables 1 and 2. No between-group differences were statistically significant. The participants in the more condensed group were, on average. somewhat more impaired (WMFT, 40.0 ± 33.1) than those in the less condensed group (25.2 ± 31.1), but the difference failed to achieve significance. Likewise, the participants in the placebo group were, on average, more severely impaired (WMFT, 38.6 ± 32.2) than those in the d-cycloserine group (26.7 ± 32.7), but again the difference was not significant.

Figure 1.

CONSORT diagram.

Table 1.

Baseline Values: More Versus Less Condensed^a.

Variable	All (N = 24)	More Condensed (n = 12)	Less Condensed(n = 12)	P
Age	58.2 ± 8.9	58.0 ± 8.4	58.4 ± 9.7	.911
Female	12 (50)	6 (50)	6 (50)	1.000
Race
Afro-American	6 (25)	3 (25)	3 (25)	1.000
Caucasian	18 (75)	9 (75)	9 (75)
Hispanic	3 (12.5)	0	3 (25)	.217
Right handedness	23	12	11
Education, y	14 ± 2.0	14.1 ± 2.0	13.9 ± 2.1	.874
Months poststroke	36.3 ± 4.1	36.6 ± 4.3	35.9 ± 41.6	.969
Side of paresis
Left	10 (41.7)	6 (50)	4 (33.3)	.408
Right	14 (58.3)	6 (50)	8 (66.7)
Type of stroke
Hemorrhage	5 (2.8)	2 (16.7)	3 (25)	.221
Lacune	8 (33.3)	6 (50)	3 (16.7)
Large vessel	11 (45.8)	4 (33.3)	7 (58.3)
Previous CIMT, Yes	4 (16.7)	2 (16.7)	2 (16.7)	1.000
Treatment site
Malcom Randall VA	3 (12.5)	1 (8.3)	2 (16.7)	.537
Brooks Rehabilitation	21 (87.5)	11 (91.7)	10 (83.3)
WMFT	32.6 ± 32.3	40.0 ± 33.1	25.2 ± 31.1	.271
MAL-Q	1.2 ± 0.8	0.9 ± 0.6	1.5 ± 0.9	.078
FM-UE	38.4 ± 12.9	33.9 ±13.0	42.9 ± 11.5	.086
GDS	10.0 ± 6.3	9.4 ± 6.2	10.7 ± 6.6	.630
SIS-participation	24.1 ± 6.2	22.6 ± 5.5	25.7 ± 6.7	.228
SIS-recovery	53.0 ± 23.4	49.1 ± 2.6	56.5 ± 26.1	.461

Abbreviations: CIMT, constraint-induced movement therapy; GDS, Geriatric Depression Scale; FM-UE, Fugl-Meyer Motor Scale–Upper Extremity; MAL-Q, Motor Activity Log–quality; SIS, Stroke Impact Scale; WMFT, Wolf Motor Function Test.

Values in No. (%) or mean ± SD. Caregiver Strain Index data were unavailable for 6 participants who did not have caregivers; therefore, we have not reported these results.

Table 2.

Baseline Values: Placebo Versus D-cycloserine^a.

Variable	All (N = 24)	Placebo (n = 12)	D-cycloserine (n = 12)	P
Age	58.2 ± 8.9	57.3 ± 11.5	59.1 ± 5.6	.639
Female	12 (50)	5 (41.7)	7 (58.3)	.414
Race
Afro-American	6 (25)	4 (33.3)	2 (16.7)	.640
Caucasian	18 (75)	8 (66.7)	10 (83.3)
Hispanic	3 (12.5)	2 (16.7)	1 (8.3)	1.000
Right handedness	23	12	11
Education, y	14 ± 2.0	14.1 ± 2.4	13.9 ±1.8	.877
Months poststroke	36.3 ± 40.1	45.9 ± 49.2	26.6 ± 27	.246
Side of paresis
Left	10 (41.7)	5 (41.7)	5 (41.7)	1.000
Right	14 (58.3)	7 (58.3)	7 (58.3)
Type of stroke
Hemorrhage	5 (20.8)	3 (25)	2 (16.7)	.468
Lacune	8 (33.3)	5 (41.7)	3 (25)
Large vessel	11 (45.8)	4 (33.3)	7 (58.3)
Previous CIMT
Yes	4 (16.7)	3 (25)	1 (8.3)	.273
Treatment site
Malcom Randall VA	3 (12.5)	2 (16.7)	1 (8.3)	.537
Brooks Rehabilitation	21 (87.5)	10 (83.3)	11 (91.7)
WMFT	32.6 ± 32.3	38.6 ± 32.2	26.7 ± 32.7	.381
MAL-Q	1.2 ± 0.8	1.0 ± 0.8	1.3 ± 0.8	.370
FM-UE	38.4 ± 12.9	34.5 ± 10.7	42.3 ± 14.0	.139
GDS	10.0 ± 6.3	11.1 ± 7.5	9.1 ± 5.1	.456
SIS-participation	24.1 ± 6.2	25.9 ± 6.5	22.3 ± 5.5	.159
SIS-recovery	53.0 ± 23.4	48.8 ± 22.9	57.5 ± 24.2	.380

Values in No. (%) or mean ± SD.

The results of the analysis of treatment effects for primary and secondary outcome measures are detailed in Tables 3 and 4. The general linear model demonstrated no significant effects of training distribution group, d-cycloserine, or their interaction on WMFT score at 3 months posttreatment, controlling for baseline WMFT (Table 5). There was a robust benefit from CIMT, as reflected in substantial gains in all subgroups shown in Table 3 and 4, and a decrease in WMFT scores of 11.7 ± 15.7 (P = .0013; ${\hat{η}}^{2}$ = 0.75).

Table 3.

Outcomes: More Versus Less Condensed.

Variable	All (N = 24)	More Condensed (n = 12)	Less Condensed (n = 12)	P
WMFT
Baseline	32.6 ± 32.3	40.0 ± 33.1	25.2 ± 31.1	.271
Posttreatment	21.9 ± 26.8	28.1 ± 30.0	15.7 ± 22.7	.265
3 mo posttreatment	20.9 ± 27.6	27.6 ± 31.6	14.3 ± 22.3	.246
Baseline–posttreatment Δ	−10.8 ± 13.1	−11.9 ± 14.3	−9.6 ± 12.2	.665
Baseline–3 mo posttreatment Δ	−11.7 ± 15.7	−12.4 ± 17.5	−11.0 ± 14.5	.823
Fugl-Meyer–upper extremity
Baseline	38.4 ± 12.9	33.9 ± 13.0	42.9 ± 11.5	.086
Posttreatment	42.6 ± 12.4	38.8 ± 12.1	46.4 ± 12.0	.135
3 mo posttreatment	42.0 ± 12.2	37.2 ± 11.9	46.8 ± 10.9	.051
Baseline–posttreatment Δ	4.1 ± 6.6	4.8 ± 6.1	3.4 ± 7.2	.616
Baseline–3 mo posttreatment Δ	3.6 ± 7.1	3.3 ± 7.2	3.9 ± 7.4	.832
MAL-Q
Baseline	1.2 ± 0.8	0.9 ± 0.6	1.5 ± 0.9	.078
Posttreatment	2.5 ± 0.9	2.2 ± 0.6	2.8 ± 1.1	.083
3 mo posttreatment	2.1 ± 1.1	1.5 ± 0.5	2.6 ± 1.3	.012
Baseline–posttreatment Δ	1.3 ± 0.7	1.3 ± 0.7	1.4 ± 0.8	.760
Baseline–3 mo posttreatment Δ	0.9 ± 0.8	0.6 ± 0.6	1.1 ± 0.9	.095
Grip
Baseline	12.7 ± 8.5	12.0 ± 10.2	13.4 ± 6.7	.687
Posttreatment	13.4 ± 9.9	11.5 ± 9.1	15.2 ± 10.6	.370
3 mo posttreatment	14.6 ± 11.0	17.2 ± 14.5	11.9 ± 5.3	.248
Baseline–posttreatment Δ	0.7 ± 5.4	−0.5 ± 4.9	1.8 ± 5.9	.317
Baseline–3 mo posttreatment Δ	1.8 ± 10.5	5.2 ± 13.9	−1.5 ± 3.2	.116
SIS-participation
Baseline	24.1 ± 6.2	22.6 ± 5.5	25.7 ± 6.7	.228
3 mo posttreatment	26.9 ± 6.9	26.2 ± 5.5	27.6 ± 8.2	.616
Baseline–3 mo posttreatment Δ	2.8 ± 6.4	3.6 ± 4.5	2.0 ± 8.0	.546
SIS-recovery
Baseline	52.8 ± 22.9	49.1 ± 19.6	56.5 ± 26.1	.440
3 mo posttreatment	58.0 ± 19.8	50.5 ± 15.7	65.5 ± 21.2	.061
3 mo posttreatment Δ	5.2 ± 23.2	1.4 ± 24.2	9.0 ± 22.5	.435

Abbreviations: MAL-Q, Motor Activity Log–quality; SIS, Stroke Impact Scale; WMFT, Wolf Motor Function Test.

Table 4.

Outcomes: Placebo Versus D-cycloserine.

Variable	All (N = 24)	Placebo (n = 12)	D-cycloserine (n = 12)	P
WMFT
Baseline	32.6 ± 32.3	38.6 ± 32.2	26.7 ± 32.7	.381
Posttreatment	22.3 ± 26.8	26.7 ± 29.4	18.0 ± 24.3	.441
3 mo posttreatment	20.9 ± 27.6	22.9 ± 28.9	19.0 ± 27.3	.738
Baseline–posttreatment Δ	−10.3 ± 12.2	−11.9 ± 12.3	−8.7 ± 12.4	.532
Baseline–3 mo posttreatment Δ	−11.7 ± 16.2	−15.7 ± 14.5	−7.7 ± 17.4	.235
Fugl-Meyer–upper extremity
Baseline	38.4 ± 12.9	34.5 ± 10.7	42.3 ± 14.0	.139
Posttreatment	42.2 ± 12.4	38.8 ± 10.9	45.7 ± 13.3	.181
3 mo posttreatment	42.0 ± 12.2	39.2 ± 11.5	44.8 ± 12.6	.265
Baseline–posttreatment Δ	3.8 ± 6.0	4.3 ± 6.5	3.3 ± 5.7	.695
Baseline–3 mo posttreatment Δ	3.6 ± 7.5	4.7 ± 5.6	2.5 ± 9.1	.484
MAL-Q
Baseline	1.1 ± 0.7	1.0 ± 0.7	1.3 ± 0.7	.283
Posttreatment	2.6 ± 1.0	2.3 ± 1.0	3.0 ± 0.8	.068
3 mo posttreatment	2.0 ± 1.2	1.8 ± 1.2	2.3 ± 1.2	.349
Baseline–posttreatment Δ	1.5 ± 0.7	1.3 ± 0.6	1.7 ± 0.8	.197
Baseline–3 mo posttreatment Δ	0.9 ± 0.8	0.8 ± 0.7	1.0 ± 1.0	.704
Grip
Baseline	12.7 ± 8.5	13.8 ± 9.6	11.6 ± 7.5	.529
Posttreatment	13.3 ± 9.9	12.8 ± 7.4	13.7 ± 12.2	.838
3 mo posttreatment	14.6 ± 11.0	15.5 ± 13.8	13.6 ± 7.8	.691
Baseline–posttreatment Δ	0.6 ± 5.5	−1.0 ± 5.0	2.1 ± 5.8	.173
Baseline–3 mo posttreatment Δ	1.8 ± 10.2	1.6 ± 13.5	2.1 ± 6.0	.924
SIS-participation
Baseline	24.1 ± 6.2	25.9 ± 6.5	22.3 ± 5.5	.159
3 mo posttreatment	26.9 ± 6.9	28.3 ± 6.4	25.5 ± 7.4	.344
Baseline–3 mo posttreatment Δ	2.8 ± 6.4	2.4 ± 6.3	3.2 ± 6.7	.751
SIS-recovery
Baseline	53.1 ± 22.9	48.8 ± 22.9	57.5 ± 23.0	.358
3 mo posttreatment	58.0 ± 19.7	59.5 ± 17.4	56.4 ± 22.4	.701
Baseline–3 mo posttreatment Δ	4.8 ± 22.7	10.8 ± 23.8	−1.2 ± 20.9	.203

Abbreviations: MAL-Q, Motor Activity Log–quality; SIS, Stroke Impact Scale; WMFT, Wolf Motor Function Test.

Table 5.

Regression Analysis: Wolf Motor Function Test Results at 3 Months Posttreatment.

Parameter	Estimate	SE	F	P	Semipartial Eta-Square^a
Wolf Motor Function Test baseline	0.76	0.10	61.21	<.0001	0.707
Group (less condensed = 1)	0.83	8.54	0.12	.736	0.001
D-cycloserine (placebo = 1)	−7.80	8.47	1.23	.281	0.014
Group × d-cycloserine	2.45	11.76	0.04	.837	0.001

Total variation accounted for.

At the end of treatment, participants were asked whether they thought they had been receiving drug or placebo. Of the 12 participants taking d-cycloserine, 10 thought that they were receiving placebo, and all 12 participants taking placebo thought that they were receiving placebo.

Safety

One participant experienced a stroke during the trial (not during actual treatment). There were no adverse events associated with d-cycloserine. One participant in the more condensed training group experienced a small skin tear and a sore elbow.

Discussion

We posited that less condensed training would yield greater retention of results of treatment 3 months later because of the greater likelihood of commonality between treatment conditions and test conditions and the greater number of intervening nights of sleep. We also posited that d-cycloserine would increase learning because of its potentiation of sodium-calcium influx through NMDA-glutamate voltage-gated sodium-calcium channels, which are crucial to learning. Neither hypothesis was supported by the study results. Yet, our results provide no support for the concept that CIMT must be concentrated into 2 weeks of training. The results of distributing the 60 hours of treatment over 10 weeks were the same as when the therapy was provided over 2 weeks. D-cycloserine proved to be safe.

Potential Reasons for Treatment Failure

Both the adjuvants to CIMT employed in this trial could have failed if participants failed to improve with CIMT because both adjuvants potentially affect only the neuroplastic response of the central nervous system. However, our results demonstrate that there was a robust benefit from CIMT, as reflected in substantial gains on all measures and a substantial, statistically significant decline in WMFT score (effect size, 0.75). Thus, there was a substantial neuroplastic response susceptible to influence by the adjuvants.

It is possible that the adjuvants tested here could achieve a small but significant impact in a less severely impaired population. Mean baseline WMFT score in this study was 32.6 ± 32.3, and baseline upper extremity Fugl-Meyer score was 38.4 ± 12.9, reflecting moderate to severe impairment for a CIMT-eligible population.

Both adjuvants could have failed if the neuroplastic response elicited by CIMT was approaching ceiling. There is increasing evidence that paresis caused by stroke is, in nearly all cases, primarily due to a white matter lesion.⁴⁷ The loss of most corticospinal and corticobulbar fibers projecting from the hemisphere to the brainstem may impose strong limits on just how much recovery can be achieved by behavioral treatments, however intensive and extended, as perhaps best reflected in the results of recent phase III trials of intensive therapy for upper extremity paresis and gait impairment following stroke.^1-4

The apparent failure of both adjuvants to augment acquisition and retention of functional gains achieved during therapy could reflect inadequate statistical power. This goes to the heart of a methodological issue that poses a major challenge for neurorehabilitation science. The actual cost of delivering treatment over a less condensed schedule or coadministering a drug such as d-cycloserine would be very small and could readily be defended on the basis of small marginal gains. However, to achieve 80% power to detect the observed gains in WMFT attained by administration of d-cycloserine (−11.9 vs −8.7) and less condensed training (−11.9 vs −9.6), a randomized controlled trial identical in design to ours would require 474 and 1038 participants, respectively.

Greater distribution of treatment over time could have failed to yield a greater retention of benefit for a number of reasons. Our interval from posttreatment to retention test (3 months) may have been too short to allow for differential attenuation of gains in the more condensed treatment group. The effects of more distributed learning may have been too small in a clinical population to be measurable, given all the other potential sources of variance. Less condensed treatment could have been associated with reduced concentration on therapy, loss of continuity, reduced involvement by caregivers, and loss of motivation. However, these effects should have been reflected in poorer acquisition by this group, something not apparent in our data. There is evidence that the benefits of distributed learning are relatively less in older adults than in young adults.⁴⁸ Gains in function achieved with CIMT substantially involve procedural memory acquisition. It is possible that contextual variability effects on procedural memory acquisition (in contrast to declarative memory acquisition) are too small to be detected in a small study such as ours or, conceivably, too small to be relevant to any rehabilitation that predominantly involves procedural memory acquisition. Finally, it is possible that distributing therapy over a longer period was not a sufficiently potent factor in establishing commonality between circumstances of treatment and retention. Conducting the treatment in the participants’ homes might be more effective in achieving this objective. This may, for example, have been an important factor in the LEAPS trial, in which the participants, though encouraged to walk outside formal therapy, actually received impairment-based therapy delivered entirely in their homes that included no walking. They did as well as those receiving functional task practice in the clinic, which included body weight–supported treadmill training and overground walking.¹

D-cycloserine treatment could have failed to yield greater learning and retention for several reasons. First, learning processes engaged by CIMT may already have been optimal at the molecular level and may not have been susceptible to improvement by a drug. Second, the dose of d-cycloserine that we used may not have been optimal, despite our efforts to pick the best dose based on extant animal studies and human subject trials. However, a recent study suggests that a higher dose (250 mg) may be required to facilitate declarative memory acquisition,³⁶ even as lower doses (eg, 100 mg) may suffice for enhancement of procedural memory formation.³⁷ Finally, repeated drug exposure may have led to tachyphylaxis, particularly in the more condensed treatment group, as has been reported in animal studies.^49,50 However, given the 10-hour half-life of d-cycloserine (http://www.medicines.org.uk/emc/medicine/20415/SPC), one would expect that 80% of the drug would be eliminated by the next daily dose. Furthermore, we did not observe a differential response to the drug in the more and less condensed groups, even though those in the less condensed group also received more total drug.

Finally, because of the shortcomings of home diaries, our study could not quantify the potential effects of ambient therapy—that is, therapy implicit in activity occurring outside treatment sessions. Ambient therapy could have made a major contribution to outcome,^5,51 particularly in less condensed therapy groups. It is likely that quantitative methods employing such devices as accelerometers will be needed to adequately measure ambient therapy in trials of upper extremity rehabilitation.

Trial Design

Because of the theoretical orthogonality of distribution of practice and d-cycloserine effects, we set up a 2 × 2 trial design in the interests of greater efficiency: we could perform 2 parallel-group randomized controlled trials at the same time, each involving 24 subjects. Our a priori assumption of orthogonality was borne out by the lack of a significant interaction effect. Notwithstanding the small cell size, the use of covariate adaptive randomization could have yielded a better balance of severity across groups.

Conclusions

In this pilot randomized controlled trial, we found no impact of greater distribution of CIMT over time or the administration of d-cycloserine on the retention of functional improvement achieved through therapy. However, we also found that there was no advantage to the massed practice of traditional CIMT programs.

Footnotes

Acknowledgements

We are grateful to Matt Morrow, the study pharmacist; Floris Singletary, our clinical coordinator at Brooks Rehabilitation Hospital; Carolyn Hanson, our clinical coordinator at the Malcom Randall VA Medical Center; and the therapists who made this study possible—Molly Dunn, Brooke Hoisington, and Margo Fitch.

Authors’ Note

Study data can be obtained by contacting the principal investigator.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Supported by Department of Veterans Affairs Rehabilitation R&D grant B6346R: Improving stroke rehabilitation: spacing effect and d-cycloserine (Clinicaltrials.gov NCT00711568).

References

Duncan

Sullivan

Behrman

. Body-weight–supported treadmill rehabilitation after stroke. N Engl J Med. 2011;364:2026-2036.

Guarino

Richards

. Effectiveness of robot-assisted neurorehabilitation for upper extremity impairment in chronic stroke: VA Cooperative Study Program #558. N Engl J Med. 2010;362:1772-1783.

Wolf

Winstein

Miller

. Effect of constraint-induced movement therapy on upper extremity function 3 to 9 months after stroke: the EXCITE randomized clinical trial. JAMA. 2006;296:2095-2104.

Nadeau

Dobkin

. Functional walking ability 2 to 6 months after stroke: effects of progressive task-specific and impairment-based exercise rehabilitaiton programs compared to usual care. Neurorehabil Neural Repair. 2013;27:370-380.

Nadeau

. CIMT as a behavioral engine in research on physiological adjuvants to neurorehabilitation: the challenge of merging animal and human research. NeuroRehabilitation. 2006;21:107-30.

Dempster

. Distributing and managing the conditions of encoding and practice. In: Bjork

Bjork

, eds. Memory. San Diego, CA: Academic Press; 1996:317-344.

Donovan

Radosevich

. A meta-analytic review of the distribution of practice effect: now you see it, now you don’t. J Appl Psychol. 1999;84:795-805.

Ebbinghaus

. Memory: A Contribution to Experimental Psychology. New York, NY: Dover; 1964.

Cepeda

Coburn

Rohrer

Wixted

Mozer

Pashler

. Optimizing distributed practice: theoretical analysis and practical implications. Exp Psychol. 2009;56:236-246.

10.

Cepeda

Pashler

Vul

Wixted

Rohrer

. Distributed practice in verbal recall tasks: a review and quantitative synthesis. Psychol Bull. 2006;132:354-380.

11.

Pyle

. Economical learning. J Educ Psychol. 1913;4:148-158.

12.

Murphy

. Distribution of practice periods in learning. J Educ Psychol. 1916;7:150-162.

13.

Shea

Lai

Black

Park

J-H

. Spacing practice sessions across days benefits the learning of motor skills. Hum Mov Sci. 2000;19:737-760.

14.

Shebilske

Goettl

Corrington

Day

. Interlesson spacing and task-related processing during complex skill acquisition. J Exp Psychol Appl. 1999;5:413-437.

15.

Childers

Tomasello

. Two-year-olds learn novel nouns, verbs, and conventional actions from massed or distributed exposures. Dev Psychol. 2002;38:967-978.

16.

Bahrick

Phelps

. Retention of Spanish vocabulary over 8 years. J Exp Psychol Learn Mem Cogn. 1987;13:344-349.

17.

Baddeley

Longman

DJA

. The influence of length and frequency of training session on the rate of learning to type. Ergonomics. 1978;21:627-635.

18.

Cermak

Verfaellie

Lanzoni

Mather

Chase

. Effect of spaced repetitions on amnesia patients’ recall and recognition performance. Neuropsychology. 1996;10:219-227.

19.

Goverover

Arango-Lasprilla

Hillary

Chiaravalloti

DeLuca

. Application of the spacing effect to improve learning and memory for functional tasks in traumatic brain injury: a pilot study. Am J Occup Ther. 2009;63:543-548.

20.

Hillary

Schultheis

Challis

Carnevale

Galshi

DeLuca

. Spacing of repetitions improves learning and memory after moderate and severe TBI. J Clin Exp Neuropsychol. 2003;25:49-58.

21.

Sumowski

Wood

Chiaravalloti

Wylie

Lengenfelder

DeLuca

. Retrieval practice: a simple strategy for improving memory after traumatic brain injury. J Int Neuropsychol Soc. 2010;16:1147-1150.

22.

Goverover

Hillary

Chiaravalloti

Arango-Lasprilla

DeLuca

. A functional application of the spacing effect to improve learning and memory in persons with multiple sclerosis. J Clin Exp Neuropsychol. 2009;31:513-522.

23.

Cepeda

Vul

Rohrer

Wixted

Pashler

. Spacing effects in learning: a temporal ridgeline of optimal retention. Psychol Sci. 2008;19:1095-1102.

24.

Raaijmakers

JGW

. Spacing and repetition effects in human memory: application of the SAM model. Cogn Sci. 2003;27:431-452.

25.

Glenberg

. Component-levels theory of the effects of spacing of repetitions on recall and recognition. Memory Cognition. 1979;7:95-112.

26.

Glenberg

Lehmann

. Spacing repetitions over 1 week. Mem Cognit. 1980;8:528-538.

27.

Braun

Rubin

. The spacing effect depends on an encoding deficit, retrieval, and time in working memory: evidence from once-presented words. Memory. 1998;6:37-65.

28.

Naqib

Sossin

Farah

. Molecular determinants of the spacing effect. Neural Plast. 2012; article ID 581291.

29.

Dudai

. The restless engram: consolidations never end. Ann Rev Neurosci. 2012;35:227-247.

30.

Saletin

Walker

. Nocturnal mnemonics: sleep and hippocampal memory processing. Front Neurol. 2012;3:1-12.

31.

Wang

Grone

Colas

Appelbaum

Mourrain

. Synaptic plasticity in sleep: learning, homeostasis and disease. Trends Neurosci. 2011;34:452-463.

32.

Bell

Kawadri

Simone

Wiseheart

. Long-term memory, sleep, and the spacing effect. Memory. 2014;22:276-283.

33.

Buonomano

Merzenich

. Cortical plasticity: from synapses to maps. Ann Rev Neurosci. 1998;21:149-186.

34.

D’Souza

Gil

Cassello

. IV glycine and oral D-cycloserine effects on plasma and CSF amino acids in healthy humans. Biol Psychiatry. 2000;47:450-462.

35.

Hood

Compton

Monahan

. D-cycloserine: a ligand for the N-methyl-D-aspartate coupled glycine receptor has partial agonist characteristics. Neurosci Lett. 1989;98:91-95.

36.

Onur

Schlaepfer

Kukolja

. The N-methyl-D aspartate receptor co-agonist d-cycloserine facilitates declarative learning and hippocampal activity in humans. Biol Psychiatry. 2010;67:1205-1211.

37.

Kuriyama

Honma

Koyama

Kim

. D-cycloserine facilitates procedural learning but not declarative learning in healthy humans: a randomized controlled trial of the effect of d-dycloserine and valproic acid on overnight properties in the performance of non-emotional memory tasks. Neurob Learn Mem. 2011;95:505-509.

38.

Jones

Wesnes

Kirby

. Effects of NMDA modulation in scopolamine dementia. Ann NY Acad Sci. 1991;640:241-244.

39.

Taub

Miller

Novack

. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil. 1993;74:347-354.

40.

Taub

Uswatte

Pidikiti

. Constraint-induced movement therapy: a new family of techniques with broad application to physical rehabilitation—a clinical review. J Rehabil Res Develop. 1999;36:237-251.

41.

Winstein

Miller

Blanton

. Methods for a multisite randomized trial to investigate the effect of constraint-induced movement therapy in improving upper extremity function among adults recovering from a cerebrovascular stroke. Neurorehabil Neural Repair. 2003;17:137-152.

42.

Wolf

Lecraw

Barton

Jann

. Forced use of hemiplegic upper extremities to reverse the effect of learned nonuse among chronic stroke and head-injured patients. Exp Neurol. 1989;104:125-132.

43.

Fugl-Meyer

Jaasko

Leyman

. The post-stroke hemiplegic patient: a method for evaluation of physical performance. Scand J Rehabil Med. 1975;7:13-31.

44.

Duncan

Wallace

Lai

Johnson

Embretson

Laster

. The Stroke Impact Scale version 2.0: evaluation of reliability, validity, and sensitivity to change. Stroke. 1999;30:2131-2140.

45.

Robinson

. Validation of a caregiver strain index. J Gerontol. 1983;38:344-348.

46.

Yesavage

Brink

Rose

. Development and validation of a geriatric depression screening scale. J Psychiatr Res. 1982-1983;17:37-49.

47.

Hedna

Jain

Rabbani

Nadeau

. Mechanisms of arm paresis in middle cerebral artery distribution stroke: a pilot study. J Rehabil Res Develop. 2013;50:1113-1122.

48.

Simone

Bell

Cepeda

. Diminished but not forgotten: effects of aging on magnitude of spacing effect benefits. J Gerontol Series B Psychol Soc Sci. 2013;68:674-680.

49.

Parnas

Weber

Richardson

. Effects of multiple exposures to d-cycloserine on extinction of conditioned fear in rats. Neurobiol Learn Mem. 2005;83:224-231.

50.

Quartermain

Mower

Rafferty

Herting

Lanthorn

. Acute but not chronic activation of the NMDA-coupled glycine receptor with d-cycloserine facilitates learning and retention. Eur J Pharmacol. 1994;257:7-12.

51.

McClung

Rothi

LJG

Nadeau

. Ambient experience in restitutive treatment of aphasia. Front Hum Neurosci. 2010;4:1-19.

A Pilot Randomized Controlled Trial of D-cycloserine and Distributed Practice as Adjuvants to Constraint-Induced Movement Therapy After Stroke

Abstract

Keywords

Distributed Practice

D-cycloserine

Methods

Participants

Randomization and Drug Treatment

Outcome Measures

Statistical Analysis

Results

Safety

Discussion

Potential Reasons for Treatment Failure

Trial Design

Conclusions

Footnotes

Acknowledgements

Authors’ Note

Declaration of Conflicting Interests

Funding

References