The current standard of care for rehabilitation of patients with aphasia after stroke is conventional speech and language therapy (SLT). Due to economic realities on most stroke units, SLT can often not be given with optimal intensity in the first weeks after the stroke. Developing new adjuvant therapies which may render SLT sessions more effective is thus one approach to improve rehabilitation outcome. Recent functional imaging studies in post-stroke aphasia have shown that the recruitment of brain regions in the unaffected hemisphere seems to be an inferior strategy for recovery of language function as compared to re-activation of brain regions in the vicinity of the infarct. Non-invasive brain stimulation techniques, such as repetitive transcranial magnetic stimulation (rTMS) or transcranial direct current stimulation (tDCS) are new methods to modulate brain activity. Evidence from our own feasibility study in sub-acute stroke suggests that these new techniques, when applied in conjunction with conventional SLT, may help to normalize brain activation patterns and might yield better rehabilitation outcome than SLT alone. With NORTHSTAR, we propose a multicenter proof-of-concept study to investigate the safety, feasibility and efficacy of these new non-invasive brain stimulation methods as adjuvant therapies for subacute post-stroke aphasia.

Our goal is to determine the most effective brain stimulation modality to decrease inhibition onto the left side of the brain. We will assess if a combination of brain stimulation and speech and language therapy will improve language recovery. We will quantify language recovery (expressive and comprehensive skills) using specific tests, commonly used by speech and language therapists.

We will invite patients recently admitted to the stroke unit of the study centers to participate in our research project. Once patients consent to our study we will randomly assign them to one of three experimental groups. For 12 days, all groups of patients will be setup with brain stimulation during their usual rehabilitation sessions. Two of those groups (treatment groups) will each receive a different type of brain stimulation (rTMS and tDCS), in the third group, patients will not receive real stimulation (placebo group). By comparing the extent of aphasia recovery between groups, we will determine the benefits attributable to brain stimulation relative to SLT alone.

Aim 1: To test the hypothesis that any type of non-invasive brain stimulation in combination with conventional SLT results in better aphasia recovery at 1 and 30 days after end of treatment than SLT with sham stimulation.

Aim 2: To test the hypothesis that tDCS in combination with SLT is non-inferior to rTMS in combination with SLT but both are superior to SLT with sham stimulation.

Aim 3: To test the hypothesis that there will be no difference in the frequency of adverse events between the treatment groups.

Aim 4: To test the hypothesis that there is a differential influence of attention, time to treatment, infarct location and bilingualism on the treatment effect.

Aim 5: To test the hypothesis that the extent of functional and structural connectivity between the stimulation site over the triangular part of the right inferior frontal gyrus and the left hemisphere primary language centers determines effectiveness of the therapy.

Description of study population

Additional information: The eligibility criteria have been selected such that the resulting study population will reflect a typical clinical population of aphasic stroke patients. Stroke location and aphasia type are thus not inclusion or exclusion criterion but will be documented in order to ensure a similar distribution between treatment groups and allow for subgroup analysis.

Study design

Additional information: All patients will receive 45 minutes of model-oriented individualized aphasia therapy administered by a certified therapist and focused on the individual specific linguistic problems as determined by the therapeutic team according to SLT best-practice guidelines.

Stimulation procedure

Transcranial magnetic stimulation (TMS): Using a figure-of-8 shaped electromagnetic coil, a rapidly changing (1Hz) magnetic field is applied to the scull over the target stimulation site (Broca Area's homologue in the non-dominant hemisphere, non-dominant BA45). At a frequency of 1Hz, these currents have been shown to interfere with normal neuronal activity by down regulating it. The suppressive effect of 20 minutes of 1Hz rTMS with a stimulation intensity that would elicit muscle contractions (motor evoked potentials, MEP) when applied over the motor cortex (resting motor threshold, RMT) lasts about 45 minutes. SLT sessions will be given immediately following the stimulation procedure. For sham-stimulation, the coil will be placed over the inter-hemispheric fissure at the vertex and stimulation will be performed with low intensity (10% RMT). This will cause similar skin sensations as real stimulation but will not induce currents in language relevant areas.

Transcranial direct current stimulation (tDCS): For tDCS a 5 cm2 sponge electrode (cathode) will be placed over the target area (Broca Area's homologue in the non-dominant hemisphere) and the anode on the forehead over the contralateral eye. A direct current of 2mA will be applied between electrodes. A negative current applied to the brain area under the cathode will cause a decrease in resting membrane potential of the underlying neurons thus reducing their excitability. tDCS will start immediately before the SLT session and last throughout the session. tDCS only evokes a tingling sensation on the skin when turned on and off. For sham stimulation the current will be turned on for 30 seconds to elicit a typical skin sensation and then turned off for the duration of the therapy. The same will be done at the end of the session.

Localization of stimulation sites: Stimulation sites will be localized using a modification of the surface distance measurements (SDM) method. Study centres will send the patient's T1-weighted MRI to the study coordinating centre. Surface renderings of the MRI will be generating using the 3D-Tool. Reference measures relative to the nasion-inion line will be identified on these reconstructions and sent to the participating centre. The person preparing the stimulation can verify the distances on the real patient's head and determine the respective stimulation site. The method has an accuracy of approx. 8 mm for localizing Broca's Area when compared to neuronavigation. Having designed the study with broad and cost-effective applicability in mind, we chose this approach although the centres do have advanced neuronavigation available.

Determination of resting motor threshold (RMT): The RMT will be defined iteratively as previously described over the right M1 prior to each treatment session. Within M1 the position which elicits the highest MEP-Amplitude in the left first dorsal interosseus muscle (FDI) with 80% or higher stimulator output will be maintained and stimulation intensity will be reduced first in 5% intervals until less than 5 consecutive MEPs with amplitude greater 50µV in 10 stimulation trials are obtained. Stimulation intensity will then be set up in 5% intervals until 5 consecutive MEPs with amplitude greater 50µV in 10 stimulation trials are obtained. The intensity at this point is used as RMT.

MR imaging procedures: MRI will be performed on a 3T SIEMENS MAGNETOM TrioTim syngo MR B17. We will acquire volumetric data-sets with 1mm resolution comprising one T1-weighted image for determination of stimulation site (ADNI-MPRAGE, 10 minutes, 192 sagittal slices, 1mm thickness, TR 2300 ms, TE 2.98ms), fluid attenuated inversion recovery sequences (axial FLAIR, 3.5 minutes, 60 slices, 2mm thickness, flip angle 150 deg, TR 900ms, TE 70ms) for infarct localization and quantification of infarct volume followed by two 10 minute resting state fMRI sequences (BOLD MOSAIC 64, 10 minutes, 34 slices, 4mm thickness, 90deg flip angle, TR 1810ms, TE 30ms) and a 10 minute diffusion sequence (63 slices, 2mm thickness, TR 8400ms, TE 90ms, 64 directions)

Measurements and study instruments

The present study will use three tests of elementary language function as primary outcome variables (verbal fluency, comprehension and picture naming) for which comparable age-matched norms for all languages (English, French and German) of the participating countries exist. Language specific test batteries as well as the Bilingual Aphasia Test (BAT) will be used as secondary outcome measures and for classification of aphasia types.

In the absence of a single test for aphasic impairment validated in all three languages, measures of aphasic impairment will only be considered as secondary outcome measures. Each participant will be assessed with a test battery that is commonly used by the majority of speech and language therapists in the corresponding languages: the Aachen Aphasia Test (AAT) in German; the Protocole Montréal-Toulouse 86 (MT) in French; and the Western Aphasia Battery (WAB) in English. The tests will be repeated 1 and 30 days after the last treatment session. In addition the Bilingual Aphasia Test (BAT) will be used, in order to validate its sensitivity to detected changes in language performance during aphasia recovery for future studies.

In addition, the following tests will be performed:

Initial Testing (baseline) - before treatment: Demographics (including education) and medical history questionnaire, Neurological exam, cognitive testing (MoCA), NIHSS, CETI, Barthel Index, modified Rankin scale After each treatment session: Assessment of adverse events (AE) and serious adverse events (SAE) 1 and 30 days after last treatment session: Neurological exam, cognitive testing (MoCA), NIHSS, Communicative Effectiveness Index (CETI), Barthel Index, modified Rankin scale.

Data analysis plan

Primary hypothesis:

1 - I Patients treated with any type of non-invasive brain stimulation will show a significantly larger increase in picture naming, token-test or semantic verbal fluency scores than any sham treated patients at 1 day and 30 days after last therapy session

1 - II tDCS will not be inferior to rTMS but superior to sham stimulation with respect to treatment effects at 1 day and 30 days after last therapy session

1. - III There will be no difference in the frequency of AE/SAE between treatment groups

   Secondary hypothesis:

2. - I Both treatment groups will show a larger increase in the global test scores than sham-groups at 1 day and 30 days after last treatment.

2 - II Patients with lesions in the anterior MCA-territory (frontal and basal ganglia) will show greater treatment effects in verbal fluency than patients with lesions in the posterior MCA territory at 1 day after last session.

2 - III Patients with lesions in the posterior MCA territory (temporo-parietal) will show greater treatment effects in Token test than patients with lesions in the anterior MCA territory at 1 day after last session.

2 - IV There will be differential influence of attention, time to treatment, infarct location and bilingualism on the improvement in picture naming, token-test and semantic verbal fluency in the three treatment groups.

2 - V Time series of BOLD-signal fluctuations measured in the triangular part of the right inferior frontal gyrus will be 2-VI fractional anisotropy values in transcallosal fiber tracts between primary language areas will be correlated with improvement in clinical outcome measures.

In order to test these hypotheses, the following primary outcome variables will be derived from the primary outcome measures: difference between semantic verbal fluency test scores at baseline and and 1 and 30 days after completion of the treatment period (dSF1, dSF30), difference between picture naming test scores at baseline and 1 and 30 days after completion of the treatment period (dPN1, dPN30), difference between Token test scores at baseline and 1 and 30 days after completion of the treatment period (dTT1, dTT30) and a cumulative number of AE and SAE during 10 days of therapy (AE1) and during the 30 days follow-up period (AE30).

Secondary outcome variables will be: Percent difference of global aphasia test scores at follow-ups (1 and 30 days after completion of the treatment period) relative to baseline (dAT1, dAT30). The stratification variables will be TRT: treatment (levels: any stimulation, sham), STT: stimulation type (levels: rTMS, tDCS, sham) and LOC: infarct location (levels: MCA-ant, MCA-post).

Planned statistical tests for primary and secondary outcome variables are ANOVA for between group effects with the respective stratification variables as factors as required by each of the hypotheses. A significance level of P<0.01 corrected for multiple comparisons will be accepted.

Confounding variables: sex, age, intelligence, socioeconomic status and aphasia type have been shown to correlate with initial aphasia severity but NOT with recovery. We will thus control for these variables by randomization. Multivariate analysis: We will use multiple regression to test whether time of therapy onset after stroke, attention, bilingualism and infarct location are differentially correlated with recovery in the treatment groups.

Resting state fMRI data will be analyzed following a previously published protocol for functional connectivity in language networks. Time series of BOLD-signal fluctuations within the right inferior frontal gyrus (pars triangularis, seed region) will be correlated with signal fluctuations in the non-infarcted gray matter of the left hemisphere on voxel by voxel basis using the MATLAB toolbox for functional connectivity implemented in SPM. We expect interhemispheric connectivity of the right inferior frontal gyrus with left hemisphere language cortex to be higher in those patients who will show the greatest response to therapy.

Diffusion Tensor Imaging data will be analyzed following our previously established protocol. Transcallosal fibers will be traced between homotopic primary language regions in both hemispheres and fractional anisotropy will be measured within those traced fibre bundles as surrogate marker of fibre tract integrity as previously described. FA values will then be correlated with clinical primary outcome measures in each treatment group to test the hypothesis that fibre tract integrity predicts therapeutic response.

Inclusion Criteria:

• ischemic stroke in the left MCA territory
• between 5 and 30 days post stroke
• right-handedness
• English, French or German as language of daily use
• score below the lower limit of the norm on at least one of the primary outcome measures

Exclusion Criteria:

• prior symptomatic ischemic or hemorrhagic stroke
• severe comprehension deficit that may compromise informed consent or understanding of instructions
• contraindications to MRI and/or TMS/tDCS
• neurodegenerative or psychiatric disease
• epilepsy or EEG-documented epileptic discharges
• chronic renal or liver failure
• life-threatening diseases
• auditory or visual deficits that cannot be corrected and might impair testing