29 June 2021: Clinical Research
Intermittent Theta Burst Stimulation (iTBS) for Treatment of Chronic Post-Stroke Aphasia: Results of a Pilot Randomized, Double-Blind, Sham-Controlled TrialJerzy P. Szaflarski 12ABDEFG* , Rodolphe Nenert 1BCDE , Jane B. Allendorfer 13BE , Amber N. Martin 1BE , Amy W. Amara 13BE , Joseph C. Griffis 14BE* , Aimee Dietz 5AE , Victor W. Mark 146DE , Victor W. Sung 1BE , Harrison C. Walker 16BE , Xiaohua Zhou 6BE , Christopher J. Lindsell 7ACE
Med Sci Monit 2021; 27:e931468
BACKGROUND: Research indicates intermittent theta burst stimulation (iTBS) is a potential treatment of post-stroke aphasia.
MATERIAL AND METHODS: In this double-blind, sham-controlled trial (NCT 01512264) participants were randomized to receive 3 weeks of sham (G₀), 1 week of iTBS/2 weeks of sham (G₁), 2 weeks of iTBS/1 week of sham (G₂), or 3 weeks of iTBS (G₃). FMRI localized residual language function in the left hemisphere; iTBS was applied to the maximum fMRI activation in the residual language cortex in the left frontal lobe. FMRI and aphasia testing were conducted pre-treatment, at ≤1 week after completing treatment, and at 3 months follow-up.
RESULTS: 27/36 participants completed the trial. We compared G0 to each of the individual treatment group and to all iTBS treatment groups combined (G₁₋₃). In individual groups, participants gained (of moderate or large effect sizes; some significant at P<0.05) on the Boston Naming Test (BNT), the Semantic Fluency Test (SFT), and the Aphasia Quotient of the Western Aphasia Battery-Revised (WAB-R AQ). In G₁₋₃, BNT, and SFT improved immediately after treatment, while the WAB-R AQ improved at 3 months. Compared to G₀, the other groups showed greater fMRI activation in both hemispheres and non-significant increases in language lateralization to the left hemisphere. Changes in IFG connectivity were noted with iTBS, showing differences between time-points, with some of them correlating with the behavioral measures.
CONCLUSIONS: The results of this pilot trial support the hypothesis that iTBS applied to the ipsilesional hemisphere can improve aphasia and result in cortical plasticity.
Keywords: Aphasia, Language, Magnetic Resonance Imaging, Rehabilitation, Transcranial Magnetic Stimulation
Approximately 30% of the ~800 000 people who suffer from ischemic stroke each year present with aphasia . Stroke with aphasia is more costly than stroke without aphasia, with an additional annual cost in excess of $2B . The time course and the degree of post-stroke aphasia recovery have been examined in several short- and intermediate-term follow-up studies . Although the subacute post-stroke period is typically associated with spontaneous recovery, minimal or no spontaneous recovery of aphasia is expected beyond the first 6–12 months after presentation [4,5]. Occasionally, aphasia may recover several years after stroke [6,7]. In a recent individual participant data meta-analysis, several factors have been shown to significantly affect outcomes of post-stroke aphasia rehabilitation, including younger age at stroke occurrence and better baseline performance on Boston Naming Test (BNT), Western Aphasia Battery Aphasia Quotient (WAB AQ), and Aachen Aphasia Test (AAT) Spontaneous-Speech Communication subscale . The presence of persistent aphasia is frequently noted as a primary cause of post-stroke social isolation, struggles with mood and depression, and perceived or real cognitive impairments, all of which significantly reduce quality of life and resumption of pre-stroke life activities .
Recently, non-invasive brain stimulation has received interest as an intervention for improving chronic post-stroke aphasia and for priming the brain prior to behavioral interventions [10–12]. More specifically, neurostimulation is thought to facilitate neuroplasticity through either changing the perilesional canonical language networks or recruiting compensatory networks [12,13]. In other words, the post-stroke recovery takes advantage of either other networks adapting to perform damaged functions or networks that were dormant prior to stroke activating in the face of increased difficulty in performing a task because of the stroke-related damage . In the realm of neurostimulation, excitatory repetitive transcranial magnetic stimulation (rTMS), which exerts its effects via decreased local GABA-ergic inhibition and increased direct long-term potentiation , has been shown to enable language processing in health and disease, including post-stroke aphasia . The use of fMRI to guide selecting the site of stimulation is another advance for administering rTMS. However, another approach is to apply rTMS to the unaffected (right) hemispheric language homologue (see  for an extensive review). In a neurostimulation study, stimulating the affected (20Hz) and unaffected (1Hz) hemispheres sequentially followed by rehabilitation for 10 days resulted in an improvement on several measures of aphasia when compared to sham . Similarly, substantial linguistic gains followed applying 1Hz rTMS to the area of maximum fMRI activation in either the left or right hemisphere prior to a 10-day behavioral intervention in participants with fluent and non-fluent aphasia .
Despite the therapeutic potential, many participants report adverse effects from rTMS, including headache, muscle twitches, and residual local hypersensitivity [12,13]. A recent development in the realm of neurostimulation is intermittent theta burst stimulation (iTBS) [16–18]. This typically more comfortable and better tolerated form of neurostimulation mimics the electrical firing of the hippocampus that underlies long-term potentiation [17,19]. Based on its mechanism of action, iTBS has high relevance to learning and memory by affecting synaptic plasticity via producing consistent, long-lasting, and powerful effects on behavior and physiology after an application period of 20 to 190 s . However, it is well recognized that excessive or prolonged stimulation can have the reverse effect . Similar to conventional rTMS, iTBS induces changes in the underlying cortex for ~60 min that are associated with benefits such as motor improvement [20,21].
Recently, we have shown that 10 sessions of fMRI-guided iTBS applied to the peristroke region showing maximum language fMRI activation, typically at or near the inferior frontal gyrus (IFG), changed fMRI activation patterns and was associated with both improved communication skills and a trend towards improved aphasia testing (AT); the participants did not report any adverse events . Additionally, fMRI-guided iTBS applied to the peristroke language activation area followed by constraint-induced aphasia therapy (CIAT) resulted in significant fMRI changes and communication gains . ITBS has been also shown to change brain structure and function when applied to the perilesional language area [18,22–24]. Together, these studies support the hypothesis that iTBS applied to the peristroke regions is safe, well-tolerated, not associated with severe adverse events (eg, seizures), and may be associated with linguistic gains through its facilitation of the perilesional canonical language networks; these effects may be independent of or occur in conjunction with cognitive rehabilitation [18,25,26]. Clinical trial data are needed to promote translation of these discoveries into practice.
The aim of the present study was to demonstrate the feasibility of conducting a pilot randomized, double-blind, sham-controlled trial comparing the effects of sole fMRI-guided iTBS to sole sham iTBS on chronic aphasia in patients with a single left (dominant) hemispheric ischemic stroke, and to determine whether any observed initial treatment response would be sustained. We specifically excluded providing cognitive intervention in this trial (ie, using iTBS as a primer for cognitive intervention); a separate open-label study was recently completed by our group to evaluate the potential efficacy of combination of iTBS and constraint-induced aphasia therapy [18,25]. We also aimed to assess the relationship between the duration of iTBS treatment (1 vs 2 vs 3 weeks) and the linguistic and fMRI outcomes. We hypothesized that iTBS alone, when compared to sham alone treatment, would improve linguistic performance in patients with aphasia in a dose-dependent fashion.
Material and Methods
PARTICIPANTS: Clinicians identified and referred eligible participants from their neurology and rehabilitation clinics. Eligible participants were at least 1 year after a single left, middle cerebral artery (LMCA) ischemic stroke with documented persistent aphasia and had not received speech-language therapy within the 3 months preceding study enrollment. Medical history was confirmed by records review including admission notes and imaging of the brain (CT or MRI). We used the most recent available brain imaging results to confirm the diagnosis and to exclude participants with more than 1 stroke. Potential participants underwent a screening Token Test (TT) and were qualified for study participation if the results showed at least mild aphasia (TT ≤40) . Patients were excluded if they had a history of a neurodegenerative (eg, dementia), metabolic (eg, encephalopathy), or supervening medical disorder (eg, brain tumor or other cancer), history of severe depression or other mental illness, contraindication to 3T MRI, or positive pregnancy test on the day of MRI scanning in women of childbearing age. The Institutional Review Boards of the participating institutions approved the study, and the trial was registered at clinicaltrials.gov (NCT 01512264). All participants (or their legal representatives if the participants were judged clinically to have impaired speech comprehension) signed the informed consent prior to initiating any study procedures.
STUDY DESIGN: Participants were randomized into 1 of 4 therapy groups (Figure 1): Group 0 (G0) received 3 weeks of sham iTBS, Group 1 (G1) received 2 weeks of sham and 1 week of iTBS, Group 2 (G2) received 1 week of sham and 2 weeks of iTBS, and Group 3 (G3) received 3 weeks of iTBS. Randomization envelopes were prepared prior to initiating the study by the study statistician (CJL); envelopes containing group assignment were opened sequentially. All study staff with the exception of the iTBS treatment staff were blinded to group allocation; however, they were blinded to the results of the pre-enrollment (t1) AT. Blinding was maintained until all participants completed the study. Each participant received fMRI and AT within 1 week prior to initiating the intervention (Figure 1; t1), within 1 week following the intervention (t4), and 3 months later (t5). AT was also performed at the end of each treatment week (t2 and t3). Since the optimal number of therapy sessions needed for the improved language outcomes is unknown, treatment dosing (1, 2, or 3 weeks) was also implemented in the design of the trial.
ASSESSMENTS: Due to the pilot nature of the study, we did not select a specific primary outcome measure. Rather, we used a suite of AT to broadly gauge language comprehension and production, and to explore which measures may be most sensitive to iTBS. The Aphasia Quotient (AQ) of the Western Aphasia Battery-Revised (WAB-R)  and fMRI were administered at baseline (t1), after intervention (t4), and at 3 months (t5). The Boston Naming Test (BNT) , Semantic Fluency Test (SFT) , and Controlled Oral Word Association Test (COWAT)  were administered at baseline (t1), after each week of treatment (t2–4), and at 3 months (t5). For each group, paired samples t tests were performed to examine changes between time-points, and effect sizes (Cohen’s d) were computed for each.
TRANSCRANIAL MAGNETIC STIMULATION: ITBS sessions were given for 5 consecutive weekdays over 3 weeks, resulting in 15 treatment sessions. Each session involved either stimulation or sham, depending on group assignment. Prior to the first treatment session [22,23], we established the resting (RMT) and active motor (AMT) thresholds by applying a single-pulse TMS to the right hemisphere motor cortex (MRI-guided localization) using a Magstim Rapid2® figure-of-eight coil (Magstim Co., Wales, UK) with EMG leads placed over the first dorsal interosseous (FDI) muscle of the left hand. The stimulation coil was placed tangentially to the skull with the handle parallel to the sagittal axis and over the primary motor cortex in the right (unaffected) hemisphere at the optimal site for obtaining a motor evoked potential in the FDI muscle. After the RMT and AMT were determined from the right hemisphere motor cortex, participants were given a 10-minute break while the iTBS treatment staff opened the randomization envelope that indicated group assignment. Thereafter, iTBS or sham iTBS was performed over the left hemisphere using either the stimulation or sham Magstim Rapid2® coils, respectively, with intensity set at 80% of AMT obtained from the right hemisphere. Stimulation was targeted towards the residual language-responsive cortex in the left frontal lobe, typically at or near the inferior frontal gyrus (IFG). The precise location was based on fMRI results (see below) with the individual peak activation results entered into the Brainsight neuronavigation system (Rogue Research, Inc., Montreal, Canada). Stimulation parameters were selected based on Huang et al  and consisted of 600 iTBS pulses, with 3 pulses at 50 Hz given every 200 milliseconds in 2-second trains at 10-second intervals over a 200-second period . Each session took about 10–15 minutes, with participants monitored for adverse events during and after each session.
FUNCTIONAL MRI BLOCK-DESIGN TASKS: Participants completed 2 runs of a well-established semantic decision/tone decision (SDTD) task that was presented in 30-second blocks with 2 alternating conditions: the control (TD, tone decision) and the active condition (SD, semantic decision) [32–34]. Each run included eleven 30-second blocks, starting with a TD block, for a total of 330 seconds. In the TD condition, subjects heard brief sequences of four to seven 500- and 750-Hz tones every 3.75 seconds and responded with a left-hand button press (ie, index finger) to any sequence containing 2 750-Hz tones. In the SD condition, subjects heard spoken English nouns designating animals every 3.75 seconds and responded with the same left-hand button press to stimuli that met 2 criteria of being native to the United States and commonly used by humans (eg, cow). If the presented items did not fulfill the corresponding criteria, participants pressed the “no” button (ie, middle finger) also with the left hand. While performance was measured by recording responses to the SD and TD conditions, the overarching goal was to engage the brain language area that remained functional after the stroke to provide a target for stimulation [18,22].
MRI DATA ACQUISITION:
Data on the initial 3 participants were acquired on a 3.0 Tesla research-dedicated Philips MRI system using an 8-channel coil. For these subjects, EPI fMRI scans were performed using thirty-two 4 mm thick axial slices covering the entire brain. EPI images were obtained using a T2*-weighted gradient-echo EPI pulse sequence (TR/TE=2000/38 ms, FOV=24.0×24.0 cm, matrix=64×64, slice thickness=4 mm). In addition, a high-resolution T1-weighted three-dimensional anatomical scan was obtained (TR/TE=8.1/3.7 ms, FOV 25.0×21.1×18.0 cm, matrix 252×211, flip angle 8°, slice thickness=1 mm) for localizing brain regions. On the remaining subjects, we performed imaging on research-dedicated 3.0 Tesla MR Siemens systems, initially using a circular polarized head coil (Allegra) and a 20-channel head coil after scanner upgrade (Prisma). Echo planar imaging (EPI) fMRI scans were performed using thirty 4-mm-thick axial slices covering the entire brain. EPI images were obtained using a T2*-weighted gradient-echo EPI pulse sequence (TR/TE=2000/38 ms, FOV=24.0×24.0 cm, matrix=64×64, slice thickness=4 mm; after scanner upgrade, TR/TE=2000/35 ms, FOV=24.0×24.0 cm, matrix=64×64, slice thickness=4 mm). In addition, a high-resolution T1-weighted three-dimensional anatomical scan was obtained (TR/TE=2300/2.17 ms, FOV 25.6×25.6×19.2 cm, matrix 256x256, flip angle 9°, slice thickness=1 mm; after scanner upgrade, TR/TE=2300/3.37 ms, FOV 25.6×25.6×19.2 cm, matrix 256×256, flip angle 9°, slice thickness=1 mm) for localization of brain regions. For each fMRI run, 165 whole-brain scans were acquired.
MRI DATA PREPROCESSING AND STATISTICAL ANALYSIS: FMRI data preprocessing and modeling were completed using MATLAB toolbox SPM12 (http://www.fil.ion.ucl.ac.uk/spm/software/spm12/) [18,35–37]. The processing followed standard steps that included discarding the first 30 seconds of the control block, followed by co-registering and aligning all scans using the coregister and realign functions in SPM. Unified segmentation computed on the anatomical scan was used to normalize functional scans . Functional scans were then spatially smoothed with an 8 mm full-width half-maximum kernel, and general linear modeling (GLM) was performed using the fMRI time series from the block-design task as boxcar regressor convolved with the canonical hemodynamic response function (HRF). In addition, we utilized the 24-parameter model  to regress out head motion effects from the realigned data (ie, 6 head motion parameters, 6 head motion parameters 1 time-point before, and the 12 corresponding squared items) plus 2 regressors accounting for the number of runs. Group random effects were computed using one-sample t tests. Data were compared between time-points using paired t tests. To avoid confounds from participants’ individual brain lesions, combined lesion-frequency maps were used as a mask to exclude lesioned voxels from group statistical analyses. Finally, all imaging data analyses were co-varied for the type of scanner used in this study.
LATERALIZATION INDEX (LI): It is well recognized that chronic stroke directly affects language lateralization [40–42]. The LI is commonly used in functional neuroimaging studies to describe the hemispheric or regional distribution of activations in response to specific tasks (eg, language or memory). It ranges from −1 (pure right-hemispheric) to 1 (pure left-hemispheric) [42,43]. For each map, a threshold was computed using a bootstrap algorithm with values above an internal threshold added together to generate a global value for each hemisphere within a region of interest (mask) . Then, the LI was calculated using the LI-toolbox on contrast maps obtained by combining HRF and derivatives contrasts . Three different atlas-based masks were used for calculating LI, including frontal, cerebellum, and whole-brain (frontal+temporal+parietal) masks . For all masks, voxels within a 20-mm area around the midline (10 mm left and 10 mm right) in an axial plane, and the parts of the ROIs that were affected by the stroke were excluded from analyses . For each group, paired samples t tests were performed to examine changes between time-points.
LESION-FREQUENCY MAPS: A previously developed MATLAB plugin (R2017b, MathWorks) was used to compute the stroke-induced lesion area for each stroke patient . Probabilistic tissue segmentation and image algebra with naïve Bayes classification were used to create feature maps encoding information about missing and abnormal tissue. All maps were binarized then summed into 1 image (separately for each group) in which the value of each voxel represents the frequency of lesion at this particular cortical location (Figure 2).
CONNECTIVITY ANALYSIS: Generalized psychophysiologic interaction (gPPI) is an analysis for fMRI data that is conducted between regions of interest, is context-dependent, and is performed to assess dynamic changes over time. Exploratory gPPI analysis was conducted using the toolbox in SPM  to model context-specific changes (ie, changes in AT measures) in the relationship between activity in 1 seed brain region and activity in the other brain regions by including a term specifying an interaction effect between the seed region time series and the task time series in each first-level GLM . The gPPI effects are interpreted as changes in interregional connectivity that are driven by psychological states related to factors such as the AT measure [47,49]. This makes gPPI an appropriate tool for testing the hypothesis that longitudinal changes occur in functional connectivity of cognitive probes such as language tasks in response to iTBS. For each participant, the first principal component of the time series from each scan was extracted from the right and left anatomically defined IFG (WFU_pickatlas toolbox: https://www.nitrc.org/projects/wfu_pickatlas/) and entered as a seed time series for the gPPI analysis. Cerebrospinal fluid (CSF) and white-matter signals were included as nuisance variables in the gPPI model in order to reduce the influence of non-neural signals on estimates of task-dependent connectivity . Next, for each participant, gPPI estimates quantifying the level of condition-dependent connectivity from right and left IFG to the rest of the cortex during each session were extracted from the gPPI model into connectivity maps. These connectivity maps were used in paired t test analyses to compare connectivity between time-points within stroke groups.
RELATIONSHIP BETWEEN GPPI CONNECTIVITY, LI, AND BEHAVIORAL SCORES:
To investigate the relationship between the changes in connectivity and behavior over time, regression analyses were performed in SPM. For each participant and each seed region, a difference in connectivity between pairs of time-points was computed and regressed with a difference of each behavioral score between the same pair of time-points. Results were corrected for multiple comparison (FWE), with significance set a
BEHAVIORAL RESULTS: Behavioral results, including effect sizes (Cohen’s d), are presented in detail in Table 2A (due to the debilitating nature of their condition, some participants were not able to complete some of the behavioral tests; most of the effect sizes were medium or large. Missing scores are reported in Table 2B and mean and standard deviation scores are reported in Table 2C. One-way ANOVAs showed no significant differences between groups in any of the behavioral measures at t1 except for SFT (F=4.13, P=0.017). Post hoc tests showed that G1 and G2 had lower SFT scores compared to G0 (P=0.038 and P=0.02, respectively) but there was no significant difference between G0 and G3 or G1–3. Due to the small number of subjects, we analyzed the data in groups, and we also combined groups 1–3 into 1 treatment group (G1–3). Paired samples t tests were used to compare scores between time-points (t1, t4 and t5) for G0 and G1–3. Results are reported in Table 2A. For BNT, several of the groups, including G0 and G1–3, significantly improved between baseline (t1) and t4 or t5. The SFT scores increased between t1 and t4 for the combined group (P=0.046). The WAB-R AQ significantly improved between t1 and t5 for the combined group (P=0.007). The COWAT did not significantly change except at 1 point for the G0 group (t4–t3; P=0.03). Due to the pilot nature of the study, corrections for multiple comparisons were not performed.
TASK FMRI RESULTS: Second level paired t tests were used to assess fMRI (SDTD task) changes between time-points within groups on a voxel-wise basis. A cluster-wise FDR algorithm was used to correct for multiple comparisons in SPM (cluster P<0.05). Results are depicted in Figure 4 and Table 3. The BOLD signal significantly increased in the right lingual gyrus in G3 between t1 and t4. G1–3 analysis showed a significant decrease between t1 and t5 in left middle temporal gyrus and in right medial fronto-orbital gyrus.
LATERALITY INDEX RESULTS: LIs were computed for each participant, each fMRI, each session, and each region of interest (Figure 5). The LIs for the G1–3 did not differ from G0 at any time-point for any mask. Paired t tests revealed a significant increase between t1 and t4 for G3 (3 weeks of iTBS, P=0.02), with the cerebellum mask indicating stronger lateralization to the right cerebellar hemisphere between these 2 time-points. Also noted was a significant increase between t4 and t5 for G0 (P=0.04) with the whole-brain mask indicating stronger lateralization to the left cerebral hemisphere between these 2 time-points.
GPPI CONNECTIVITY RESULTS: The G0 exhibited decreases in functional connectivity for the left IFG as seed region: between t1 and t5, functional connectivity decreased in the right supramarginal, middle occipital, and angular gyri (Figure 6A). For the left IFG seed, between t4 and t5, functional connectivity decreased also in the right inferior frontal region (Figure 6A). With seed in the right IFG, between t1 and t5 functional connectivity decreased in the right pre- and post-central gyri and in left and right paracentral lobules (Figure 6B).
In contrast to connectivity decreases in G0, G1 significantly increased in functional connectivity between t1 and t5 both with left and right IFG as seed region in multiple cortical areas (Figure 6, Table 4). G2 showed decreased functional connectivity with left IFG as seed region between t4 and t5 in left middle occipital gyrus, left middle temporal gyrus, left lingual gyrus, left calcarine, and left frontal inferior operculum (Figure 6A, Table 4), and increased functional connectivity with right IFG as seed region between t1 and t5 in left lingual gyrus (Figure 6B, Table 4). Functional connectivity did not significantly change for G3 or for G1–3.
RELATIONSHIP BETWEEN GPPI CONNECTIVITY AND BEHAVIORAL SCORES: For the left IFG as seed, significant results were found for G1, G2, and G1–3, with a difference between t4 and t1 (Figure 7A, Table 5). For G1, a difference in connectivity in the left middle temporal gyrus and right precuneus was negatively correlated with SFT score. For G2, a difference in connectivity in left occipital inferior gyrus was negatively correlated with COWAT score. For G1–3, a difference in connectivity in the left superior frontal gyrus was negatively correlated with BNT score.
For the right IFG as seed, significant results were found between t4 and t1 for G1 and between t5 and t4 for G3. For G1 (Figure 7B), the differences in connectivity in left planum temporale end left precuneus were positively correlated with COWAT scores, respectively. For G3, a difference in connectivity in right anterior insula was positively correlated with BNT score.
RELATIONSHIP BETWEEN LATERALITY INDEX AND BEHAVIORAL SCORES:
No significant correlation was found between changes in LI in cerebellum and behavioral scores. Significant correlations were found for G0 and G2: for G0, changes over time in LI with frontal mask were significantly correlated with BNT (t1–t5, r=−0.95,
IMPROVEMENTS IN LANGUAGE SKILLS IN RESPONSE TO ITBS: Our observed improvements in AT in response to iTBS were variable when comparing groups G1–G3, with the most consistent improvements noted for the BNT (Table 2A), where the statistically significant changes were of medium or large sizes. In the combined G1–3 group, the BNT responses significantly improved after the intervention (t4), which was sustained at 3-month follow-up (t5). This is consistent with the results of a combined CIAT and iTBS intervention we recently reported [18,25] and implies that while iTBS can improve language in patients with chronic stroke-induced aphasia, additional treatments some weeks to months after initial therapy to maintain the gains may be needed. The notion of booster treatment is similar to the idea of a transfer package, which consists of a set of techniques designed to maintain and possibly improve gains from the initial treatment. A transfer package is commonly used in rehabilitation studies to reinforce maintenance and possible further improvement of the initial post-intervention function increases [52,53].
However, it is important to note for future studies that while the collected measures improved immediately after treatment (t4), there was some regression at follow-up (t5). We further note that the BNT remained improved for the 3 months after therapy (P=0.056), with the WAB-R score also significantly improved at the t5 time-point (P=0.007), indicating that some improvements may be sustained over time and may parallel the changes observed in the fMRI measures. It is also important to acknowledge that G0 showed some gains as well, indicating that some of the changes observed in BNT could be due to test-retest learning effects or spontaneous recovery. While some of the tests we used may be subject to practice effects, this was minimized by rotating different versions of each test. In addition, recent research has shown relative stability of the WAB-R when repeated over short periods . Finally, our AT results are consistent with the results of approximately 20 uncontrolled studies that have assessed short- and intermediate-term effects of TMS on language improvement in post-stroke aphasia patients .
Only 4 of the previous neurostimulation studies used an approach similar to ours (ie, applying rTMS to the lesioned rather than the intact hemisphere) [14,15,18,55]. In the first study, trends for language improvement were observed in 8 patients who received the same treatment protocol as in the present study . In the second study, low-frequency (1 Hz) rTMS was applied to either the inferior frontal or superior temporal gyri (based on the type of aphasia and the region most-activated with fMRI) followed by 60 minutes of intensive speech therapy . Participants in the study significantly improved in several aphasia measures. The third study’s participants were randomized to rTMS and language training vs sham and language training . Combined low-frequency rTMS over the non-dominant hemisphere’s IFG (inhibition) and higher frequency rTMS over the dominant hemisphere’s IFG (excitation) in 30 patients with post-stroke non-fluent aphasia, followed by speech/language training, resulted in improved language after active vs sham treatment in the language section of the Hemispheric Stroke Scale and other measures of post-intervention outcomes (eg, NIHSS). Finally, in our previous study that combined iTBS immediately followed by CIAT, the WAB-R AQ improved after 10 treatment sessions . While these studies have used different TMS stimulation parameters, guidance methods, targets, and additional interventions, their results converge with the current pilot RCT results, suggesting that neurostimulation can improve short- and intermediate-term language outcomes in patients with chronic post-stroke aphasia.
EFFECTS OF TMS ON LANGUAGE LATERALIZATION AND CONNECTIVITY IN POST-STROKE APHASIA: In our prior research on iTBS, we reported changes in fMRI language lateralization and activation in the dominant and non-dominant hemispheres after stimulation of the left-hemisphere targets, indicating both local and global effects of the stimulation . We have subsequently documented additional effects of iTBS on brain anatomy and structural connectivity in patents with post-stroke aphasia [23–25]. It is important to note that while the majority of the TMS intervention studies in aphasia used structural MRI to localize the stimulation area, only a few studies have used other neuroimaging methods for this purpose [15,18,22,56,57]. Inour studies, we have used the SDTD fMRI task to visualize the target for stimulation on pre-intervention imaging. Others have used either O15PET and a verb generation task , or fMRI or functional near-infrared spectroscopy (fNIRS) with a word repetition task [15,56].
Functional neuroimaging has proven effective for identifying the active language area in the peristroke area and in the contralateral unaffected hemisphere . Beyond targeting therapy, conducting follow-up neuroimaging allows exploration of the cortical changes in response to an intervention . Our study was specifically designed to test and is the first to report the effect of neurostimulation as the sole rehabilitative modality; other studies have used neurostimulation to prime the brain prior to a behavioral intervention. Here, although iTBS was the sole intervention, we showed cortical plasticity with non-significant shifts of language lateralization to the dominant (affected) hemisphere and concurrent shifts in lateralization to the right cerebellar hemisphere. Additionally, we showed significant correlations between changes in connectivity over time and changes in behavioral performance using gPPI for all active treatment groups, while these effects were not observed in the sham group (G0). Moreover, the vast majority of the significant results were found between t1 and t4, which are the time-points around the stimulation period. This suggests that TMS can have a positive impact on cerebral networks linked to behavioral performance and that the observed effect can partially dissipate with time after completion of the treatment, again suggesting the need for additional (booster) therapy to possibly maintain the gains from the initial intervention. Significant correlations were also found between changes in LI over time in frontal and whole-brain mask and behavioral scores. These results possibly suggest that laterality of cortical activity plays a role in the observed behavioral performance, which is a phenomenon previously observed in neuroimaging studies [7,60]. These findings underscore the importance of the canonical language regions for post-stroke language recovery [37,58,61]. They further underscore the importance of brain plasticity in post-stroke recovery, which is a concept that has been at the core of post-stroke rehabilitation [7,62]. Whether the shifts in activation patterns and eventual recovery rely on the peristroke areas, other areas of the ipsilateral hemisphere, or the contralateral hemisphere may depend on multiple factors including handedness , preservation of the white-matter tracts (bottlenecks) [63,64], age at the time of the stroke , lesion extent , intensity and duration of the intervention , or other currently unknown factors such as genetics .
This study’s limitations need to be considered for future neurostimulation and post-stroke aphasia rehabilitation studies. First, recruitment and retention in randomized trials are challenging, and although we identified the planned number of participants, a number of screening failures and dropouts decreased our sample size, which limited our ability to detect differences between groups and time-points and prevented a meaningful comparison of treatment duration effects. Blinding is important and was maintained through separating the treatment team from the testing/imaging team, and via the use of a sham coil. However, we did not ask the participants if they could determine whether they received experimental treatment vs sham, and it is possible that covert awareness of the form of treatment could have inadvertently unblinded them. Further, a “transfer package” is being used in many rehabilitation studies to reinforce maintenance and possible further improvement of the initial treatment effects . Such a package, whether behavioral or TMS-based, needs to be developed and validated for aphasia rehabilitation to sustain treatment gains. We also recommend measuring discourse productivity as a functional outcome in patients who undergo post-stroke aphasia rehabilitation . Participants who received sham treatment were noted to have improvements in BNT. This finding questions the validity of using BNT in rehabilitation studies and underscores the importance of validating measures for longitudinal use prior to selecting them as the primary outcome measures. Of note is that there were baseline differences between groups, including degree of aphasia and level of performance on AT. These differences may disappear when larger samples are included or may need to be controlled for in subsequent studies; level of education will need to be either controlled for or included as a co-variate in final analyses. Further, the AT results analyses were not adjusted for multiplicity and we cannot rule out the possibility of Type I error. However, the consistency of direction, magnitude, and timing of effects and their biological plausibility support those exploratory findings. Finally, in larger studies, stratification by potential confounders such as education, age, size and location of the lesion, aphasia type (eg, fluent vs non-fluent), and aphasia severity may increase the ability to observe between-group effects.
The results of this randomized, double-blind, sham-controlled pilot study support the hypothesis that neurostimulation, as the sole therapeutic approach, can improve post-stroke aphasia and induce short- and intermediate-term cortical plasticity in human brain networks involved in language function. Understanding the strengths and potential limitations of the current study will inform the design of future trials.
FiguresFigure 1. Diagram of the randomized, double-blind, sham-controlled treatment protocol, and associated testing (*AT – aphasia testing/AT – reflects AT without obtaining the Aphasia Quotient (AQ) of the Western Aphasia Battery-Revised (WAB-R), iTBS – intermittent theta burst stimulation; fMRI – functional magnetic resonance imaging; t1–5 – study time-points; G0–3 – study groups). Figure 2. Lesion maps for the 27 included stroke patients. Each voxel value is the number of participants whose stroke lesion extends to that particular voxel (all pictures in neurological convention – left in the figure corresponds to left in the brain). Top lesion map depicts all participants together, then divided by groups. Figure 3. Consort Statement (fMRI – functional MRI; TT – token test; F/U – follow-up; TMS – transcranial magnetic stimulation; ST – sham treatment). Figure 4. FMRI semantic decision/tone decision (SDTD) task result (general linear model). Paired t tests were computed between time-points for each iTBS group (contrast SD >TD). For every contrast, corrected data are provided (Cluster-wise FDR, P<0.05) with peak coordinates in Table 3. All pictures are in neurological convention (left in the figure corresponds to left in the brain). Figure 5. Results of Laterality Index (LI) analyses. Results are depicted for frontal mask (left), cerebellum mask (center), and whole-brain mask (right). Paired t tests revealed a significant increase between t1 and t4 for G3 (3 weeks of iTBS, P=0.02) with cerebellum mask, and a significant increase between t4 and t5 for G0 (P=0.04) with whole-brain mask. Figure 6. GPPI connectivity results. For each participant, the first principal component of the BOLD time series from each scan was extracted from the left (A) and right (B) inferior frontal gyrus (IFG) and entered as a seed time series for the gPPI analysis. Location of the BOLD signal changes is provided in Table 4. Red and Blue frames refer to increases and decreases over time, respectively. Figure 7. Results of regression analysis between change over time in gPPI connectivity coefficients and change over time in behavioral measures. Results for changes in the left IFG seed are shown in (A) and for the right IFG seed are shown in (B). Location of the BOLD signal changes is provided in Table 5. Red and Blue frames refer to positive and negative regression coefficients, respectively.
TablesTable 1. Demographic data of stroke participants included in the analyses (N=27). Table 3. Main peak coordinates of paired t tests on general linear model analysis results included in Figure 4. Significant differences in cortical activity were found for Group 3 (G3), between pre-treatment and post-treatment (t4 >t1) and for the combined group (G1–3) between pre-treatment and 3-month follow-up (t1 >t5). Table shows all local maxima separated by more than 1 mm. Table 4. Main peak MNI coordinates (x, y, z) for significant results of gPPI connectivity analysis. For each participant, the first principal component of the BOLD time series from each scan was extracted from the right and left inferior frontal gyrus (IFG) and entered as a seed time series for the gPPI analysis. Table 5. Main peak MNI coordinates (x, y, z) for significant results of regression analyses between changes over time in gPPI connectivity coefficients and changes over time in behavioral measures. Results are depicted in Figure 7.
1. Code C, Papathanasiou I, Rubio-Bruno S: International patterns of the public awareness of aphasia: Int J Lang Commun Disord, 2016; 51(3); 276-84
2. Boehme AK, Martin-Schild S, Marshall RS, Lazar RM: Effect of aphasia on acute stroke outcomes: Neurology, 2016; 87(22); 2348-54
3. Tippett DC, Hillis AE: Where are aphasia theory and management “headed”?: F1000Res, 2017; 6; F1000 F; aculty Rev-1038
4. Pedersen PM, Jørgensen HS, Nakayama H: Aphasia in acute stroke: Incidence, determinants, and recovery: Ann Neurol, 1995; 38(4); 659-66
5. Robey RR: A meta-analysis of clinical outcomes in the treatment of aphasia: J Speech Lang Hear Res, 1998; 41(1); 172-87
6. Johnson L, Basilakos A, Yourganov G: Progression of aphasia severity in the chronic stages of stroke: Am J Speech Lang Pathol, 2019; 28(2); 639-49
7. Stefaniak JD, Halai AD, Lambon Ralph MA: The neural and neurocomputational bases of recovery from post-stroke aphasia: Nat Rev Neurol, 2020; 16(1); 43-55
8. Stefaniak JD, Halai AD, Lambon Ralph MA: Predictors of poststroke aphasia recovery: A systematic review-informed individual participant data meta-analysis: Stroke, 2021; 52(5); 1778-87
9. Doogan C, Dignam J, Copland D, Leff A: Aphasia recovery: When, how and who to treat?: Curr Neurol Neurosci Rep, 2018; 18(12); 90
10. Dionísio A, Duarte IC, Patrício M, Castelo-Branco M: Transcranial magnetic stimulation as an intervention tool to recover from language, swallowing and attentional deficits after stroke: A systematic review: Cerebrovasc Dis, 2018; 46(3–4); 178-85
11. Norise C, Hamilton RH: Non-invasive brain stimulation in the treatment of post-stroke and neurodegenerative aphasia: Parallels, differences, and lessons learned: Front Hum Neurosci, 2017; 10; 675
12. Shah PP, Szaflarski JP, Allendorfer J, Hamilton RH: Induction of neuroplasticity and recovery in post-stroke aphasia by non-invasive brain stimulation: Front Hum Neurosci, 2013; 7; 888
13. Bates KA, Rodger J: Repetitive transcranial magnetic stimulation for stroke rehabilitation-potential therapy or misplaced hope?: Restor Neurol Neurosci, 2015; 33(4); 557-69
14. Khedr EM, Abo El-Fetoh N, Ali AM: Dual-hemisphere repetitive transcranial magnetic stimulation for rehabilitation of poststroke aphasia: A randomized, double-blind clinical trial: Neurorehabil Neural Repair, 2014; 28(8); 740-50
15. Abo M, Kakuda W, Watanabe M, Morooka A: Effectiveness of low-frequency rTMS and intensive speech therapy in poststroke patients with aphasia: A pilot study based on evaluation by fMRI in relation to type of aphasia: Eur Neurol, 2012; 68(4); 199-208
16. Barker AT: The history and basic principles of magnetic nerve stimulation: Electroencephalogr Clin Neurophysiol Suppl, 1999; 51; 3-21
17. Gamboa OL, Antal A, Moliadze V, Paulus W: Simply longer is not better: Reversal of theta burst after-effect with prolonged stimulation: Exp Brain Res, 2010; 204(2); 181-87
18. Szaflarski JP, Griffis J, Vannest J: A feasibility study of combined intermittent theta burst stimulation and modified constraint-induced aphasia therapy in chronic post-stroke aphasia: Restor Neurol Neurosci, 2018; 36(4); 503-18
19. Huang YZ, Edwards MJ, Rounis E: Theta burst stimulation of the human motor cortex: Neuron, 2005; 45(2); 201-6
20. Talelli P, Greenwood RJ, Rothwell JC: Exploring Theta Burst Stimulation as an intervention to improve motor recovery in chronic stroke: Clin Neurophysiol, 2007; 118(2); 333-42
21. Ziemann U, Corwell B, Cohen LG: Modulation of plasticity in human motor cortex after forearm ischemic nerve block: J Neurosci, 1998; 18(3); 1115-23
22. Szaflarski JP, Vannest J, Wu SW: Excitatory repetitive transcranial magnetic stimulation induces improvements in chronic post-stroke aphasia: Med Sci Monit, 2011; 17(3); CR132-39
23. Allendorfer JB, Storrs JM, Szaflarski JP: Changes in white matter integrity follow excitatory rTMS treatment of post-stroke aphasia: Restor Neurol Neurosci, 2012; 30(2); 103-13
24. Griffis JC, Nenert R, Allendorfer JB, Szaflarski JP: Interhemispheric plasticity following intermittent Theta Burst Stimulation in chronic poststroke aphasia: Neural Plast, 2016; 2016; 4796906
25. Allendorfer JB, Nenert R, Nair S: Functional magnetic resonance imaging of language following constraint-induced aphasia therapy primed with intermittent Theta Burst Stimulation in 13 patients with post-stroke aphasia: Med Sci Monit, 2021; 27; e930100
26. Szaflarski JP, Eaton K, Ball AL: Poststroke aphasia recovery assessed with functional magnetic resonance imaging and a picture identification task: J Stroke Cerebrovasc Dis, 2011; 20(4); 336-45
27. De Renzi E, Vignolo LA: The token test: A sensitive test to detect receptive disturbances in aphasics: Brain, 1962; 85; 665-78
28. Kertesz A:: Western aphasia battery-revised, 2006
29. Kaplan E, Goodglass H, Weintraub S:: Boston Naming Test, 1983
30. Kozora E, Cullum C: Generative naming in normal aging: Total output and qualitative changes using phonemic and semantic constraints: Clin Neuropsychologist, 1995; 4(9); 313-20
31. Lezak M:: Neuropsychological assessment, 1995
32. Karunanayaka P, Kim KK, Holland SK, Szaflarski JP: The effects of left or right hemispheric epilepsy on language networks investigated with semantic decision fMRI task and independent component analysis: Epilepsy Behav, 2011; 20(4); 623-32
33. Kim KK, Karunanayaka P, Privitera MD: Semantic association investigated with functional MRI and independent component analysis: Epilepsy Behav, 2011; 20(4); 613-22
34. Szaflarski JP, Binder JR, Possing ET: Language lateralization in left-handed and ambidextrous people: fMRI data: Neurology, 2002; 59(2); 238-44
35. Nenert R, Allendorfer JB, Martin AM: Neuroimaging correlates of post-stroke aphasia rehabilitation in a pilot randomized trial of constraint-induced aphasia therapy: Med Sci Monit, 2017; 23; 3489-507
36. Nenert R, Allendorfer JB, Martin AM: Age-related language lateralization assessed by fMRI: The effects of sex and handedness: Brain Res, 2017; 1674; 20-35
37. Nenert R, Allendorfer JB, Martin AM: Longitudinal fMRI study of language recovery after a left hemispheric ischemic stroke: Restor Neurol Neurosci, 2018; 36(3); 359-85
38. Ashburner J, Friston KJ: Unified segmentation: Neuroimage, 2005; 26(3); 839-51
39. Friston KJ, Williams S, Howard R: Movement-related effects in fMRI time-series: Magn Reson Med, 1996; 35(3); 346-55
40. Dietz A, Vannest J, Maloney T: The calculation of language lateralization indices in post-stroke aphasia: A comparison of a standard and a lesion-adjusted formula: Front Hum Neurosci, 2016; 10; 493
41. Saur D, Hartwigsen G: Neurobiology of language recovery after stroke: Lessons from neuroimaging studies: Arch Phys Med Rehabil, 2012; 93(1 Suppl); S15-25
42. Szaflarski JP, Allendorfer JB, Byars AW: Age at stroke determines post-stroke language lateralization: Restor Neurol Neurosci, 2014; 32(6); 733-42
43. Bigras C, Shear PK, Vannest J: The effects of temporal lobe epilepsy on scene encoding: Epilepsy Behav, 2013; 26(1); 11-21
44. Wilke M, Schmithorst VJ: A combined bootstrap/histogram analysis approach for computing a lateralization index from neuroimaging data: Neuroimage, 2006; 33(2); 522-30
45. Wilke M, Lidzba K: LI-tool: A new toolbox to assess lateralization in functional MR-data: J Neurosci Methods, 2007; 163(1); 128-36
46. Griffis JC, Allendorfer JB, Szaflarski JP: Voxel-based Gaussian naïve Bayes classification of ischemic stroke lesions in individual T1-weighted MRI scans: J Neurosci Methods, 2016; 257; 97-108
47. McLaren DG, Ries ML, Xu G, Johnson SC: A generalized form of context-dependent psychophysiological interactions (gPPI): A comparison to standard approaches: Neuroimage, 2012; 61(4); 1277-86
48. Friston KJ, Buechel C, Fink GR: Psychophysiological and modulatory interactions in neuroimaging: Neuroimage, 1997; 6(3); 218-29
49. O’Reilly JX, Woolrich MW, Behrens TE: Tools of the trade: psychophysiological interactions and functional connectivity: Soc Cogn Affect Neurosci, 2012; 7(5); 604-9
50. Bartels A, Zeki S: The chronoarchitecture of the human brain – natural viewing conditions reveal a time-based anatomy of the brain: Neuroimage, 2004; 22(1); 419-33
51. Oldfield RC: The assessment and analysis of handedness: The Edinburgh inventory: Neuropsychologia, 1971; 9(1); 97-113
52. Morris DM, Taub E, Mark VW: Protocol for a randomized controlled trial of CI therapy for rehabilitation of upper extremity motor deficit: The Bringing Rehabilitation to American Veterans Everywhere Project: J Head Trauma Rehabil, 2019; 34(4); 268-79
53. Rijntjes M, Haevernick K, Barzel A: Repeat therapy for chronic motor stroke: A pilot study for feasibility and efficacy: Neurorehabil Neural Repair, 2009; 23(3); 275-80
54. Duncan ES, Schmah T, Small SL: Performance variability as a predictor of response to aphasia treatment: Neurorehabil Neural Repair, 2016; 30(9); 876-82
55. Szaflarski JP, Vannest J, Wu SW: Excitatory repetitive transcranial magnetic stimulation induces improvements in chronic post-stroke aphasia: Med Sci Monit, 2011; 17(3); CR132-39
56. Hara T, Abo M, Kakita K: The effect of selective transcranial magnetic stimulation with functional near-infrared spectroscopy and intensive speech therapy on individuals with post-stroke aphasia: Eur Neurol, 2017; 77(3–4); 186-94
57. Heiss WD, Hartmann A, Rubi-Fessen I: Noninvasive brain stimulation for treatment of right- and left-handed poststroke aphasics: Cerebrovasc Dis, 2013; 36(5–6); 363-72
58. Griffis JC, Nenert R, Allendorfer JB: The canonical semantic network supports residual language function in chronic post-stroke aphasia: Hum Brain Mapp, 2017; 38(3); 1636-58
59. Eliassen JC, Boespflug EL, Lamy M: Brain-mapping techniques for evaluating poststroke recovery and rehabilitation: A review: Top Stroke Rehabil, 2008; 15(5); 427-50
60. Szaflarski JP, Allendorfer JB, Banks C: Recovered vs. not-recovered from post-stroke aphasia: The contributions from the dominant and non-dominant hemispheres: Restor Neurol Neurosci, 2013; 31(4); 347-60
61. Saur D, Lange R, Baumgaertner A: Dynamics of language reorganization after stroke: Brain, 2006; 129(Pt 6); 1371-84
62. Heiss WD, Kessler J, Thiel A: Differential capacity of left and right hemispheric areas for compensation of poststroke aphasia: Ann Neurol, 1999; 45(4); 430-38
63. Griffis JC, Nenert R, Allendorfer JB, Szaflarski JP: Damage to white matter bottlenecks contributes to language impairments after left hemispheric stroke: Neuroimage Clin, 2017; 14; 552-65
64. Middlebrooks EH, Yagmurlu K, Szaflarski JP: A contemporary framework of language processing in the human brain in the context of preoperative and intraoperative language mapping: Neuroradiology, 2017; 59(1); 69-87
65. Heiss WD, Thiel A: A proposed regional hierarchy in recovery of post-stroke aphasia: Brain Lang, 2006; 98(1); 118-23
66. Enard W, Przeworski M, Fisher SE: Molecular evolution of FOXP2, a gene involved in speech and language: Nature, 2002; 418(6900); 869-72
67. Medina J, Norise C, Faseyitan O: Finding the right words: Transcranial magnetic stimulation improves discourse productivity in non-fluent aphasia after stroke: Aphasiology, 2012; 26(9); 1153-68
22 September 2021 : EditorialEditorial: Global Regulatory Initiatives Deliver Accelerated Approval of the First Bispecific Therapeutic M...
Med Sci Monit In Press; DOI: 10.12659/MSM.934854
08 September 2021 : Clinical ResearchAcceptance of COVID-19 Vaccination and Its Associated Factors Among Cancer Patients Attending the Oncology ...
Med Sci Monit In Press; DOI: 10.12659/MSM.932788
16 July 2021 : Review articleSilent Hypoxemia in Patients with COVID-19 Pneumonia: A Review
Med Sci Monit In Press; DOI: 10.12659/MSM.930776
23 Sep 2021 : Animal ResearchForkhead Box Protein M1 Promotes Nasopharyngeal Carcinoma Cell Tumorigenesis Possibly via the Wnt/β-Catenin...
Med Sci Monit In Press; DOI: 10.12659/MSM.931970
22 Sep 2021 : EditorialEditorial: Global Regulatory Initiatives Deliver Accelerated Approval of the First Bispecific Therapeutic M...
Med Sci Monit In Press; DOI: 10.12659/MSM.934854
22 Sep 2021 : Clinical ResearchA Comparative Study of the Clinical Efficacy and Safety of Pneumatic Trabeculoplasty and Selective Laser Tr...
Med Sci Monit In Press; DOI: 10.12659/MSM.933454
20 Mar 2020 : Clinical ResearchSocial Capital and Sleep Quality in Individuals Who Self-Isolated for 14 Days During the Coronavirus Diseas...
Med Sci Monit 2020; 26:e923921
15 Apr 2020 : Clinical ResearchPsychological Impact and Coping Strategies of Frontline Medical Staff in Hunan Between January and March 20...
Med Sci Monit 2020; 26:e924171
05 May 2020 : Review articleAn Evidence Based Perspective on mRNA-SARS-CoV-2 Vaccine Development
Med Sci Monit 2020; 26:e924700