Article Text

Download PDFPDF

Research paper
Constraint-induced aphasia therapy following sub-acute stroke: a single-blind, randomised clinical trial of a modified therapy schedule
  1. Almut Sickert1,
  2. Lutz-Christian Anders2,
  3. Thomas F Münte3,
  4. Michael Sailer1,4
  1. 1Neurological Rehabilitation Centre, MEDIAN Kliniken, Magdeburg, Germany
  2. 2Department of Speech Science and Phonetics, Martin-Luther University, Halle-Wittenberg, Germany
  3. 3Department of Neurology, University of Lübeck, Lübeck, Germany
  4. 4Department of Restorative Neurology, Otto-von-Guericke-University, Magdeburg, Germany
  1. Correspondence to Dr Almut Sickert, Neurologisches Rehabilitationszentrum Magdeburg, Gustav-Ricker-Straße 4, Magdeburg 39120, Germany; almut.sickert{at}


Background and purpose The trend towards a shorter stay in rehabilitation clinic has implications for future language therapy. Constraint-induced aphasia therapy (CIAT) is administered 3 h per day for a total of 30 h of treatment. It was evaluated for patients with chronic aphasia. In the present study we investigated the efficacy of a modified CIAT schedule and included patients with sub-acute stroke. We conducted a randomised, single-blind, parallel-group study. The results were compared to those of patients who received identically intensive treatment focusing on conventional aphasia therapy.

Methods Fifty patients were treated with our modified version of CIAT and 50 received a standard aphasia therapy at the same intensity and duration. Inclusion criteria were clinical diagnosis of first-ever stroke, aphasia in the sub-acute stage and German speakers. Language function was evaluated using the Aachen Aphasia Test and the Communicative Activity Log directly before therapy onset, after the training period and at 8-week and 1-year follow-ups.

Results Patients of both groups improved significantly in all sub-tests of the Aachen Aphasia Test Battery. The improvements remained stable over a 1-year follow-up period. Patients and relatives of both groups rated daily communication as significantly improved after therapy.

Conclusions Both CIAT and conventional therapy performed with equal intensity are efficacious methods for patients with sub-acute aphasia. The modified CIAT schedule is practical in an everyday therapeutic setting. Our results indicate that a short-term intensive therapy schedule in the early aphasia stage leads to substantial improvements in language functions.

Clinical Trial Registration Information Clinical Trial Registration-URL: Unique identifier: NCT01625676

  • Aphasia
  • Speech Therapy
  • Speech
  • Randomised Trials
  • Rehabilitation

Statistics from


Aphasia is a language disorder that results from damage to areas of the brain that are responsible for language. Approximately 80% of language disorders arise as a consequence of stroke.1 Constraint-induced aphasia therapy (CIAT) is founded on the principles of the constraint-induced movement therapy (CIMT).2 The therapy includes an intensive, massed practice. Patients receive therapy each weekday for a 2-week period for up to 3 h per day.3 Complying with the principles of the phenomenon termed ‘learned non-use’ described by Taub et al,4 ,5 persons with aphasia often use non-verbal communication as a compensatory mechanism, for example, drawing or pantomiming to avoid spoken communication. To prevent or overcome learned non-use, the language exercises are realised in the form of a game. Communication between the patients has to be performed in spoken words or sentences. Compensatory mechanisms are constrained. Depending on the severity of the initial deficit, patients receive an intensive shaping therapy to improve their individual abilities to communicate.3 The effectiveness of CIAT in treating patients with chronic aphasia has been shown in a number of controlled clinical trials.3 ,69 The results indicate that massed practice paradigms are successful regardless of the method used.

In many facilities it is evidently difficult to administer CIAT in the original protocol described by Pulvermüller et al3 because of the intensity of practice and the limited number of patients in the groups. Additionally, patients with aphasia in the early sub-acute stage were often excluded from clinical trials up till now. The present study was designed to adapt the CIAT protocol to the patient's abilities in the early sub-acute stage. We considered the resources for executing CIAT in the rehabilitation centre and included patients with global aphasia. The aim of this study was to assess the feasibility of modified CIAT in the early sub-acute stage and to examine the efficacy of modified CIAT versus standard treatment with the same intensity of intervention.



One hundred patients (40 females/60 males), mean age 60.4±11.9 years, participated in this study (see table 1 for patient characteristics). Patients were recruited from the local rehabilitation centre (Neurologisches Rehabilitationszentrum Magdeburg, MEDIAN Kliniken). All patients suffered a first-ever stroke with aphasia in the sub-acute stage defined as time since lesion onset from 1 to 4 months post-stroke.1 The average duration of time since lesion was 34.8 days. In 85 patients aphasia was caused by an ischaemic stroke in the territory of the left middle cerebral artery and in 15 patients by a haemorrhagic stroke affecting left hemispheric areas. The following exclusion criteria were used: presence of a residual aphasia, dysarthria (scale values 0–3, dysarthria rating scale10) and apraxia of speech.

Table 1

Patient characteristics

Aphasia was diagnosed according to the guidelines of the standardised neurolinguistic German aphasia test battery, the Aachener Aphasia Test (AAT).11 All patients had to understand the rules of the game for the CIAT group. This requirement was assessed by a test game. If the patients satisfied the criteria of understanding the aim of the game, naming of items with therapeutic help and identifying one of four presented cards with object drawings, they were included in the study. All participants were native German speakers. Pre-morbid handedness was assessed with the Edinburgh Handedness Inventory.12 Patients pre-morbid educational level ranged from 6 to 12 years (M=9.2). We conducted a randomised, single-blind, parallel-group study. All trial participants were randomly assigned using a computer-generated randomisation code to either CIAT or standard treatment. There were no significant differences in participant characteristics between the two participant groups before participation. The experimental procedure was approved by the Ethics Committee of the University of Konstanz. Written informed consent was obtained from every patient or caregiver.

Language functions were evaluated by the AAT11 and the Communicative Activity Log (CAL). Tests were administered by speech therapists who were not involved in the study before training onset and after the 3-week training period. After the post-training assessment, patients were invited to an 8-week and a 1-year follow-up visit. The CAL is a questionnaire, which rates the quality and the amount of daily communication. It is designed by Pulvermüller et al3 to be analogous to the Motor Activity Log (MAL).13 We used a shortened version of CAL, including a patient self-evaluation and an evaluation by relatives of the amount and comprehension of everyday conversation.6 Patients and relatives rated independently of each other. All questions were answered on a 6-point scale (0=never; 5=same as pre-stroke).

A total of 100 patients underwent pre- and post-treatment testing. A group of 26 participants (CIAT: N=15; standard treatment group: N=11) were tested at both follow-up stages (8-week and 1-year follow-up visit). The aphasic syndromes were determined as: CIAT: amnesic (N=8), Wernicke (N=2), Broca (N=3), global (N=2); Standard treatment group: amnesic (N=4), Wernicke (N=4), Broca (N=2), global (N=1). Twenty-seven patients discontinued the therapy after our 3-week training period because they were satisfied with the outcome. The remaining patients did not attend the follow-up visit because of no aphasia/residual aphasia (N=17), severe illness (eg, reinfarction) (N=22), having moved (N=3) or mortality (N=5).


Each patient received 2 h of training over 15 days. In contrast to model-oriented aphasia therapy (MOAT) by Barthel et al,7 CIAT as well as the conventional aphasia therapy was performed as a group communication treatment. During the training period, participants did not receive any other language therapy, but participated in the standard occupational and physical treatment programme. Patients were randomly assigned to groups independent of their aphasia characteristics. A person who did not have patient contact used a computer-generated random number. Every group consisted of 4–6 patients and a speech therapist. Furthermore, two patients without aphasia participated in the CIAT group setting. These patients were recruited from the medical professional rehabilitation team. In contrast to Meinzer et al6 and Barthel et al,7 the home practice programme was not used in the CIAT setting, but patients and relatives of both groups received professional advice.

The conventional therapy was guided by a standard aphasia treatment focusing on training specific deficits. Contents of the conventional therapy were, for example, exercises including sentence completion, improving patients’ retrieval of words, learning sentence patterns, conversation on current topics, listening to words, and repeating and following instructions. The therapist initiated the communicative activities. The interventions aimed to target several modes of communication. In contrast to CIAT, the participants were permitted to use any communication mode, including non-verbal communication.

CIAT is based on therapeutic language games. The key principles of the treatment were massed practice, shaping and constraint of non-verbal strategies.3 The treatment stimuli used for the game consisted of cards with object drawings, photographs of everyday situations and a module of written language, established by Meinzer et al.6 Additional writing in the sense of computer-generated phonemic cues was added.14 If the patient was able to name the picture after using the self-cue, the written cues were eliminated. Participants who were able to observe all rules and obtained a higher-than-average achievement were asked to write down dictated words during the game. In this way, we included the ability to write for the first time.

The therapist presided over the treatment setting but did not participate as a co-player. The rules of communication were formulated, individualised for each participant and were gradually increased. The therapist provided as much cueing as necessary, depending on the level of each participant's verbal ability, for a successful response. Following the procedure by Pulvermüller et al,3 we placed a visual barrier between the participants so they could not see each other's hands or cards.

The patients from the medical professional rehabilitation had to consider all rules of the game. When the difficulty level of the exercises was increased, these patients took the first turn and, if the communication was unsuccessful, they gave the players feedback. In contrast to Pulvermüller et al,3 Meinzer et al6 and Maher et al,8 the patients from the medical professional rehabilitation were not appointed to be co-therapists but rather to be co-players with an exemplary function.

After discharge from the programme, participants received some outpatient treatment at a comparable intensity between groups (CIAT: 1.9 h/week on average; standard group: 2.13 h/week on average). This treatment was controlled for (1) professional speech therapists, (2) method based on a standard approach widely used in Germany and (3) time and intensity.

Data analyses

The statistical analysis was carried out using the concept of descriptive data analysis.15 To compare the results of the AAT we used t-transformed raw values.11 Differences between clinical and demographical parameters were analysed by unpaired t-tests. The Mann–Whitney test was used for between-group comparisons of the AAT-spontaneous speech scale. Individual improvements in the AAT were assessed according to the guidelines11 on the basis of critical differences. Between-group comparisons with regard to changes of language functions in the sub-tests of the AAT and the CAL were analysed by repeated measures analysis of variance. For within-group comparisons, paired t-tests were used. The stability of language output across the follow-up measurements was analysed by Friedman's test and Dunn's multiple comparison test. We used the χ2-test for differences between intervention groups in categorical variables. To analyse the associations between demographical parameters and improvement in the AAT, we allocated all 100 participants independent of the training group to a sample (age: ≤60 and >60 years; educational level: ≤8 and ≥10 years; sex: females and males, aphasia syndrome). Associations between changes in language function and aphasia syndrome were evaluated for patients with global aphasia, both by treatment group and independently.


A total of 100 patients received treatment and were analysed after the 3-week training period. Thirty-six patients of the CIAT treatment group and 37 of the standard group members showed significant improvements on at least one sub-test of the AAT immediately after the training. Both groups, independent of the treatment schedule, showed immediately after the training significant improvements in all sub-tests of the AAT (CIAT: t [49]>5, p<0.001; standard treatment group: t [49]>3, p<0.001; see table 2 for speech and language characteristics). We tried to estimate the magnitude of performance increase in all sub-tests of the AAT on the basis of percentile values. Within each group, language skills improved from pre- to post-treatment. Spontaneous speech improved by a mean of 15%, repetition 10%, Token Test 27%, written language 30%, naming 32% and comprehension 12% in patients receiving constraint-induced therapy. Standard treatment patients’ performance increased by a mean of 15% in spontaneous speech, 11% in repetition, 32% in the Token Test, 37% in written language, 31% in naming and 15% in comprehension. No significant differences were observed across treatment groups (Token Test: F [1, 99]=0.02; p>0.05; repetition: F [1, 99]=0.019; p>0.05; written language: F [1, 99]=0.593; p>0.05; naming: F [1, 99]=0.160; p>0.05; comprehension: F [1, 99]=0.034; p>0.05).

Table 2

Speech and language characteristics pre- and post-therapy

The CAL consisted of a version for patient self-evaluation and one for evaluation by relatives. Each version included 11 items that addressed the amount of everyday conversation and 10 items rating the language comprehension. The presence of a severe aphasia and the absence of some caregivers was a limiting factor in the completion of the questionnaire. We could analyse 44 self-evaluations and 45 ratings by relatives before and after treatment in the CIAT group. In the standard treatment group we could analyse 41 self-evaluations and 38 ratings by relatives. The patients and relatives of both groups rated, independent of each other, comprehension and amount of communication. The patients reported a significant increase in the amount of communication and in language comprehension in everyday life after treatment (amount of communication: CIAT: t [43]>4, p=0.0001; standard treatment group: t [40]>2, p=0.047; language comprehension: CIAT: t [43]>2, p=0.0063; standard treatment group: t [40]>2, p=0.028). Significant improvement was confirmed by the relatives in both groups (amount of life situation communication: CIAT: t [44]>3, p=0.001; standard treatment group: t [38]>2, p=0.013; language comprehension: CIAT: t [44]>2, p=0.005; standard treatment group: t [37]>3, p=0.0009). There were no significant differences concerning the CAL results between the intervention groups (F<1) (table 2). The ratings of patients and relatives were comparable in the within-group analysis (CIAT: all t<1.04, p>0.05; standard treatment group: all t<0.65, p>0.05).

A group of 26 patients were tested at an 8-week and a 1-year follow-up visit. Compared with the post-assessment, patients of the CIAT group showed no further significant improvement in the Token Test. Results remained stable at a significant level (p<0.001). Participants trained by the standard treatment schedule showed a further significant improvement above immediate post-treatment values (8 week vs 1 year p<0.01). In the repetition, written language and naming sub-tests, we observed a commensurable advancement in both groups. The improvements remained highly significant over the 1-year follow-up period (p<0.001). A comparison with post-treatment scores demonstrated no further change in language function (p>0.05). We observed an additional significant increase in the language comprehension sub-test between post-assessment and the last follow-up visit in the CIAT group (p<0.01). The results of the AAT-spontaneous speech scale remained at the same high level in both intervention groups (p<0.001). There were no significant differences in the follow-up assessment values between both intervention groups (see figure 1A,B for language test results at both follow-up periods).

Figure 1

(A and B) Results (mean scores) of the constraint-induced aphasia therapy and the standard treatment group at both follow-up periods (follow-up 1=after 8 weeks; follow-up 2=after 1 year). The results of the Aachener Aphasia Test remained stable in both groups.

Neither intervention group differed significantly on any test value before and after therapy. We also sought to compare the therapy outcome of patients with different aphasia syndromes. Therefore in an additional analysis the participants were grouped depending on aphasia syndrome regardless of treatment schedule. Importantly, the distribution of aphasia syndromes was not different in both groups before treatment (p>0.4). The results documented that there were no significant differences in the outcome of the patients in all sub-tests of the AAT (p>0.05). Both treatment schedules were equally effective for patients regardless of aphasia syndrome.

Age and sex had no important influence on the degree of improvements in language skill measured by the AAT (p>0.05). Both treatment schedules were suitable for all participants regardless of their educational level. A higher educational level (≥10 years) had a positive impact on the results of written language and spontaneous speech analysis of the AAT.


We used two short intensive treatment approaches, a constraint-induced schedule (CIAT) and a standard group schedule, for the treatment of patients with aphasia. We found that both therapies lead to significant improvements in language function. Importantly the patient's abilities improved to the same extent for the two therapies.

Complying with the principles of CIAT as described by Pulvermüller et al,3 we additionally included writing as an aid to word retrieval into the therapy. Although our standard treatment includes more writing exercises, no statistically significant differences were evident when comparing the standard therapy with our modified CIAT. The improvement in written language was more pronounced in those patients who received standard therapy. The CIAT group in the study by Pulvermüller3 did not practice written language. Meinzer et al6 included a module of written language, while our modified CIAT therapy involved writing to a greater extent. Here the focus was on spoken communication. In the Model MOAT by Barthel et al,7 writing exercises were most frequently practised, because this programme was designed as an individual therapy. This explains the significant differences concerning the better AAT results in the written language sub-test for MOAT versus CIAT (in the original protocol), and the tendency towards a significant difference between MOAT and CIAT (modified by Meinzer et al6). In our study we found only slight but not statistically significant differences between CIAT and the standard treatment group.

The improvements in language functions were uncorrelated with age, sex and educational level. In contrast to the original protocol described by Pulvermüller et al,3 we also included patients with global aphasia. We checked compliance every session and asked patients about difficulties and their frustration level. Importantly, no patient, whether in the CIAT or in the standard treatment group, felt overwhelmed or stopped the therapy. Both treatment schedules were equally well received by the patients. Interestingly, the participants showed a substantial improvement in language skills, irrespective of the initial aphasia syndrome. The response to the therapy did not vary with the type of aphasia syndrome.

The presence of a severe impairment in language comprehension was a limiting factor for the completion of the CAL. It is particularly problematic to use a questionnaire in a group of aphasic patients, because there is no possibility to make sure whether all participants precisely understood the items. Furthermore, patients in the early sub-acute stage often do not realise their disturbance in its entirety. Caregivers indicated that the separation of patients and relatives during the week made it difficult to complete some items. Different demands on communication are a general problem when using questionnaires.

After intervention we observed significant improvements in the CAL scores in both groups. However, only those patients who trained with CIAT noticed a more pronounced increase in the amount of communication. These results are comparable to those reported in previous clinical trials.6 ,7

In contrast to Meinzer et al and Barthel et al, we did not include home practice or the assistance of family members in the therapies. Patients and relatives of both groups received professional advice on request. Barthel et al7 concluded that an intervention which involves the assistance of family members, or additional training in everyday communication, is more effective. Some caregivers of the MOAT group were incapable of participating in partner-oriented training in communication skills. Even so, patients achieved significant improvements. Therefore the additional training in everyday communication by family members or caregivers cannot be the crucial factor for the improvement in CAL results. Indeed, CIAT cannot be more effective in facilitating the transfer of therapeutic gains from exercising to everyday communication, because there were no statistically significant differences between the intervention groups. Thus a short-term intensive therapy schedule seems to be therapeutically effective. However, Bowen et al16 have demonstrated that aphasia therapy with lower therapy intensity failed to achieve comparable improvements.

Patients of both groups received a commensurable quantity of therapy after intervention. Consequently, there were no significant differences in the AAT results between the intervention groups after the follow-up period. The previous effect remained stable. Pulvermüller17 recommended have a group of patients with similar problems of communication. This was the case in their CIAT studies. The present study, by contrast, was designed in that patients were grouped independently of their aphasic characteristics. Thus no patient had to wait for a suitable group. In comparison to homogeneous groups we observed that the participants were able to motivate each other much better, probably because patients learned from each other and they were in contact with each other during the whole hospital stay. Springer18 described the benefits of heterogeneous groups too, if the focus was on spoken communication. It is well conceivable that patients with mild aphasia could have an exemplary function for severely impaired patients. Even patients with mild aphasia can benefit from heterogeneous groups. Depending on individual skills, they might be able to use more semantically strategies to explain co-players unknown items and to give them feedback if the response is not successful. Importantly, the CIAT schedule is individually adjusted to the possibilities of each patient and especially suitable for heterogeneous groups.

The effectiveness of CIAT in treating patients with chronic aphasia has been shown in a number of controlled clinical trials. In the current study we demonstrated the feasibility of the method in the early sub-acute stage patients for the first time. We determined the efficacy of two therapies under real conditions in a rehabilitation centre. Nevertheless, we cannot completely rule out the influence of spontaneous remission because of the early sub-acute stage of the patients. For ethical reasons we did not have a no treatment group.

Though we did not find an advantage for CIAT it offered many possibilities to stimulate language effectively in a playful way. Like the conventional treatment with equal intensity it was very well tolerated by patients. Therefore, our results suggest the implementation of a short-term intensive aphasia therapy schedule during the early sub-acute rehabilitation stay.


Both treatment schedules improved language skills significantly. The current results suggest that CIAT and conventional therapy performed with equal intensity are effective methods for the treatment of patients with sub-acute aphasia. Both patients and relatives rated everyday communication as significantly increased. As a therapy with a massed treatment schedule the modified CIAT is feasible under real conditions in a rehabilitation centre. This therapy was well tolerated in the sub-acute stage, regardless of aphasia syndrome. In contrast to the original protocol described by Pulvermüller et al3 that required a high amount of time and personnel resources, our modified CIAT is less intensive with regard to these costly factors.


We are grateful to the team of speech and language therapists, especially to Susanna Pohl-Jakobs, for their support of this project.



  • Contributors AS, L-CA, TFM and MS have substantially contributed to the manuscript. They conceived and designed the study. They participated in the acquisition and interpretation of data as well as in drafting and revising the article. All authors have approved the final version of the manuscript.

  • Competing interests None.

  • Ethics approval University of Konstanz.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.